MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

[BACK AGAIN] Eurorack: Haible Living VCOs by R*S
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author [BACK AGAIN] Eurorack: Haible Living VCOs by R*S
nordlead
Hey guys,

Jürgen Haible's Living VCOs main pcbs - including TRI, SINE, SYNC per VCO as well as the Driver Section and Haible's amazing VCA/Inverter/Overdrive module) are already and officially available again.

Now it looks like the front panel and matching panel component pcb for Euro are ready to go in production. Based on some of your comments I decided to drop the "Fine" and "Coarse" labels altogether, so now there's total freedom of whether you use one or two large knobs for those instead of the normal Davies:



As the main pcb is 8"(!) wide (and pretty dense), the front panel is 42hp. Here's a picture of the component pcb prototype:



Front panels will be 2mm aluminum, screen printed for a nice black and consistent look with the Serge Euro range by R*S,
pcbs made in Europe. Due to the large format, production is not cheap, I'm aiming for € 84 for a set of front panel + component pcb (8" x 4.1") - not including the main pcb.

If everything goes well, these can be shipped before Xmas. w00t

Not accepting orders yet, but please feel free to express comments / interest here - thanks! Guinness ftw!

Made some demos showing the new features a little while ago:

[soundcloud url="https://api.soundcloud.com/tracks/221578572" params="auto_play=false&hide_related=false&show_comments=true&show_use r=true&show_reposts=false&visual=true" width="100%" height="450" iframe="true" /]

[soundcloud url="https://api.soundcloud.com/tracks/216023073" params="auto_play=false&hide_related=false&show_comments=true&show_use r=true&show_reposts=false&visual=true" width="100%" height="450" iframe="true" /]
Joe.
No wiring at all?
Grumskiz
Very nice! I love that you added Sync and the VCA looks like a nice addition smile
Hope I will be able to afford one of these after christmas ^^;
nordlead
LoFi Junglist wrote:
No wiring at all?


Exactly - the main pcb is a monster, wouldn't be too keen on wiring this by hand ... so the component pcb takes care all wiring screaming goo yo
PWM
What Is the spacing between the jack sockets? 12,7 mm?
Jarno
nordlead wrote:
Exactly - the main pcb is a monster, wouldn't be too keen on wiring this by hand ...


Exactly, this is the reason why a fully stuffed PCB has been sitting in a drawer for 5 years already.
Still am going to finish it though (g*df*ckingd*mmit).
PWM
Go, Jarno! We believe in you! (He said thinking about his, yet to populate PCB..)
nordlead
PWM wrote:
What Is the spacing between the jack sockets? 12,7 mm?


Horizontal space between the Thonkiconn jacks (within one section) is 11 mm.
Altitude909
Yay! Finally done right. I'm in for one..
cane creek
Im interested, as long as I've still got some money over crimbo Mr. Green
oberkorn
yes! I'm in for one set
sumdumguy
I am interested. thanks.
Reality Checkpoint
Yes please, 1 set for me. thumbs up
fonik
LoFi Junglist wrote:
No wiring at all?

no wiring at all.

i had the privilege to built the prototype. it is a smooth build!

i hope to do a few sound snippets over the weekend.
audiohawk
Damn! Another one I can't resist!
Count me in for one set...
Thank you for your work we're not worthy
woodster
Really looking forward to getting hold of one of these thumbs up
DabiDabDab
Standard question; can i get only panel component pcb no panel?
simfonik
So that connects with this, right?

http://randomsource.net/haible/living_vcos

I'm in for 1 set.
apoisontree
I hope there will still be main PCBs with transistors arrays available by the time you the panel and control boards.

I'd like to avoid paying 9EUR postage twice, if possible.
hellpony
Im in for one set!
ablearcher
oh man, my backlog is already threatening to bury me without adding another giant project... but I'm probably in, this seems like one of those too good not to build ones.
evengravy
I'll take two sets of panel plus panel pcb to match the PCBs I have,

They look stunning, very well done

applause

edit; by the way, any chance of a sneak peek at the BOM so that I could preempt it with a component order?
PWM
BOM is on the site.
http://randomsource.net/haible/lvcos/LivingVCOs_BOM.pdf
evengravy
Yep, got that one thanks though, I meant a bom for the pots/jacks. I have the vco parts and pcbs. I'm assuming 9mm alphas but values. I assume 1M log (1 per osc) plus 100Ks everywhere else but just wanted to check. Also thonk jacks?
limpmeat
interested.....
cyphersum
Definitely interested in getting one!
ZatsuOn
I'm in for a set!
vicpop
Very interested
YannickD
totally interested in this!
jonesboy
I'm also very interested.
spotta
Eagerly waiting for ordering to open thumbs up
AlanP
Hopefully when they come out I won't be quite so buggered for money smile
search64
Tell us more about that Serge Resonant Equalizer smile applause
hpsounds
search64 wrote:
Tell us more about that Serge Resonant Equalizer smile applause


love eek! hyper Guinness ftw! thumbs up

H.
sy_arp
I WOULD LIKE A PANEL & PCBS PLEASE THANKS praise randomsource
calaveras
interest expressed!
Reality Checkpoint
Does anyone know the number and values of the pots required for this? help

(Would like to get an order in with Tayda before the discount window closes!)
didahdrieghe
I'll have a set please!
donttouchmyrice
interested
muffdiver
Would love to grab one upon restock.
didahdrieghe
Does anyone have a saved mouser cart for this project?
If not I'll make one and post it.
antf4rm
Count me in for 1 full set please! Looks great!
YannickD
didahdrieghe wrote:
Does anyone have a saved mouser cart for this project?
If not I'll make one and post it.


Would be much appreciated, currently doing a Reichelt cart. Will post it once it's complete.
nordlead
Front panels and component pcbs are available now w00t



Unfortunately the main pcbs have already been sold out, but should be back in about 3-4 weeks - but I doubt that the panels will last till then ... d'oh!

Speaking of the main pcbs, thinking about doing a small batch with 3x THAT300 soldered in - unfortunately the THAT candy is not cheap, so price would end up around € 60 per (main) board (including all the bypass caps in SMT). Please let me know if there is interest - thanks!
antf4rm
Pm me and I'll pay now for a panel and control pcb.
antf4rm
Definitely interested in a deluxe main pcb with expensive candy.
spotta
Is there going to be a build thread - or will this turn into the build thread?
audiohawk
Panel & component pcb arrived blazing fast.
Looking great!
Thank you for your work. we're not worthy
HipDestroyer
I would love to have a go at this beast with the smd components on board and both pcbs.
spotta
Started on mine now thumbs up
How 'confirmed' are the +/-12 value substitutions - are there any that would be a good idea to socket for now to try out different values?
spotta
More questions!
I choose the THAT300 option over the BCM847DS for no other reason than THAT chips seem to have a good reputation. Neither have been fitted yet, so I'm still free to choose, can anyone tell me what differences I can expect between the two choices? I'm guessing there must be a difference otherwise why the choice.
Also the BC550/560 - Should these be matched?
AlanP
Got a few THAT300's on the way as well, and what orientation are they? The silkscreen is not clear on this...
spotta
I think pin 1 has a little mark on the inside of the pad - it almost looks like a mark on the tempco silkscreen marking. hopefully Ralf can confirm.
Altitude909
So is this the build thread? I have my Mouser BOM that I can share..
shabax
Altitude909 wrote:
So is this the build thread? I have my Mouser BOM that I can share..

Sounds like it is.
Please do share.
Hazium
Altitude909 wrote:
So is this the build thread? I have my Mouser BOM that I can share..



I'm into taking a look at that. What did it end up costing?
Altitude909
$166 with no pots, jacks, and DIL power header. Everything else for the voice board and euro control surface are there.

http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=bd822 dece0
didahdrieghe
Altitude909 wrote:
$166 with no pots, jacks, and DIL power header. Everything else for the voice board and euro control surface are there.

http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=bd822 dece0


You rock Altitude909! Have you built yours?
Altitude909
^waiting for parts

It's a pretty straightforward BOM, I opted for the super-matched pairs and nicer opamps where it made sense. Dont forget pots, I use the smallbear alphas so they are not included
nordlead
Hey,

just found the notifications of the last 12 posts or so in my Spam-folder! Sorry about not having seen this.

Picked up another batch of main pcbs today - including 3 boards with MAT14 w00t - probably overkill, but the specs of the MAT14 are so impressive, thought this is worth a try ...



Why use super-matched transistors? There should be some improvement in tracking (due to better log conformity) and possibly temperature drift (due to better matching) - whether you notice either in real life is hard to say, also whether it's worth the added $$$ ... might not make any sense at all, still I know I need to build me a MAT14 version!
Altitude909
Working on mine now. First BOM Error, TEMPCO, says quantity is 1, it is 3
Altitude909
Question: On the control surface PCB, Should I populate just the headers that have a foot print? The long one has 35 holes but the footprint is 33 holes and the two extra are connected to the 12 pin one to its left. I assumed this was since a 35 pin header as easier to find than a 33 pin but that's not really the case since no one has odd numbered ones that long anyway (My bom has a 36 pin)
nordlead
Altitude909 wrote:
Question: On the control surface PCB, Should I populate just the headers that have a foot print? The long one has 35 holes but the footprint is 33 holes and the two extra are connected to the 12 pin one to its left. I assumed this was since a 35 pin header as easier to find than a 33 pin but that's not really the case since no one has odd numbered ones that long anyway (My bom has a 36 pin)


Yes and no - I generally use those precision sockets (headers) that can be cut to length, so the 35 pins is not related to part choices.

Those 2 extra pins you noticed are not required, but an option (that's probably not relevant to most builders) - these two pins bring out the 4th (unused) input of the mixer in the driver section, in case someone wanted to experiment / hardwire another signal into that mixer. The outmost (lowest) pad is the actual prolonged input (it's a GND-pad next to it).

To connect the panel pcb to the main pcb, only the pads that exist on both pcbs should (can) be used - I recommend actually stacking the boards together with the pin strips and connectors without soldering first and check if everything is correct / correctly aligned. Also if you use precision sockets, it will be very hard to disconnect the boards again (it is possible, but go very slowly, mm by mm), so (finally) glueing the boards together is done at the very end - e.g. solder those connectors only onto one of the pcbs and leave the other side unsoldered till the end.

Hope this helps!
Altitude909
Thanks.

One other thing: Are all the SMD bypass pads supposed to be populated? There are two on the back of my board that dont have parts on them (coming off the power header to the VCOs)
nordlead
Altitude909 wrote:
Thanks.

One other thing: Are all the SMD bypass pads supposed to be populated? There are two on the back of my board that dont have parts on them (coming off the power header to the VCOs)


No need to worry about the SMD pads (those pads you mention aren't for bypass caps, but an alternative footprint for SMT beads. I suggest to go with normal (TH) beads).
Altitude909
sorry, I wasnt clear. I ordered the board WITH the BCM matched pairs and the 1206 bypass parts soldered. Everything on the back has parts already except those two..
nordlead
Altitude909 wrote:
sorry, I wasnt clear. I ordered the board WITH the BCM matched pairs and the 1206 bypass parts soldered. Everything on the back has parts already except those two..


(?!) No, that's exactly what I understood / assumed. Your version has all the SMT you need already installed w00t , just ignore these 2 (optional / extra) SMT pads.
Altitude909
cool. thanks!
Altitude909
alright, results!

Notes on my Mouser cart:

1; 1uF Caps are on the large side, if someone has something else that's available at mouser, I'll switch them. They fit the foot prints but CT2a/b/c is a different footprint and that one is tight,

2: Ferrite beads are big but could be made to fit like I did. Those are my go to parts and I'm not going to switch them out unless someone has a proven part that is resistor size.

3: The TI TL074 parts are abnormally long for some reason and wont fit without a socket and that's all the Mouser stocks in PTH. I lifted mine so they float about the resistor and diode at each end


Hazium
This looks like an involved build. 5+ hours I'm sure. It looks like you are soldering the IC's directly to the board. I'm new to this so I don't know all the techniques. Are there benefits to doing this, other than some saved $?
Altitude909
Not that long, it's a lot of triples so it goes quick. 4 hrs maybe. I prefer to solder the chips down, if you're not confident in your soldering, use sockets.

Are calibration instructions in the works? The JH page is a bit terse.

Also, there is no mention of where the -12V pin is, I needed to look this up since I have not dealt with a MOTM connector before and if one should use a keyed header for the euro power connector, it will be opposite of the MOTM connector footprint silkscreen
spotta
Hazium wrote:
This looks like an involved build. 5+ hours I'm sure. It looks like you are soldering the IC's directly to the board. I'm new to this so I don't know all the techniques. Are there benefits to doing this, other than some saved $?

Soldering directly is fine until you have to remove/replace/troubleshoot. Then that socket earns the 20pence it costs 1000 times over
thumbs up
spotta
What's the orientation of Q4 & Q5? or what does the square hole represent?
nordlead
spotta wrote:
What's the orientation of Q4 & Q5? or what does the square hole represent?


Square pad = Emitter (followed by B C )
spotta
nordlead wrote:
spotta wrote:
What's the orientation of Q4 & Q5? or what does the square hole represent?


Square pad = Emitter (followed by B C )
thumbs up
Monobass
Now in stock at Thonk, a set including the Panel and both PCBs.

https://www.thonk.co.uk/shop/living-vcos/
PWM
Is the kit available without the main PCB? I have that one already.
spotta
so is CD31 meant to be a styrene?
nordlead
spotta wrote:
so is CD31 meant to be a styrene?


C0G or Film should be fine, if you can find 47N in Styrene (and fit it in here), it will work too, of course, but I'd probably rather use it elsewhere (e.g. Serge Resonant EQ!)
spotta
Cheers Ralf. I only have film in the parts box for that value ;(
qp
Completed and calibrated everything last night. All went pretty smooth for the most part. I had forgotten to solder in the tempo's initially so the scaling was crazy off.

altitude: CD31 (47N cap) is missing from your BOM. Thankfully I had one laying around. It's nothing fancy but it does the trick.

FYI for euro power: red stripe lines with the bottom of the header


Some notes on calibration:

Setting up the v/oct took a bit to get to work with me. Usually I start at a low frequency and with a guitar tuner I adjust the vco for C, then play 1 octave up and adjust the scale. For this, it was fighting me the whole time. Oddly what I found to work better was playing the 1 octave higher C, tune the VCO then play 1 oct lower C and adjust the scale. I didn't have to mess with the high freq adjustment much, but I did a bit.

I set the Tri level to match the Saw output with an oscilloscope.

For the Sine Round and Symmetry, you really need a scope to set it proper. Adjust both until you get a nice rounded sine.

The PW adjustment is P40T1-3. Be sure to adjust it to the PWIDTH pot and not the PWM pot, as I did and was cursing the lack of range. Adjust for 50% duty with pot at 12 o'clock.

The OCT DN & OCT UP screenprint are backwards. Adjust the DN for UP and the UP for DN. One thing I notice, which may be part of the design, is when I switch between octaves, there is a glide applied to the voltage, so it doesnt snap in octaves.

I didn't mess with the VCA much. With the scope I didn't notice much of difference when I adjusted the VCA trimmer. For the VCA CV I turned the drive down all the way and applied a looping envelope to the CV. There's a point where the VCA starts to bleed when adjusting. I went to that point and just backed it down a bit. It's very possible that I'm doing this part all wrong, so someone correct me if so.

Couple of things I would like to modify eventually:

The fine tune is reaaaaally fine. Like only a couple of cents. I'm assuming the 1.5M resistor is for the fine tune, but not certain. If so, I would like to drop this down to at least a 1M. (EDIT: I checked the build doc and the 1.5M is for the ratio between coarse and fine, so a lowering the ohms should spread the ratio out more. I haven't tested this yet though). [EDIT] I added a 1M in parallel (soldered to the back side) which changes the value to 600k. I tried several values and this was what seemed to work best for me. It's still a pretty tight spread but with a bit more play to keep from having to rely on the coarse tune as much.

The Variable Depth knob should be a different taper. It does nothing almost half way then it kicks in and has a very small sweet spot. I'm thinking a 10KC (reverse audio) may be better suited here. [EDIT]I followed Dave Brown's mod of adding a 10k resistor between the center lug of the pot and ground. This does help open the sweet spot up a bit more and is a better altertnative than removing the panel just to swap out the pot. I'd still suggest trying different tapers if you haven't mounted the panel yet.

I installed the TUNE pot in the normal position rather than "reversed" and the pot works High CCW and low CW, so the "normal" seems to be the reverse of what I would expect. I'll probably leave it that way as I really don't feel like taking the panel off now, but keep it in mind when doing your build.

If I did anything wrong here please let me know. Hope this info helps! Guinness ftw!



spotta
Good to know, some of that info will help me out, a couple of bits would have been nice to know last week before I soldered all my pots wink
Just waiting for the last few components that I didn't have here and then we'll see how I get on thumbs up
Altitude909
BOM fixed. Thanks
qp
Totally forgot I have a box of rogans, so finally finished it up today. I keep those reserved only for the poshest of modules. love


PWM
Looks great, qp!

Just ordered the panel/panel PCB and a triple chorus PCB (have 3 TDA1022s for ages)
My backlog is taking on epic proportions since I joined Muff's.. eek!
qp
PWM wrote:
Looks great, qp!

Thanks! Guinness ftw!

considering the complexity of this module, it was one of the most painless builds. Top notch design by R*S! we're not worthy thumbs up

took it for a spin the other day. MY ASS IS BLEEDING
sduck
I've been trying to find a wiring diagram that corresponds to the current main pcb. Or work one out, but some things are being difficult. Does a schematic for the adaptor board exist?
artQ
My first post here ... So hello folks.

Just wanted to say thanks to altitude909 for the BOM and qp for the pics n cal infos.

This will make things much easyer.

Got the stuff from mouser allready ... so ready to go :-)

Keep up the good work
spotta
Done, have the 2n7 styrenes on back order, so I deliberately fitted the cheapest alternatives had in stock as I've always wondered if my ears could tell the difference between 3 caps that cost .7 pence for three and 3 caps that cost 1000 times more!

Does anyone know the power specs for this? I would like to test and calibrate on my module tester (max 200ma)
nordlead
Please pay attention to the orientation of both power connectors:

+V is at the top, -V at the bottom / edge of the board:



The board marking (indicated cutout) is incorrect / misleading for euro power connectors (10 pin connectors).

Make sure you (double) check before powering up the board!

Big thanks to spotta for pointing this out! thumbs up thumbs up thumbs up
spotta
qp - did you have any issues with dc offset on your waveforms?
Mine all have some offset, the saws being the worst
antf4rm
That offset is not good. Hopefully you figure that one out.

I'm going to start building this week. Just got my voice pcb.
PWM
Spotta, it looks like your scope is in DC mode. Are you measuring DC there?
spotta
Switched to AC mode, still have offset on all outputs, here is the pulse

(Not sure why it's put it upside down, it's the correct way up on my iPad!)
spotta
antf4rm wrote:
That offset is not good. Hopefully you figure that one out.

I tried a few things that Ralf suggested, in the end decided to go ahead with calibration, it's all about what they sound like that matters.
PWM
So? How do they sound?
iirussell
I'm tempted to try this. Would you consider it an intermediate or advanced project?
spotta
PWM wrote:
So? How do they sound?

The three of them beating together is lovely. I need to carry out the fine tune mod that qp mentions though, havent got round to any other calibration yet either.

iirussell wrote:
I'm tempted to try this. Would you consider it an intermediate or advanced project?

If you can solder than personally I would say its quite an easy project to build with the component panel and Mouser carts already posted, calibration however is another matter.
PWM
Never mind..
sduck
iirussell wrote:
I'm tempted to try this. Would you consider it an intermediate or advanced project?


If you've got solid soldering skills it's an intermediate project. There's a lot of tight soldering though, you don't have room for sloppy soldering. Other than that it's just a LOT of stuff to solder.
antf4rm
evengravy wrote:
Yep, got that one thanks though, I meant a bom for the pots/jacks. I have the vco parts and pcbs. I'm assuming 9mm alphas but values. I assume 1M log (1 per osc) plus 100Ks everywhere else but just wanted to check. Also thonk jacks?


I would still like to see a bom for the pots. I don't see a listing of values for the oscillator pots anywhere. seriously, i just don't get it

B100K - 1
A100K - 1
A1M - 1
B10K - 2
???? - 18

The original schematic shows a 10k 10turn pot for the freq of each oscillator.

I'd love to make the smallbear order....
spotta
http://randomsource.net/haible/lvcos/RandomSource_LVCOs_Euro.pdf
thumbs up
antf4rm
spotta wrote:
http://randomsource.net/haible/lvcos/RandomSource_LVCOs_Euro.pdf
thumbs up


thanks i missed that.

Dang smallbear is out of 10k lin maybe ill get em from thonk with the thonkiconns
iirussell
iirussell wrote:
I'm tempted to try this. Would you consider it an intermediate or advanced project?

If you can solder than personally I would say its quite an easy project to build with the component panel and Mouser carts already posted, calibration however is another matter.[/quote]

Good to hear, I'll try it in the near future then.
guitarfool
Okay, I started populating the main board today, and completely bogged down. There are errors in the resistor section of the "LivingVCOs_BOM.pdf".
I figures the board designations followed Juergen's schematics (page 1) with A,B and C suffixes, and that the same applies to the additional schematics for the wave shaper circuits, but I'm finding too many exceptions, and I really don't feel like tracing the board to try to match board designations with the schematics. very frustrating

How did you guys get through this? Is there another BOM that matches the board designations?
qp
guitarfool wrote:
Okay, I started populating the main board today, and completely bogged down. There are errors in the resistor section of the "LivingVCOs_BOM.pdf".
I figures the board designations followed Juergen's schematics (page 1) with A,B and C suffixes, and that the same applies to the additional schematics for the wave shaper circuits, but I'm finding too many exceptions, and I really don't feel like tracing the board to try to match board designations with the schematics. very frustrating

How did you guys get through this? Is there another BOM that matches the board designations?


The original design is for +\-15v. The R*S BOM has alternatives marked for +\-12v: http://randomsource.net/haible/lvcos/LivingVCOs_BOM.pdf
guitarfool
Yes, I am aware of all that. That is the same document I'm talking about. But where are resistors RS10, R49, RS19, RS23, RS24 (for example) on the PCB? The RS resistors (from the Sine shaper, I presume) only go to 8 on the schematic. If you look at the BOM 10k section, there is "... RS7, RS7A, RS8A, RS8B, RS19...". It should be RS7A, RS7B, RS7C, RS8A, RS8B, RS8C ... but there is no RS19 anywhere on the board, or an RS19 or R19 either.

My concern is that some of the designations on the PCB that appear to match the schematics, don't.

Is there another BOM, or a value designated version of the board layout somewhere?
sduck
I found it really hard to make out everything on the pcb itself, with my relatively old eyesight, but fortunately there's a LARGE PRINT version on the r*s site - http://randomsource.net/haible/lvcos/LVCOs_top_names_updated.png

I can't find RS19 either, although I have it checked off as installed on my board, somewhere. Wonder what I've done wrong...
guitarfool
Yeah, I've been using that too. I guess I'll just keep muddling through. When I finish (and it's working), I think I'll correct the BOM and report back here.
antf4rm
guitarfool wrote:
Yeah, I've been using that too. I guess I'll just keep muddling through. When I finish (and it's working), I think I'll correct the BOM and report back here.


Please do!
sduck
Ok, don't have a cow, man. Ok ok, you can go ahead and have a cow, I just always wanted to use that phrase for some reason.

Figured it out. It puzzled me deeply, as I know I counted out 35 10K resistors and found correct places to put them all, although a quick look at that reference sheet was bamboozling me, as many of those resistors weren't on it. The trick is, I have a prototype pcb.

It seems the prototype boards used some different reference designators than the final production boards. But the BOM never got updated, at least the one currently available, to reflect that change. So, the easy way to fix this is to use the old reference sheet, which is listed first on the r*s website - http://randomsource.net/haible/lvcos/LVCOs_top_names.png - and put the resistors where it says they go. Make sense?

I only checked out the 10K's, as that's where most of the discrepancies seemed to be. Here's where they go...

guitarfool
Ahhh, mystery solved! I figured there was some reason why no one had reported issues before, so I'm not surprised that the board designations had changed but not the BOM. I'll just use the old designation sheet that you so thoughtfully provided instead of what is on the board itself. Thanks!

But why can't I have a cow? I've heard that all of my life and always wondered, does it mean "give birth" to a cow, "acquire" a cow, or have one in the biblical sense? Miley Cyrus
sduck
Bart Simpson Bart Simpson Bart Simpson Bart Simpson Bart Simpson Bart Simpson Bart Simpson

dunno it's a simpsons thing I never watch that show

Bart Simpson Bart Simpson Bart Simpson Bart Simpson Bart Simpson Bart Simpson Bart Simpson
antf4rm
My silkscreen also matches this image. Cool, now i know what to do. Thanks guys.

[/quote]
apoisontree
How do you solder the tempcos with the BCM847DS ?
sduck
Like this? You'd probably be better served by using smaller tempcos, but these big ones, left over from when I got them for the original LVCO build, were what I had -



(note that I'm building a non-euro version, so my headers are on the "wrong" side)
antf4rm
Just successfully soldered these guys in. Those are the PMP4201Y matched pairs for QM1, QM, QMC. Dang those are small! That was scary eek!
You can test them by probing the holes for the alternative BC550C at Q4A, Q5A... to make sure they are not solder bridged.

apoisontree
@sduck: so you just covered one BCM847DS with the tempco for each core.
sduck
Yes - that's all I could figure out to do with it. If anyone has a better idea, I'd love to know!
Styrofoam
All done and working! Debugging turned out to be nothing more than two unsoldered joints. Sounds and looks gorgeous.

Now on to calibration.

sduck
I'm just about done fixing stuff and calibrating my LVCOs. Note that I actually built a 5U version that uses 15v, but most of what I found should work for 12v versions also - so these are just some addendums to Qp's excellent post:

Calibrating the 3 VCOs - I didn't do anything to the 3 TR8 - Freq trimmers. I think they're just general offsets so you can set a standard lowest pitch for all 3 VCOs if you're so inclined. Mine all go to the same almost-a-pitched note but not-quite-a-LFO frequency by default, so didn't experiment further.

I used qp's method of working down in pitch to get started with the scale calibration, which works well. In a separate session a day later I was able to fine tune them the normal way, and get them to track reasonably well up to 5 octaves. I did have to do some trimming of the HF trimmers to really get a somewhat linear kind of tracking - I found the pitch going a bit flat at 2 octaves, but a bit sharp at 4 octaves up, depending on which octave I trimmed it for. I turned the HF trimmer a bit in the CCW direction, and it seemed to even the curve out a bit. This may well work differently for 12v - experimenting is the only to find out. It takes forever to get these things calibrated - plan accordingly. My studio cat really hated all this.

Oh - and even really rigorous calibration is only going to go so far - these things are unstable by design. If you calibrate out of the cabinet, it's quite likely that keeping it in a warm cabinet is going to throw things off a bit. That's just the way of this design. If you need an ultra stable VCO, get a Synthesis Technology MOTM 300.

The linear detune knob is wired backwards. At least it is if you're anal about wanting all your tuning knobs to work in the same direction. Considering that it only changes the pitch less than 10 cents total, it's not really a big deal.

The fine tune knob - that 1.5M resistor is just way wrong. It basically makes the fine tune knob work the same range as the detune knob. I tried a 1M resistor instead at first, but still found it to not be nearly enough range to be useful. So I clipped in a 1M pot across that connection, and experimented with it to see what value actually worked how I wanted it - a whole step up or down from the middle position. 75K turned out to be exactly right. Didn't have any 75K resistors, so I used 100K, and it's fine. Now I don't have to be super careful with the main tune pot before dialing in the exact pitch with the fine tune pot.

The pulse width knob doesn't seem to work correctly, although I double checked the schematic and it seems to be identical to JH's design, so I guess it's working the way it's supposed to (I had some weird issue with the square wave circuit that I had to kind of work around, so my experience with this may be a fluke). But, if I set the trimmer to the duty is right in the middle at 12 oclock on the knob, I lose the signal completely at about 10 oclock. If I set the knob full CCW, and adjust the trimmer so I just start getting a spike (and some sound), then the actual center point for the duty cycle on the knob is about 10 oclock. Not really a big deal. I also found that generous amounts of cv will still knock the spike off the scale, so to speak - the sound will die out. I guess it's just a matter of using the PWM pot carefully.

The VCA drive knob is weird, and may be something that works differently in a 12v system - I haven't experimented with it yet. What I'm experiencing is that you can use that part of the circuit as a VCA, or a drive circuit, but not both. It works well as a basic VCA, but once you start dialing in the drive, the cv you're applying stops working as a vca, and turns into more of a modulation sort of thing. Not how I was thinking it was going to work, but still fun and useful.

As qp noted, the octave up and down trimmers are labeled backwards. Not a big deal, as long as you know about it.

And yes, the vibrato depth knob is weird. I put a 10K log one in there, but that's just worse than a linear one maybe - Nothing happens until you turn it to about 3 oclock, and then it's full on at about 3:30. I'm going to experiment further with this one tonight. Edit: after further experimentation I ended up with a 100K linear pot and a 7.5K resistor between the center lug and ground. Works really well - the vibrato starts coming on at about 9 oclock, and ramps up very gradually to full mayhem. I experimented with a bunch of pots and handful of pomono test clips - note that turning both pots all the way CCW at the same time will make sparks and magic smoke happen. I mean, it could happen, but it definitely didn't happen here. Nope. Never happened.

qp wrote:

FYI for euro power: red stripe lines with the bottom of the header[/b]

Some notes on calibration:

Setting up the v/oct took a bit to get to work with me. Usually I start at a low frequency and with a guitar tuner I adjust the vco for C, then play 1 octave up and adjust the scale. For this, it was fighting me the whole time. Oddly what I found to work better was playing the 1 octave higher C, tune the VCO then play 1 oct lower C and adjust the scale. I didn't have to mess with the high freq adjustment much, but I did a bit.

I set the Tri level to match the Saw output with an oscilloscope.

For the Sine Round and Symmetry, you really need a scope to set it proper. Adjust both until you get a nice rounded sine.

The PW adjustment is P40T1-3. Be sure to adjust it to the PWIDTH pot and not the PWM pot, as I did and was cursing the lack of range. Adjust for 50% duty with pot at 12 o'clock.

The OCT DN & OCT UP screenprint are backwards. Adjust the DN for UP and the UP for DN. One thing I notice, which may be part of the design, is when I switch between octaves, there is a glide applied to the voltage, so it doesnt snap in octaves.

I didn't mess with the VCA much. With the scope I didn't notice much of difference when I adjusted the VCA trimmer. For the VCA CV I turned the drive down all the way and applied a looping envelope to the CV. There's a point where the VCA starts to bleed when adjusting. I went to that point and just backed it down a bit. It's very possible that I'm doing this part all wrong, so someone correct me if so.

Couple of things I would like to modify eventually:

The fine tune is reaaaaally fine. Like only a couple of cents. I'm assuming the 1.5M resistor is for the fine tune, but not certain. If so, I would like to drop this down to at least a 1M. (EDIT: I checked the build doc and the 1.5M is for the ratio between coarse and fine, so a lowering the ohms should spread the ratio out more. I haven't tested this yet though). [EDIT] I added a 1M in parallel (soldered to the back side) which changes the value to 600k. I tried several values and this was what seemed to work best for me. It's still a pretty tight spread but with a bit more play to keep from having to rely on the coarse tune as much.

The Variable Depth knob should be a different taper. It does nothing almost half way then it kicks in and has a very small sweet spot. I'm thinking a 10KC (reverse audio) may be better suited here. [EDIT]I followed Dave Brown's mod of adding a 10k resistor between the center lug of the pot and ground. This does help open the sweet spot up a bit more and is a better altertnative than removing the panel just to swap out the pot. I'd still suggest trying different tapers if you haven't mounted the panel yet.

I installed the TUNE pot in the normal position rather than "reversed" and the pot works High CCW and low CW, so the "normal" seems to be the reverse of what I would expect. I'll probably leave it that way as I really don't feel like taking the panel off now, but keep it in mind when doing your build.

antf4rm
Thanks sduck for the excelent report!
DrReverendSeance
Thank you qp and sduck for the very detailed reports!

sduck wrote:

The fine tune knob - that 1.5M resistor is just way wrong. [...] 75K turned out to be exactly right. Didn't have any 75K resistors, so I used 100K, and it's fine.

qp wrote:


Couple of things I would like to modify eventually:

The fine tune is reaaaaally fine. Like only a couple of cents. I'm assuming the 1.5M resistor is for the fine tune, but not certain. If so, I would like to drop this down to at least a 1M. (EDIT: I checked the build doc and the 1.5M is for the ratio between coarse and fine, so a lowering the ohms should spread the ratio out more. I haven't tested this yet though). [EDIT] I added a 1M in parallel (soldered to the back side) which changes the value to 600k. I tried several values and this was what seemed to work best for me. It's still a pretty tight spread but with a bit more play to keep from having to rely on the coarse tune as much.



I am having trouble locating the 1.5M resistor related to the fine tune knob you are both referring to. I don't see it on the current BOM at Randomsource's website or in any of the schematics from Jurgen Haible.
didahdrieghe
DrReverendSeance wrote:
Thank you qp and sduck for the very detailed reports!

sduck wrote:

The fine tune knob - that 1.5M resistor is just way wrong. [...] 75K turned out to be exactly right. Didn't have any 75K resistors, so I used 100K, and it's fine.

qp wrote:


Couple of things I would like to modify eventually:

The fine tune is reaaaaally fine. Like only a couple of cents. I'm assuming the 1.5M resistor is for the fine tune, but not certain. If so, I would like to drop this down to at least a 1M. (EDIT: I checked the build doc and the 1.5M is for the ratio between coarse and fine, so a lowering the ohms should spread the ratio out more. I haven't tested this yet though). [EDIT] I added a 1M in parallel (soldered to the back side) which changes the value to 600k. I tried several values and this was what seemed to work best for me. It's still a pretty tight spread but with a bit more play to keep from having to rely on the coarse tune as much.



I am having trouble locating the 1.5M resistor related to the fine tune knob you are both referring to. I don't see it on the current BOM at Randomsource's website or in any of the schematics from Jurgen Haible.


It's on the panel interface pcb specific to the eurorack version of the project by Randomsource, that's why it's not in the schematics.
Here's the BOM for that pcb:
http://randomsource.net/haible/lvcos/RandomSource_LVCOs_Euro.pdf
DrReverendSeance
^
Thanks didahdrieghe! That is what I was missing.
the bad producer
The PWM on the original PCB's didn't do this, at least none of my builds had this issue. I did encounter some difficulty setting this up running at 12V (specifically with the Barcode iteration of the circuit) but IIRC it was sorted in the end. I wonder if there was some discrepancy between the original PCB and the published schematic?

As Jurgen said himself:

Quote:
Even with strong pulse width modulation, you never "loose" the VCO signal, because you never get down to 0% or up to 100% pulse width.


sduck wrote:


The pulse width knob doesn't seem to work correctly, although I double checked the schematic and it seems to be identical to JH's design, so I guess it's working the way it's supposed to (I had some weird issue with the square wave circuit that I had to kind of work around, so my experience with this may be a fluke). But, if I set the trimmer to the duty is right in the middle at 12 oclock on the knob, I lose the signal completely at about 10 oclock. If I set the knob full CCW, and adjust the trimmer so I just start getting a spike (and some sound), then the actual center point for the duty cycle on the knob is about 10 oclock. Not really a big deal. I also found that generous amounts of cv will still knock the spike off the scale, so to speak - the sound will die out. I guess it's just a matter of using the PWM pot carefully.

donttouchmyrice
Any builders around here that went the MAT14 route? if so, any noticable obstacles?
I purchased one but haven't started on it yet, still waiting for som of the parts.

cheers
evengravy
Just finished up two of these inc the eurorack front panel pcb. A,lot of components to go through but smooth as silk to build. Excellent job on these nordlead! applause
HipDestroyer
How are you guys digging this module so far? I think I'm gonna build one soon if I can get my hands on the pcbs
AlanP
donttouchmyrice wrote:
Any builders around here that went the MAT14 route? if so, any noticable obstacles?
I purchased one but haven't started on it yet, still waiting for som of the parts.

cheers


I used THAT300, no noticeable problems.
PWM
HipDestroyer wrote:
How are you guys digging this module so far? I think I'm gonna build one soon if I can get my hands on the pcbs


Don't be random, buy them at the source! smile
http://randomsource.net/haible/living_vcos
HipDestroyer
PWM wrote:
HipDestroyer wrote:
How are you guys digging this module so far? I think I'm gonna build one soon if I can get my hands on the pcbs


Don't be random, buy them at the source! smile
http://randomsource.net/haible/living_vcos


Don't worry I am - just waiting to hear back from r*s about panel ocb availability :-)
Karma
Are there any alternative for the switch 2 positions? the last part that i need before soldering strips and they're available on the mid june at mouser, i can't wait any more Dead Banana
sduck
Digikey might have them. (is digikey still in europe?) Or you might try the slightly more expensive but much better built NKK ones - mouser 633-M201303 - double check the spec sheets to make sure it'll fit - I think mountain switches are a NKK knock off, so they should be fairly close, but i haven't used them in a long time.

Or, you could just leave it out for the moment, install them when you get it - the thing will still work without it, just no octave switching.
Karma
sduck wrote:
Digikey might have them. (is digikey still in europe?) Or you might try the slightly more expensive but much better built NKK ones - mouser 633-M201303 - double check the spec sheets to make sure it'll fit - I think mountain switches are a NKK knock off, so they should be fairly close, but i haven't used them in a long time.

Or, you could just leave it out for the moment, install them when you get it - the thing will still work without it, just no octave switching.


I think it's not the right one because of the size of the switch, the problem is to unmount the two pcb to solder the switch i don't want to break anything Dead Banana
HipDestroyer
Okay, I think I've got a hold of some pcbs for this beast.

I've read through the thread and noted all of the tweaks and mods suggested.

Sduck: What did you end up doing regarding the PWM?
thetwlo
Karma wrote:
Are there any alternative for the switch 2 positions? the last part that i need before soldering strips and they're available on the mid june at mouser, i can't wait any more Dead Banana


I think SmallBear in the US, but shipping probably kills it for you.
sduck
HipDestroyer wrote:


Sduck: What did you end up doing regarding the PWM?


I set the trimmer so the signal never disappears via the knob, however when set this way the center of duty is around 10 oclock. Not a big deal. When using cv the signal still disappears if it gets too high- oh well. There was some other weirdness regarding the square wave circuit, and my workaround may have caused all this, or maybe it's just a 12v-15v difference thing. I've got a euro version of this to build, maybe I'll see what's up then. May be a while though - moving right now, everything is in boxes.
HipDestroyer
thetwlo wrote:
Karma wrote:
Are there any alternative for the switch 2 positions? the last part that i need before soldering strips and they're available on the mid june at mouser, i can't wait any more Dead Banana


I think SmallBear in the US, but shipping probably kills it for you.


I ordered this instead. Can't guarantee it will fit yet but looking at the datasheets it looks like it's the same dimensions and features:

http://www.mouser.dk/Search/ProductDetail.aspx?R=108-2MS1T2B3M2QE-EVXv irtualkey12040000virtualkey108-2MS1T2B3M2QE-EVX
thetwlo
HipDestroyer wrote:

I ordered this instead. Can't guarantee it will fit yet but looking at the datasheets it looks like it's the same dimensions and features:

http://www.mouser.dk/Search/ProductDetail.aspx?R=108-2MS1T2B3M2QE-EVXv irtualkey12040000virtualkey108-2MS1T2B3M2QE-EVX


Looks like it just has a pretty short toggle.

I think this is the exact same as the Mouser one:
http://smallbear-electronics.mybigcommerce.com/spdt-on-on-mountain-10t c410/
Karma
sduck wrote:
Or, you could just leave it out for the moment, install them when you get it - the thing will still work without it, just no octave switching.


Finally it's done d'oh!

I played quickly with it, and all seems to work, the sound is just "whoooooaaaa!!!"

Now it's time to calibrate Rockin' Banana!


Edit: it's a that 300 version
HipDestroyer
I have a question regarding the transistors which is mainly rooted in my own stupidity/laziness. I acidentally bought BC560C transistors instead of BC560B - doh!

I googled it and found a post that Jurgen wrote that it usually doesn't matter too much which you use but the C has a higher hfe and judging from the datasheets it does look like there is a substantial difference in the minimum hfe of these two transistors.

I don't know enough about VCO design to actually judge whether this is one of those cases where it matters or not. Maybe I should just cave in and buy some C-transistors. Can anyone share some knowledge on the subject? Here's a clipping of the original schematic:

[/img]
HipDestroyer
I finally finished building this beautiful monster of a module. Whoever said this is a 5 hour build can fuck off – it took me quite a while. Did most of it on my balcony and got crazy sunburnt while soldering it haha!

Anyways. I used the small switches as mentioned above and they are fine, they even look kinda cute I think. I followed the tips of qp and sduck - thanks guys - and ended up using a 100k resistor on the panel pcb which works great (on 12v). Everything seems to work but calibrating this thing is a sisyphos task. Holy shit. Dead Banana



thetwlo
HipDestroyer wrote:
I finally finished building this beautiful monster of a module. Whoever said this is a 5 hour build can fuck off – it took me quite a while. Did most of it on my balcony and got crazy sunburnt while soldering it haha!
Anyways. I used the small switches as mentioned above and they are fine, they even look kinda cute I think. I followed the tips of qp and sduck - thanks guys - and ended up using a 100k resistor on the panel pcb which works great (on 12v). Everything seems to work but calibrating this thing is a sisyphos task. Holy shit. Dead Banana


Ditto!! in every aspect!
w00t just finished mine about an hour ago!

Thanks also to qp and sduck!!!

sounds fantastic, haven't tried to calibrate it--I assume that's the "5 hours"
HipDestroyer
Okay can someone help me? I'm stuck in the calibration process of the sine waves. I get a mid/high frequency clicking/noise whenever the sine wave is round, but if I calibrate it by ear to eradicate the clicking/noise the sinewave on the scope is flattened as if overdriven.

Is this normal? Here is the most perfect sine wave I could achieve with and example of how it sounds, in the scond half of the file I eq it to isolate the noise.

Here is the sound: https://soundcloud.com/gilberte666/lvco-sine-test/s-LzU4K

Here's the scope shot:
HipDestroyer
HipDestroyer wrote:
Okay can someone help me? I'm stuck in the calibration process of the sine waves. I get a mid/high frequency clicking/noise whenever the sine wave is round, but if I calibrate it by ear to eradicate the clicking/noise the sinewave on the scope is flattened as if overdriven.

Is this normal? Here is the most perfect sine wave I could achieve with and example of how it sounds, in the scond half of the file I eq it to isolate the noise.

Here is the sound: https://soundcloud.com/gilberte666/lvco-sine-test/s-LzU4K

Here's the scope shot:


And this is what the "nicest" sounding sine looks like on a scope. Does this look like an ok acceptable sine to you guys?

AlanP
Looks like a log slewed square to me... check your other waveshaping trimmers, from memory.
pix
nevermind...
cleaninglady
HipDestroyer , having a 'nice sounding' sine probably isn't the best way to calibrate, i'd say go for the one that looks best on the 'scope. The purity of Sine wave shape is more likely the important factor here.
thetwlo
still haven't tried to calibrate mine, sounds great as is--but clearly needs the Pulse tweaked.
The discrete SVVCF from Manhattan really pairs well with this as it has 3-ins.
Personally, I wish there was an LFO range switch for each VCO, other than that it's perfect!
of course, likely not useful to most.

PWM
thetwlo wrote:
I wish there was an LFO range switch for each VCO



You could run the LFO output to the FM inputs jack switches. Now you mentioned this I know I'm gonna.
pix
also finished mine yesterday, all good so far, they sound great (osc sync sounds particularly interesting)
muffdiver
Have a question about resistors RS1A, RS1B, RS1C - the BOM says these are value 2.2k but on the schematic posted it listed as one of the spots for 10k. All the photos are bit blurry. Any idea on the actual value?



sduck wrote:
Ok, don't have a cow, man. Ok ok, you can go ahead and have a cow, I just always wanted to use that phrase for some reason.

Figured it out. It puzzled me deeply, as I know I counted out 35 10K resistors and found correct places to put them all, although a quick look at that reference sheet was bamboozling me, as many of those resistors weren't on it. The trick is, I have a prototype pcb.

It seems the prototype boards used some different reference designators than the final production boards. But the BOM never got updated, at least the one currently available, to reflect that change. So, the easy way to fix this is to use the old reference sheet, which is listed first on the r*s website - http://randomsource.net/haible/lvcos/LVCOs_top_names.png - and put the resistors where it says they go. Make sense?

I only checked out the 10K's, as that's where most of the discrepancies seemed to be. Here's where they go...

glennfin
Just got my boards/panel sets last night...

Does anyone know if 2N3906's and 2N3904's can be substituted for the BC550 and BC560?
Altitude909
glennfin wrote:
Just got my boards/panel sets last night...

Does anyone know if 2N3906's and 2N3904's can be substituted for the BC550 and BC560?


different pinout afaik
HipDestroyer
Finally had the time to make some music with my Living VCO. I absolutely love this thing applause

[soundcloud url="https://api.soundcloud.com/tracks/269583779" params="auto_play=false&hide_related=false&show_comments=true&show_use r=true&show_reposts=false&visual=true" width="100%" height="450" iframe="true" /]
nordlead
HipDestroyer wrote:
Finally had the time to make some music with my Living VCO. I absolutely love this thing applause


Only came across this track now - very nice! Thanks for sharing! thumbs up thumbs up thumbs up

Actually, since I received some questions about calibration and 1V/Oct tracking, I finally got around to checking out the tracking of the Living VCO core is Spice and was quite pleasantly surprised to find out 2 things - please beware that these are findings from simulation, to be confirmed in reality -

    * the core itself seems to track very nicely with an OPA2134 (instead of the TL072 - I didn't bother to simulate the TL072) - over 8 octaves (!) - maybe further but I didn't go beyond) and

    * the HF compensation (trimpots) seems to be not needed at all, but rather detrimental to tracking - i.e. HF trimmers should be turned down all the way or replaced or bridged by a link between pin 1 and 2 (center). With that link (= without the trimmer) I had no HF deterioration in Spice up to 14.8kHz (and didn't measure beyond).
    MY ASS IS BLEEDING

So I just did a very quick test on my LVCOs (proto) board (THAT300) - put in the OPA2134, turned down HF trimpot all the way and simply adjusted the tracking trimmer and got pretty perfect results over the 5 octaves my AKAI MAX 49 delivers (top frequency tested was slightly above 2kHz).

Would love to hear if that works for you, too. Guinness ftw!

BTW: another batch of panels has been ordered and should arrive soon ...
nordlead
Based on some excellent posts and remarks is this thread I spent some time looking into the PWM / calibration aspect. As some of you noted, Jürgen's original values seem a bit odd in terms of knob range / behavior.

So here's a fresh value set to optimize Pulse / PWM control for 12V - the result of a few days in Spice and some real life testing. It looks like this should make the trimmer obsolete (simply link pins 2 and 3), so no more calibration needed. An independent validation / feedback would be very welcome!

nordlead
In case you missed the first round, the Living VCOs for Euro sets are available again! nanners

http://randomsource.net/haible/living_vcos

I've updated the panel pcb to reflect some of your comments:

- you can now pick the direction of the LIN DETUNE pots
- added a slot for a taper resistor to tweak the behavior of the PORTA DEPTH pot
- Updated docs (including a silkscreen of the main pcb with 12V values)
nurbivore
What is the input for the VCA? Is one specific waveform from each VCO being mixed together? The PCB description in the R*S store mentions an "optional mixer". Is that something that's just not exposed on this panel? I've seen other DIY versions that did have individual attenuators for each VCO, but I couldn't tell if that's something they added in custom, or even what PCB they were using.

Also, what are the two 1v/o inputs and and one 1v/o output on the left?
BasariStudios
I wonder how can i get one of these...i know its a DIY...
sduck
Sorry I didn't answer this earlier - I've been moving, and didn't have access to my modular or computer stuff for a few days. Didn't want to be making stuff up...

nurbivore wrote:
What is the input for the VCA? Is one specific waveform from each VCO being mixed together?


Depends how you wire it. If you're using the default panel and connection board, it's whatever you plug into it - nothing is normalled to it. No behind the scenes mixing is going on.

nurbivore wrote:
The PCB description in the R*S store mentions an "optional mixer". Is that something that's just not exposed on this panel? I've seen other DIY versions that did have individual attenuators for each VCO, but I couldn't tell if that's something they added in custom, or even what PCB they were using.


Yes, there's a 4 channel mixer circuit on board the main pcb, although the current euro panel (and connections pcb) doesn't use it. I'm using it with my 5U version of this - I have a switch on each VCO to select whether the square or saw output is sent to the mixer (didn't have room for a 4 way switch). Each VCOs (square or saw) output then goes through an attenuator, and then to the mixer, and then I have the output of the mixer going to a mix out jack, which is also normalled to the VCA input.

nurbivore wrote:
Also, what are the two 1v/o inputs and and one 1v/o output on the left?


Those 2 inputs are what are fed into the processor over on the left - two because 2 is better than 1 (?). The output is the result of the processor section, available for use elsewhere if desired, and also sent to the track switch on each VCO. Very handy stuff!
nurbivore
sduck wrote:
Sorry I didn't answer this earlier - I've been moving, and didn't have access to my modular or computer stuff for a few days. Didn't want to be making stuff up...


No problem, thanks for clarifying.

Still conflicted on whether to build this for eurorack with a diy mixer breakout, or 4U for a nascent boat already pretty set on oscillators. Decisions, decisions... confused
PWM
BasariStudios wrote:
I wonder how can i get one of these...i know its a DIY...


http://randomsource.net/haible/living_vcos
lfflo
HipDestroyer Did you figure out the reason of your clicking problem with the sine wave? I've got the same issue here :( To me it seems that somehow the sync pulse is leaking into the sine wave.
Any hint where to look would be highly appreciated.

Thanks in advance!
Flo
muffdiver
What's a spec on a switch that will fit for the octave switch (SPDT On-Off-On)?

I've ordered 5 different ones and every one of them is too big.


EDIT: just ordered this one:
http://www.jameco.com/z/DS-43T1B1A1QN-SPDT-Subminiature-Toggle-Switch- ON-OFF-ON-7-32-Inch-Mounting_72557.html


Hopefully this one will fit.
PWM
muffdiver, the switch you ordered is one with solder lugs and won't fit. You need a PCB version with pins. Mouser has them.
muffdiver
Thanks @PWM.

Guess I'll give it one more shot with this one:
mouser# 611-T103-001

Luckily I had another BOM waiting to be ordered - mouser shipping prices are steep.
muffdiver
It's done! By far the biggest project I've tackled so far. Everything appears to be working. Used 820k for fine tune trim. Maybe should have used a smaller value but seems to give some range. Got up early this morning to calibrate. Got a good 3 to 4 octaves in my sleepy haze. Will calibrated the PWM, Tri and Sine later.

Listening to all three oscillators harmonize together is a thing of beauty!

-



-



-



-
PWM
Great choice of knobs. Looks pretty!
muffdiver
PWM wrote:
Great choice of knobs. Looks pretty!



Thanks PWM - and thanks for the help with the switch. The one from mouser worked, just in case anyone else needs the spec.
muffdiver
Loving this thing..... It's peanut butter jelly time!

-
[s]https://soundcloud.com/ghregon3arth/06-26-2016-livingvco-02[/s]
-

-[/s]
HipDestroyer
I have this weird problem with my LVCO. All of a sudden, the middle oscillator died. I reflowed and reflowed and changed the chips around and nothing worked. Then I tried to do some exploratory poking with my dmm and I noticed that the polystyrene cap in my middle oscillator doesn't give the correct readout (says 0nf/OL) and the the working oscillators' caps have the correct value. I'm going to try and replace it and see if that is the cause of my trouble. Anyone ever heard of this before? Polystyrene caps dying like that?

Bought them from Mouser so I don't expect any fake hocuspocus here.
PWM
HipDestroyer wrote:
I have this weird problem with my LVCO. All of a sudden, the middle oscillator died. I reflowed and reflowed and changed the chips around and nothing worked. Then I tried to do some exploratory poking with my dmm and I noticed that the polystyrene cap in my middle oscillator doesn't give the correct readout (says 0nf/OL) and the the working oscillators' caps have the correct value. I'm going to try and replace it and see if that is the cause of my trouble. Anyone ever heard of this before? Polystyrene caps dying like that?

Bought them from Mouser so I don't expect any fake hocuspocus here.


They are quite sensitive to heat. Maybe it was on the verge after soldering..
You could stick another kind in to test if that cap actually is the problem. It'll work with any cap but will be less stable.
HipDestroyer
PWM wrote:
HipDestroyer wrote:
I have this weird problem with my LVCO. All of a sudden, the middle oscillator died. I reflowed and reflowed and changed the chips around and nothing worked. Then I tried to do some exploratory poking with my dmm and I noticed that the polystyrene cap in my middle oscillator doesn't give the correct readout (says 0nf/OL) and the the working oscillators' caps have the correct value. I'm going to try and replace it and see if that is the cause of my trouble. Anyone ever heard of this before? Polystyrene caps dying like that?

Bought them from Mouser so I don't expect any fake hocuspocus here.


They are quite sensitive to heat. Maybe it was on the verge after soldering..
You could stick another kind in to test if that cap actually is the problem. It'll work with any cap but will be less stable.


Thanks for the tips. I reflowed it and it seems to work for now. Hmm, but I'm keeping an eye on these styros. I wonder how much heat can fuck them up though, would a hot cabinet be able to take them out? My case can get a bit warm cause I have a lot of digital modules and stuff like that.
glennfin
Has anyone confirmed this? I'm about to place a Mouser parts order and wouldn't mind leaving off the 50k trim pots... cool

nordlead wrote:
Based on some excellent posts and remarks is this thread I spent some time looking into the PWM / calibration aspect. As some of you noted, Jürgen's original values seem a bit odd in terms of knob range / behavior.

So here's a fresh value set to optimize Pulse / PWM control for 12V - the result of a few days in Spice and some real life testing. It looks like this should make the trimmer obsolete (simply link pins 2 and 3), so no more calibration needed. An independent validation / feedback would be very welcome!

glennfin
Ok, I'm making progress... finished the component board and panel... now populating the main board starting with the caps....... ok.... there's "C3A, C3B and C3C" on the PCB but only "C3B, C3Q, C3S" on the BOM.

obviously some sort of error.... hmmm.....
peloazul
glennfin wrote:
Has anyone confirmed this? I'm about to place a Mouser parts order and wouldn't mind leaving off the 50k trim pots... cool


I've built mine like this (Nordlead's latest update you quoted) and (as far as I have used it) it does function the way I expect. I have not yet undertaken the task of tuning it.
sduck
glennfin wrote:
Ok, I'm making progress... finished the component board and panel... now populating the main board starting with the caps....... ok.... there's "C3A, C3B and C3C" on the PCB but only "C3B, C3Q, C3S" on the BOM.

obviously some sort of error.... hmmm.....



One of several discrepancies between the prototype version and the production version. Hopefully you've figured out which is which... (hint: it's pretty obvious)
glennfin
sduck wrote:
glennfin wrote:
Ok, I'm making progress... finished the component board and panel... now populating the main board starting with the caps....... ok.... there's "C3A, C3B and C3C" on the PCB but only "C3B, C3Q, C3S" on the BOM.

obviously some sort of error.... hmmm.....



One of several discrepancies between the prototype version and the production version. Hopefully you've figured out which is which... (hint: it's pretty obvious)


Thanks, yes i pretty much figured the BOM is actually referring to C3a C3b and C3c (marked on the PCB) which i assume are all 2n7 poly's.. Rockin' Banana!
antf4rm
Looks like C2A, C2B, and C2C are shifted over one pad. wink
Middle photo.






muffdiver wrote:
It's done! By far the biggest project I've tackled so far. Everything appears to be working. Used 820k for fine tune trim. Maybe should have used a smaller value but seems to give some range. Got up early this morning to calibrate. Got a good 3 to 4 octaves in my sleepy haze. Will calibrated the PWM, Tri and Sine later.

Listening to all three oscillators harmonize together is a thing of beauty!

-



-



-



-
glennfin
There's extra pads on all the non-electrolytic caps to allow the use of different size caps I imagine, three holes per cap silkscreen outline.... but yea, in that photo, C2A, C2B and C2C seem to be in the wrong position... causing the leads to be shorted together.... hmmm.....


antf4rm wrote:
Looks like C2A, C2B, and C2C are shifted over one pad. wink
Middle photo.






muffdiver wrote:
It's done! By far the biggest project I've tackled so far. Everything appears to be working. Used 820k for fine tune trim. Maybe should have used a smaller value but seems to give some range. Got up early this morning to calibrate. Got a good 3 to 4 octaves in my sleepy haze. Will calibrated the PWM, Tri and Sine later.

Listening to all three oscillators harmonize together is a thing of beauty!

-



-



-



-
muffdiver
Oh hey -- I guess I need to check this? VCO seems to work fine...hmmm.
gruebleengourd
I've been struggling for the past few days trying to figure out why I cannot adjust the pw or apply pwm on my build. Whatever the error is its in all three vcos, and they are working fine in all other respects. I built 12v euro with the adjusted 12v values in the bom. I thought if found the issue when I read the message from Nordlead above about using a 22k resistor and removing the trimpot but that has seemingly made no difference. I get a narrow pulse wave and neither the pw knob nor pwm knob on the panel makes any difference...


Edited later:

Ok I see where I went wrong. I built according to the parentheses values in the BOM which didn't include the corrections for the pwm circuit. When I saw the fix in this thread I focused on the 22k and the jumper without noticing that other values were also changed... Which of course make sense and I even wondering how just changing the one resistor and shorting the timpot would solve the issue... Of course there are more resistors to change.
Unfortunately I don't have 4m3 or 12k resistors on hand (how lame) so this will have to wait until another night.

Note to self read the entire build thread carefully before starting.

Still, it would be good if the main board BOM was updated with the 12v mods for the pwm. I see that the corrected values are in the panel build doc in the board diagram, but that was the last thing I looked at.
glennfin
....and yet more errors on the BOM... missing "R17A" ......... is there anyone working from a corrected bom?? I'd like to address these errors while I'm stuffing and soldering parts, not after the fact.. meh ...... and "R49" ??? what is that supposed to be? (edit: there is no "R49" on the PCB but it's on the BOM...) EDIT again; Looks like you have to follow this instead of the BOM...

http://randomsource.net/haible/lvcos/Living_VCOs_2016_12V_values.png


sduck wrote:
glennfin wrote:
Ok, I'm making progress... finished the component board and panel... now populating the main board starting with the caps....... ok.... there's "C3A, C3B and C3C" on the PCB but only "C3B, C3Q, C3S" on the BOM.

obviously some sort of error.... hmmm.....



One of several discrepancies between the prototype version and the production version. Hopefully you've figured out which is which... (hint: it's pretty obvious)
antf4rm
Finally finished this thing. It wasnt to terrible to tune up. It scales pretty well and it stays in tune. I can tune it to a chord and transpose it as expected. I have normaled the three sin wave outputs from the switch side of the jacks to Mix 1,2,3. Then the mix out goes to the switch jack of the vca input. This way i can make sin wav chords with less patching and without the need for another mixer. I also missed the changed values in the pwm section minus the 22k like gruebleengourd. I'll have to go back and fix that someday or live with slightly limited pwm.

Only one issue. Noticed i have a little spike in the top of my sine and triangle waves. Any suggestions for how to get rid of it?

It seems to appear in all of the wave forms. A micro spike. Same results on all three oscillators.









gruebleengourd
antf4rm wrote:
. I also missed the changed values in the pwm section minus the 22k like gruebleengourd. I'll have to go back and fix that someday or live with slightly limited pwm.

Only one issue. Noticed i have a little spike in the top of my sine and triangle waves. Any suggestions for how to get rid of it?


When I first finished the build and had PWM section resistors as per the red parenthesis in the BOM my waveshapes were perfect.

After I changed the trimpot and 22k resistor, I too got those spikes. So while it seems pretty far from the core/waveshaping, perhaps an imbalance in the PWM section is causing that spike. I'm waiting on the 4M3 resistors so I can finish up correctly, hopefully the spikes disappear then.

The other thing I did before realizing that I had to replace so many resistors for +12V PWM was swap out Q7 just because who knows why, but it didn't seem to have any influence on the spikes.

If putting in the right resistors doesn't help, I'll try beefing up the power rail caps, but I doubt that's the issue. I expect the PWM section needs to be properly balanced.
glennfin
gruebleengourd wrote:
I've been struggling for the past few days trying to figure out why I cannot adjust the pw or apply pwm on my build. Whatever the error is its in all three vcos, and they are working fine in all other respects. I built 12v euro with the adjusted 12v values in the bom. I thought if found the issue when I read the message from Nordlead above about using a 22k resistor and removing the trimpot but that has seemingly made no difference. I get a narrow pulse wave and neither the pw knob nor pwm knob on the panel makes any difference...


Edited later:

Ok I see where I went wrong. I built according to the parentheses values in the BOM which didn't include the corrections for the pwm circuit. When I saw the fix in this thread I focused on the 22k and the jumper without noticing that other values were also changed... Which of course make sense and I even wondering how just changing the one resistor and shorting the timpot would solve the issue... Of course there are more resistors to change.
Unfortunately I don't have 4m3 or 12k resistors on hand (how lame) so this will have to wait until another night.

Note to self read the entire build thread carefully before starting.

Still, it would be good if the main board BOM was updated with the 12v mods for the pwm. I see that the corrected values are in the panel build doc in the board diagram, but that was the last thing I looked at.


What "other values" have changed for the PWM section for 12v operation?... The BOM is such a mess... trying to figure this out.... disappointing that R*S haven't provided an accurate BOM.
gruebleengourd
If you look at the diagram on the last page of the euro adaptor build guide the corrections are in the circuit diagram, but not in the materials list. The area of importance is in that graphic posted above with the 22k comment and pink jumper, almost every resistor that is in the graphic has changed from what's in the vco board Bom. Maybe 1 or 2 are the same. You have to rebuild according to the diagram.
glennfin
Ok, the answer to this should end my confusion...

Are ALL the components on this "suggested values for 12v operation" silkscreen correct?

Here = http://randomsource.net/haible/lvcos/Living_VCOs_2016_12V_values.png

and this bom is incorrect? =
http://randomsource.net/haible/lvcos/LivingVCOs_BOM.pdf

I've already made the front panel board changes that were suggested above, and on the main board, changed R39A, B, and C to 820k with a wire link at R39 x, y and z...
I'm hoping everything else that is supposed to be there is on that "suggested values" diagram... if I know that's all correct at least I can follow that verses that screwy BOM confused

gruebleengourd wrote:
If you look at the diagram on the last page of the euro adaptor build guide the corrections are in the circuit diagram, but not in the materials list. The area of importance is in that graphic posted above with the 22k comment and pink jumper, almost every resistor that is in the graphic has changed from what's in the vco board Bom. Maybe 1 or 2 are the same. You have to rebuild according to the diagram.
gruebleengourd
Other than the pwm for 12v I don't think the bom was that bad. Yes there are typos but it's not to hard to figure out that the resistor values for each position were the same for each oscillator.

I built according to the red parentheses in vco board bom and everything but the pwm worked fine. I'm still waiting on some 4M3 resistors so that I can correct the pwm section, but I'll report back in a day or two unless someone else can chime in.

If you're preparing to order parts get all the red parentheses components skip the 50k trimpot and order all the resistors in pwm correction diagram.


glennfin wrote:
Ok, the answer to this should end my confusion...

Are ALL the components on this "suggested values for 12v operation" silkscreen correct?

Here = http://randomsource.net/haible/lvcos/Living_VCOs_2016_12V_values.png

and this bom is incorrect? =
http://randomsource.net/haible/lvcos/LivingVCOs_BOM.pdf

I've already made the front panel board changes that were suggested above, and on the main board, changed R39A, B, and C to 820k with a wire link at R39 x, y and z...
I'm hoping everything else that is supposed to be there is on that "suggested values" diagram... if I know that's all correct at least I can follow that verses that screwy BOM confused

gruebleengourd wrote:
If you look at the diagram on the last page of the euro adaptor build guide the corrections are in the circuit diagram, but not in the materials list. The area of importance is in that graphic posted above with the 22k comment and pink jumper, almost every resistor that is in the graphic has changed from what's in the vco board Bom. Maybe 1 or 2 are the same. You have to rebuild according to the diagram.
PWM
glennfin wrote:
Ok, the answer to this should end my confusion...

Are ALL the components on this "suggested values for 12v operation" silkscreen correct?

Here = http://randomsource.net/haible/lvcos/Living_VCOs_2016_12V_values.png


Mine is build exactly as this. The PW works but I'm still not very happy about the response of the PW pot, first half does nothing..

Also, there seems to be a problem with my driver section. My octave switch doesn't reach an octave, up or down.
Could nordlead post the schematics of this section?
glennfin
Thank you PWM... that's how I will proceed. We'll see what happens with the PW circuit on my build when I'm finished. The saving grace with issues on this build is that the main board can be separated from the front panel board and component changes/mods can be made on the bench......vs having to remove pot and jack nuts to gain access to components..... and yes, schematics would be most appreciated. Mr. Green


PWM wrote:
glennfin wrote:
Ok, the answer to this should end my confusion...

Are ALL the components on this "suggested values for 12v operation" silkscreen correct?

Here = http://randomsource.net/haible/lvcos/Living_VCOs_2016_12V_values.png


Mine is build exactly as this. The PW works but I'm still not very happy about the response of the PW pot, first half does nothing..

Also, there seems to be a problem with my driver section. My octave switch doesn't reach an octave, up or down.
Could nordlead post the schematics of this section?
glennfin
Hey grueb... yea, the bom IS pretty bad... like where is "R49" on the PCB?... I understand the common component thing between the 3 VCO's but I'm trying to understand why it's a problem to provide an accurate BOM, with the PCB's/panels that are being sold as a DIY Eurorack module.... that reflects use as a "EURORACK MODULE" ??. yea I can probably figure it out, but this back and forth between this diagram, that diagram, previous versions, ........ just give us the correct BOM (and schematic) that goes with the Eurorack PCB and Panel you're selling!!!

................. unless this is all still in development in which case forget everything I just said.. Dead Banana

Am I asking too much??? very frustrating

gruebleengourd wrote:
Other than the pwm for 12v I don't think the bom was that bad. Yes there are typos but it's not to hard to figure out that the resistor values for each position were the same for each oscillator.

I built according to the red parentheses in vco board bom and everything but the pwm worked fine. I'm still waiting on some 4M3 resistors so that I can correct the pwm section, but I'll report back in a day or two unless someone else can chime in.

If you're preparing to order parts get all the red parentheses components skip the 50k trimpot and order all the resistors in pwm correction diagram.


glennfin wrote:
Ok, the answer to this should end my confusion...

Are ALL the components on this "suggested values for 12v operation" silkscreen correct?

Here = http://randomsource.net/haible/lvcos/Living_VCOs_2016_12V_values.png

and this bom is incorrect? =
http://randomsource.net/haible/lvcos/LivingVCOs_BOM.pdf

I've already made the front panel board changes that were suggested above, and on the main board, changed R39A, B, and C to 820k with a wire link at R39 x, y and z...
I'm hoping everything else that is supposed to be there is on that "suggested values" diagram... if I know that's all correct at least I can follow that verses that screwy BOM confused

gruebleengourd wrote:
If you look at the diagram on the last page of the euro adaptor build guide the corrections are in the circuit diagram, but not in the materials list. The area of importance is in that graphic posted above with the 22k comment and pink jumper, almost every resistor that is in the graphic has changed from what's in the vco board Bom. Maybe 1 or 2 are the same. You have to rebuild according to the diagram.
gruebleengourd
glennfin wrote:

................. unless this is all still in development in which case forget everything I just said.. Dead Banana

Am I asking too much??? very frustrating



No, I actually fully agree, this one could be a lot better. It does need a revision. I just don't want to say anything too bad as I've built all the random source euro serge modules and they are fantastic... But maybe those had a second pair of eyes to catch the errors.

This is still synth diy and the community plays a role in catching mistakes and errors, so I don't sweat it too much that were are the ones contributing to making it a better document. Hopefully it gets updated soon for those building down the line.

I've done quite a few modules. But I've never felt like I needed to read the entire build thread at wigglers to do it correctly, and then some...

I'd setup the oscillator circuit on a bread board and mess with the values, myself to find the best pw solution, but I don't have the matched transistors... And for spice I've only got the free linear version. Which is only really convenient if you build exclusively with their parts. Is there another decent free spice implementation out there with gui?
glennfin
All good. I'm working from the 12v silkscreen image. Let us know if you find out anything about that PW issue... cool

gruebleengourd wrote:
glennfin wrote:

................. unless this is all still in development in which case forget everything I just said.. Dead Banana

Am I asking too much??? very frustrating



No, I actually fully agree, this one could be a lot better. It does need a revision. I just don't want to say anything too bad as I've built all the random source euro serge modules and they are fantastic... But maybe those had a second pair of eyes to catch the errors.

This is still synth diy and the community plays a role in catching mistakes and errors, so I don't sweat it too much that were are the ones contributing to making it a better document. Hopefully it gets updated soon for those building down the line.

I've done quite a few modules. But I've never felt like I needed to read the entire build thread at wigglers to do it correctly, and then some...

I'd setup the oscillator circuit on a bread board and mess with the values, myself to find the best pw solution, but I don't have the matched transistors... And for spice I've only got the free linear version. Which is only really convenient if you build exclusively with their parts. Is there another decent free spice implementation out there with gui?
gruebleengourd
So my resistors arrived today and I'm sey according to the 12v board diagram. Spent the time scaling and calibrating and now everything is great, and these oscillators certainly live up to their name.

I did notice occasional spikes at the peaks of the sine and triangle, but it's not constant. I was also having some trouble getting good contact between the headers on the two boards. Maybe it's related I stopped scoping it after I got the sines calibrated.

At first the pulse width range on its own seems odd, but when you start adding cv for pwm it helps for finding the right range with the overmodulation.

The vca is a nice addition. I do get some slewing when switching octaves. I havent touched any of the trimmers on the vca/controller portion. I'm not certain it's giving full octaves, I'll look into that later.

I'm not too keen on the vibrato depth range, but that's mostly my taste, anything beyond something subtle annoys me.

Super cool sound from these!
PWM
gruebleengourd wrote:
I'm not too keen on the vibrato depth range, but that's mostly my taste, anything beyond something subtle annoys me.


Did you ad the 10k resistor from the wiper of 'Depth' to 0V?
It does improve the response somewhat.
gruebleengourd
PWM wrote:
gruebleengourd wrote:
I'm not too keen on the vibrato depth range, but that's mostly my taste, anything beyond something subtle annoys me.


Did you ad the 10k resistor from the wiper of 'Depth' to 0V?
It does improve the response somewhat.


Yes, but maybe that's not the right value for me. I would prefer much finer resolution on the subtle end of the scale. For my taste the useful range is less than 5 degrees.

Overall I am very happy with this design. It is a bit unusual what the vca drive does, I guess it's offsetting the vca cv in? But it is a cool effect to have an envelope going in to the vca cv and have regular response at 0 and then go to more compressed dynamics and volume offset, and overdrive as you turn one knob. I kinda wish there was a patch point for it.
glennfin
She's done!!




Haven't calibrated yet... all three osc's working.. except the fine tune controls are WAY too fine.... .. I'm pretty sure during the build I changed three resistor values per someone's suggestion to help correct this but apparently it isn't helping. seriously, i just don't get it [/img]
Karma
You can decrease the value of Rfine for a better range, 1M5 seems to be a bit to high, try between 500 and 800
sduck
1M5 is way too high. 500K is better but still way too fine for my tastes. I ended up using 100K as per this post - https://www.muffwiggler.com/forum/viewtopic.php?p=2182407#2182407
cyphersum
What was used for R33A,B,C? Silkscreen shows 22k*... BOM shows 20k (33k)

Thanks!
cane creek
I'm suprised there isn't a thread about the Phaser that was released a few days ago.
glennfin
I used a 68k for the fine tune controls and it's still way too fine....... I'll try a 47k next...


sduck wrote:
1M5 is way too high. 500K is better but still way too fine for my tastes. I ended up using 100K as per this post - https://www.muffwiggler.com/forum/viewtopic.php?p=2182407#2182407
glennfin
I used 22k. I started out trying to use the BOM but quickly discovered many errors, didn't match the PCB silkscreen, etc.. I ended up completing the build just using the suggested 12v values silkscreen image from the R*S website.


cyphersum wrote:
What was used for R33A,B,C? Silkscreen shows 22k*... BOM shows 20k (33k)

Thanks!
peloazul
glennfin wrote:
I ended up completing the build just using the suggested 12v values silkscreen image from the R*S website.


I want to second this (for emphasis) as the build choices that worked for me as well.
cyphersum
thumbs up

peloazul wrote:
glennfin wrote:
I ended up completing the build just using the suggested 12v values silkscreen image from the R*S website.


I want to second this (for emphasis) as the build choices that worked for me as well.
cyphersum
Can a 1k PTC Thermistor (mouser 803-102PS1G) be used for the Tempco?
cygmu
cyphersum wrote:
Can a 1k PTC Thermistor (mouser 803-102PS1G) be used for the Tempco?


The temperature coefficient of those parts is way too high I think -- something like 8000ppm/K (1kOhm at 25C, 1040Ohm at 30C according to the data sheet.)
maxbobo
Hey, This was my first diy build. I think it was an overambitious first project because of the discrepancies between the BOM and the 12v value sheet - Confusing for a beginner. In the end I followed the 12v sheet, and the PWM updates, as closely as possible. This seemed to work well.

Build went smooth enough. Its almost completely functional .. but I have some debugging to do on the third OSC.

It works - all wave outputs work, as well as direct controls for PWM, V/Oct, Sync, and FM.. but it does not seem to be responding to the driver section (Octave, main v/oct ins, port, vibrato, etc).The 'Track' switch seems to not engage the driver section even though Im getting continuity between the switch and edge connector in the driver section.

I reflowed solder to all the edge connectors and the switch as well. When that failed I replaced the switch. No luck. Can anyone point me in the right direction as far as troubleshooting, obvious failures, etc.. the original schematic is hard to follow as all the locations are different of course.

Still have to calibrate and fix the Fine resistor values too.

PS - Just finished the Tau phaser. Easy to build and it sounds great…
maxbobo
Got it. meh

Bad solder joint on one of the pins... Everything's working now. Onto the calibration.
cyphersum
cygmu wrote:
cyphersum wrote:
Can a 1k PTC Thermistor (mouser 803-102PS1G) be used for the Tempco?


The temperature coefficient of those parts is way too high I think -- something like 8000ppm/K (1kOhm at 25C, 1040Ohm at 30C according to the data sheet.)


Thanks Cygmu... sourced the 3500ppm ones.


Another question I had on this build is regarding the 50k R40 trimpots... I see the updated build doc shows the first two trimpots using a link between pin 3 and 2 but not the third trimpot. Are we suppose to link it as well? omit? or use trimpot?

Thanks everyone, as always!
PWM
As I understand it, you can omit the PW trimpots for the 12V version and instal the link.
peloazul
Fwiw, I built with the 12V notes/parts, including the link, and the duty cycle of the pulse wave was not 50% at ~12 o'clock on the pulse width pot. It was, however, consistently at 50% at the same location on all three VCOs (between one and two o'clock, for me).

Also, goodness, I spent at least three hours vexing over the tuning of it. I think you could go down that rabbit hole indefinitely.
nordlead
Sorry guys, the last months have been crazy and I haven't managed to follow this thread closely... now that I did I just discovered that I had updated the BOM a while ago, but somehow the new file didn't make it onto the server very frustrating - very sorry about this!

I've just uploaded an updated version of the BOM where I tried to make the PWM adjustments and the tracking / HF suggestions as clear as possible.
I've also added the PWM pic from this thread and some more info:
http://randomsource.net/haible/living_vcos

Hope this clears any confusion.

In general, please shoot me an email if you detect any obvious or not so obvious errors (rather than posting here hoping that I'll eventually come across that post).

Oh, and in case you did send me an email previously in which you pointed out any errors in the BOM (which got lost or overlooked on my end), I'm happy to send you a free original Haible FS1A or String Filter pcb set or a Dual Wasp for Euro kit which has just been released ... (just forward me your mail again).

Merry Christmas!
muffdiver
Fixed those misplaced capacitors. Didn't make much of a difference sound wise - but I guess it's better than having some shorted part of the circuit and it gave me a chance to change the fine tune resistor value (1M5) on the panel pcb from 750k to 91k. Much better range now!
didahdrieghe
I'm confused about where the link should go for the HF compensation trimmers. BOM says between pin 1 and 2 but the picture shows it between pin 2 and 3. Which is correct?
nordlead
didahdrieghe wrote:
I'm confused about where the link should go for the HF compensation trimmers. BOM says between pin 1 and 2 but the picture shows it between pin 2 and 3. Which is correct?


Pins 2 and 3 of the HF trimmers are linked anyway (you should be able to see that on the bottom layer), so linking pin 1 with either 2 or 3 should do the trick, i.e. the BOM is correct. Which picture are you referring to?

In general, if you think you spotted a mistake in the docs, please send me an email (rather than post here) as I may not see a post at all.
KvotheKingkiller
Just put in my Mouser Order (thanks @Altitude909 for the mouser BOM!) got everything except some styrene caps (2700pf). Can't find an equivalent anywhere. : (

So close. Any Suggestions? Thnx
paterursus
Small Bear Electronics carries them.
KvotheKingkiller
paterursus wrote:
Small Bear Electronics carries them.


Thanks! Checked there first actually.

Maybe I'm just confused about my farads conversions?

Small bear lists the value as mF which is supposed to mean millifarads, unless they acutally mean ųF but are using "m" for "micro"? as 2700pf would be 0.0000027 mF

seriously, i just don't get it

sduck
mF is often used instead of uF, as on that page. So the one you have highlighted is correct.
KvotheKingkiller
sduck wrote:
mF is often used instead of uF, as on that page. So the one you have highlighted is correct.


Thanks for the sanity check! Guess I was just being dumb.

It's just a bit weird to me tho. Its not like this is my first rodeo, Ive built 3 synths, a drum machine, half a dozen pedals, guitar preamps, and a monome, and I haven't really come across mF = uF and milliFarads ARE a thing. But I guess I only really source parts from mouser or digikey and have only had to convert between nF pF and uF.
Altitude909
I think its for no other reason than its not a "u-F" part its, a"μ-F" (as in the greek letter) and that will cause all sorts of problems for different browsers, font sets etc so mF is used. A millifard is 1000 uF and I've never seen a capacitor marked or specd like that, usually they go right to F for big ones
Lemmy
KvotheKingkiller wrote:
Just put in my Mouser Order (thanks @Altitude909 for the mouser BOM!) got everything except some styrene caps (2700pf). Can't find an equivalent anywhere. : (

So close. Any Suggestions? Thnx


I'm having the same problem, but I'm in the UK. Can anyone suggest a UK supplier for this part? Or - what are the characteristics of a polystyrene cap, could I substitute a different type?
templar
eu supplier:

http://www.reichelt.com/gb/en/Styroflex-Capacitors/STYROFLEX-2-7N/3/in dex.html?ACTION=3&LA=2&ARTICLE=19847&GROUPID=3158&artnr=STYROFLEX+2%2C 7N&SEARCH=%252A
Lemmy
Ah, thanks templar!
Lemmy
I have another question for those who know - thonk is out of thonkiconn jacks, which are the ones specified for this project. Can I just use kobiconns from Mouser instead, or should I wait for thonk to restock?
KvotheKingkiller
Altitude909 wrote:
I think its for no other reason than its not a "u-F" part its, a"μ-F" (as in the greek letter) and that will cause all sorts of problems for different browsers, font sets etc so mF is used. A millifard is 1000 uF and I've never seen a capacitor marked or specd like that, usually they go right to F for big ones


Yea I get that its "micro" not really a "u". It was more that I was used to seeing it just being typed as "u" since most don't know or don't want to bother with the odd micro character, versus "m" for micro. It just seemed weird to be the only place I'd seen to do that.

All sorted now, thnx! Next challenge was that the resistors listed in the mouser BOM seem to be a bunch of mismatches with what the designators say. Either that or mouser sent me a lot of wrong parts and quantities (doubtful). I did a cursory check between mouser BOM and R*S docs, but I should have been more thorough.

Also there are a few values that are completely unreadable on the pdf's because the black text is obscured by the silkscreen, but hopefully can deduce from whats left over.

Gonna make a google spreadsheet that maps part ref number to value and share here, hopefully it will be helpful for someone else. Switching between the two images is too error prone for my tastes.
didahdrieghe
Remember reading somewhere that people were normalling their oscillator outputs to the vca, any tips on that? Can't figure out what to connect where on the board.
soup
didahdrieghe wrote:
Remember reading somewhere that people were normalling their oscillator outputs to the vca, any tips on that? Can't figure out what to connect where on the board.


your choice to mix1 - mix4, mix-out to vca-in
didahdrieghe
Woops yeah that makes sense. So it's okay to just connect 4 outputs to the 1 input?
sduck
No - you'll want to go through the mixer first
nurbivore
The Linear Detune is supposed to be really fine, right? I can't actually hear any effect on a single oscillator by itself, just the change in beating when combining two.

Also, I think this is going to be what finally pushes me over the edge into getting an oscilloscope. The sine output has a weird harshness to it, and I don't quite trust this oscilloscope app on my phone.
sduck
Yes, linear detune is almost uselessly fine. Considering how nonlinear the response is on these vcos, it seems pretty useless.

The sine out is pretty good on mine - better than quite a few other vcos I have.
Leverkusen
I just finished populating the PCB and wondered if it would make sense to test it with an oscilloscope before I start the wiring (no panel PCB because of doing a 5U version).

I am not sure if this is supposed to work anyway but when I connect the PCB to power the +15 rail of the PSU goes down to nearly 0 V and the current up to 4, which is somewhere near to what it can deliver. I did not find any shorts by examining the soldering and would now check the components.

But before I do this I just wanted to make sure that I am right with my assumption that the unconnected board is somehow equivalent to having all pots fully CCW and no plug connected? And that I should be able to see waveforms on the scope? Or at least no power irregularities?

hmmm.....
sduck
With my 5U version, I didn't power it up until I had connected all of the first VCO. As it mostly worked, I didn't test the power usage.
Leverkusen
I think I got it working now - this call to link two pins might have misleaded me somehow:

nordlead wrote:
didahdrieghe wrote:
I'm confused about where the link should go for the HF compensation trimmers. BOM says between pin 1 and 2 but the picture shows it between pin 2 and 3. Which is correct?


Pins 2 and 3 of the HF trimmers are linked anyway (you should be able to see that on the bottom layer), so linking pin 1 with either 2 or 3 should do the trick, i.e. the BOM is correct. Which picture are you referring to?

In general, if you think you spotted a mistake in the docs, please send me an email (rather than post here) as I may not see a post at all.


It felt a bit strange to connect all three pins and since it was the only short I could find I just cut the link between 1 and 2 and then I got the basic waveforms on all six outputs. Maybe this is new in PCB v.2017? There is a link screenprinted on the PCB that connects pins 2 and 3 TRPC just besides it. I think it wasn't there on the 2015 PCB's.

By the way - how did you do the wiring? I am using the Synthcube frontpanel w/36 jacks and 26 pots on 4 MOTM spaces and it seems as it will be the more serious kind of fun. Also the spacing on the PCB is to narrow for the MTA-headers I have here - or rather the headers do fit but not the housings then... eek!
sduck
What kind of header/connectors are you using? The ones I used fit. I think. I may have had to do some exacto knife shaving to get things all in there.

How I did the wiring -

Jurgen Haible - Random*Source Living VCOs back by Stephen Drake, on Flickr
Leverkusen
Wow, looks beautiful! I like the many colors... love

I am using molex headers like this one:



I can fit 2 or three side by side when I squeeze them but then there is not enough room for another one. Now I started to solder the wire directly to the board...

I wonder why there are so many 0V/GND pads on the PCB since all the jack rings are connected is it really mandatory to connect all single GND pads?
sduck
For those mta header/connectors I used mouser parts

571-640454-3
571-640441-3
571-640550-3

(change the last number to 2 for the 2 pin variety)

It's possible the board has changed, which would cause problems for these. Nothing wrong with direct wiring to the board, I've done that many times, you just want to make sure stuff works before committing, as it's a major chore to desolder (I've also had to do this). Or be very confident in your skills.

As far as grounds to all those jacks, I always connect the wires also, it's good practice (for me at least). While the aluminum panel does conduct electricity, it's not a very good conductor, and one shouldn't rely on it for a good path to ground. And if the bridechamber panel is painted, that's going to degrade its conductivity also.
soup
I recently finished building one and didn't have any problem with fitting standard te connectivity mta headers/housings without modification. The 3rd osc sync in is a little tight but works fine once it's in place. I built a banana frac one so I didn't connect any of the grounds and it works great...

Leverkusen
sduck wrote:
For those mta header/connectors I used mouser parts

571-640454-3
571-640441-3
571-640550-3

(change the last number to 2 for the 2 pin variety)

It's possible the board has changed, which would cause problems for these. Nothing wrong with direct wiring to the board, I've done that many times, you just want to make sure stuff works before committing, as it's a major chore to desolder (I've also had to do this). Or be very confident in your skills.

As far as grounds to all those jacks, I always connect the wires also, it's good practice (for me at least). While the aluminum panel does conduct electricity, it's not a very good conductor, and one shouldn't rely on it for a good path to ground. And if the bridechamber panel is painted, that's going to degrade its conductivity also.


I think that are the headers I use - the problem is more that the housings do not fit next to each other cause they are a bit wider then the headers.

Direct wiring just started to be fun (and somehow fast) when I noticed that I I would not get the PCB screwed to the panel anymore having knot the wiring together in a too messy way. So I had to desolder everything and start again with more attention...d'oh!

I just connected all jack ring connectors to get a good ground and hope that will work. Maybe connecting the ground pads is a good idea since I would not have to think about which to use and which not when I just connect all of them. Looks like this now:



Regarding the pre wiring testing I am still not sure if it is supposed to work w/o wiring the pots and the jack normalizations. I get something out of every oscillator and when it was connected it also reacts to the frequency knobs but not to anything from the driver. Measuring at DRVOUT gives some varying voltages and the octave switch does nothing.

Also the sine and the triangle have a big down spike at their peaks. But I am happy that it seems to partly work. I might have to trouble shoot more before I re connect everything.

@sduck: what is that resistor you put on the Osc 1 pulse output?
soup
the driver is not hard patched to the oscillators, you have to connect it to one of the CV ins.
Leverkusen
soup wrote:
the driver is not hard patched to the oscillators, you have to connect it to one of the CV ins.


Yes, I had the CV section of the driver (left part of the original Hablewiring digrams) done w/o effect - but probably have it done wrong somewhere. Also I am not sure about measuring DRVOUT with a meter.
sduck
Leverkusen wrote:

@sduck: what is that resistor you put on the Osc 1 pulse output?


Not 100% sure I remember correctly. I think it's to fix a problem with the prototype board I was using - the square out is routed through an opamp, differently that Jurgen's original design, and it didn't work, so I kludged a fix that replicated the original design. At least I think that's what thats for.
Leverkusen
I have finished the wiring and most parts seem to be running except that I cannot get any square wave out of it. It is there at the collector of Q6 (15V pp, though and with an offset of -1.4V) but it disappears on the other side of R60 (EDIT: I meant R30, of course). It is the same o all three oscillators and comparing to Jürgens original schematics there seems to be a lot happening between that point and the output, which I don't see/understand - starting with R31 not going to -15V but to zero on the PCB.

Does anyone have an idea what to look after? Or are there schematics that do fit the actual PCB (v2017)?

I am not sure if this is right but contrary to the BOM and this thread I did not connect link 1 and 2 of TRPC but linked JP beside it (which is linking 2 and 3 instead of 1 and 2). It was the only way to get the oscillators running when I finished the PCB. Removing the link made the square on Q6 worse so I guess this is the way to do it? Nevertheless none of the links gave me a square at the output...

seriously, i just don't get it

sduck
Arggh. That's the same problem I had. I thought he fixed that. I ended up running a wire from one end of R30 (you say R60 but probably mean R30) to the output. I'd have to pull mine out to really tell you 100% what worked - I think there may have been something else to the fix. I seem to recall some kind of opamp that needed to be bypassed, that wasn't in the schematic at all. There's something weird going on, as supposedly the euro version built with the same board works fine.
Leverkusen
sduck wrote:
Arggh. That's the same problem I had. I thought he fixed that. I ended up running a wire from one end of R30 (you say R60 but probably mean R30) to the output. I'd have to pull mine out to really tell you 100% what worked - I think there may have been something else to the fix. I seem to recall some kind of opamp that needed to be bypassed, that wasn't in the schematic at all. There's something weird going on, as supposedly the euro version built with the same board works fine.


Oh. Thank you for sharing this - at least it is relieving to learn that it might not be my fault. I would not understand how the design might not work at all as it seems that is MOTM with some modifications to Euro and I thought the difficulties lay in the documentation of those conversion. Maybe he lost track somewhere in the conversion process and does not know anymore what is original MOTM and what had changed due to the conversion. Did you experience only the missing pulse wave or the offsets too?

I wonder if I have made a wrong ground connection somewhere? For example I did not really understand how to wire up the sync input as it has no ground pad at all.

confused
Laughing
If the signal completely disappears, then yes, I imagine you're getting your signal shorted to ground somewhere. If the output of the system should be 10Vp-p, then you do need to keep in mind what Jurgen's written on the page.

Quote:
The waveform from the 3 VCOs are roughly +/-1.2V to be compatible with Moog systems. The board also contains 3 Amplifiers to boost this up to the MOTM (and others) standard level of +/-5V.


The amp circuit is really just a standard opamp circuit, IIRC.

I got a legacy board which only has pulse and saw outputs, soldered it together, plugged everything up, and it worked perfectly on the first go, all rather recently, too. I hope there's nothing wrong with the R*S board...
soup
I bought what I think was the latest production pcb and followed the random source BOM and didn't have to do any modifications, it took me a few trys to figure out all of the wiring but other than that it worked fine.

You can just use a different 0V for the sync jack, they are all the same.
Leverkusen
Thanks everyone for sharing your experiences - reassured that it should work in a MOTM environment I finally got it running and solved all my issues. It's peanut butter jelly time!

I had to go through all PCB layouts/versions and the original/r*s mod schematics in parallel since I got a little confused with the high frequency and the pulse width trimmer modifications.

For the vibrato width I went with C100k pot and a 7k5 resistor and I am very happy with the transition starting from around 9 o'clock with a nice transition to fully CW.

There are still some things I do not understand though:

There are two pads besides the SYNC mark on the PCB, what are they for? I only connected the one one the right, just besides the marking. Before I had the other on connected to 0V which shorted the pulse wave.

How do you use all the different pads supplying the oscillator waves - two pulses and 4 saws? One saw is not amplified but apart from this I found no differences in them.

I still have an offset voltage between .5 and 1.5 V on the outputs and the waves range between 8V pp and 12 V pp.

The main tuning pot has a quite limited range, just about half an octave in both directions, which is not much of an issue withe the range switch and the massive range of the single oscillators but still seems a bit strange, kind of a main fine tune.

The VCA trimmer has no effect on anything, has it?

I like how the Drive knob has a small spot where the patched in CV reacts as if it was inverted.

The only real issue is that I still have the occasional spike, especially in the sine wave. Anyone found a solution for this?

Apart from that it's a great machine! My first build that contained heavy wiring. I opted for a somehow skiff friendly assembly because it sits in front of the PSU and the space is a little narrow there but think I got the appeal of the MOTM bracket style now...









(Still feels a bit euroracky...hihi)

hihi
rvlt
Hi Leverkusen,

I made the same mistake as you did with the sync input: shorting the square wave.. wink took me a while to figure it out.

I almost finished my LVCO, it still has a prototype panel but apart from that it’s done. I spent a whole weekend ‚optimizing‘ it, here’s what I did:

- the sine (and partly the triangle) outputs where a bit too hot compared to the pulse/saw outputs, and the sine wave had too many overtones / noise for my taste, sounding almost like a triangle. I lowered the input of the sine shaper (if I remember correctly), you have to make a compromise between perfect sine shape and introducing more noise. It involves cutting a trace an the pcb and running a jumper cable, but it was worth it. All outputs have more or less the same amplitude now and the sine sounds more like a sine.

- after I corrected the sync wiring I got a ‚kind of‘ sync sound, but I was never really satisfied with it, especially when comparing it to other sync sounds (from a moog e.g.). Using a scope I could see that on some freq. settings it was syncing and on some/most others it was not, not really sounding like a proper hard sync. I read posts on the electro-music forum in the LVCO thread about adding a sync input, tried two different approaches of adding sync but was never 100% satisfied with it, so in the end I decided to ditch them completely.

(btw if you find the time: could you check the sync feature on a scope? I’m really interested if others have the same problem)

- didn’t notice any offset voltage

- yes, some outputs appear two or three times on the pcb, a bit confusing at first.

- if I remember correctly, yes, there was a trimmer in the VCA section that did almost nothing.

The best feature of this VCO is the Drive knob, it’s the best sounding overdrive I have, especially when you use it in a subtle way, just adding some fatness/saturation.
Leverkusen
rvlt wrote:
Hi Leverkusen,

I made the same mistake as you did with the sync input: shorting the square wave.. wink took me a while to figure it out.

I almost finished my LVCO, it still has a prototype panel but apart from that it’s done. I spent a whole weekend ‚optimizing‘ it, here’s what I did:

- the sine (and partly the triangle) outputs where a bit too hot compared to the pulse/saw outputs, and the sine wave had too many overtones / noise for my taste, sounding almost like a triangle. I lowered the input of the sine shaper (if I remember correctly), you have to make a compromise between perfect sine shape and introducing more noise. It involves cutting a trace an the pcb and running a jumper cable, but it was worth it. All outputs have more or less the same amplitude now and the sine sounds more like a sine.

- after I corrected the sync wiring I got a ‚kind of‘ sync sound, but I was never really satisfied with it, especially when comparing it to other sync sounds (from a moog e.g.). Using a scope I could see that on some freq. settings it was syncing and on some/most others it was not, not really sounding like a proper hard sync. I read posts on the electro-music forum in the LVCO thread about adding a sync input, tried two different approaches of adding sync but was never 100% satisfied with it, so in the end I decided to ditch them completely.

(btw if you find the time: could you check the sync feature on a scope? I’m really interested if others have the same problem)

- didn’t notice any offset voltage

- yes, some outputs appear two or three times on the pcb, a bit confusing at first.

- if I remember correctly, yes, there was a trimmer in the VCA section that did almost nothing.

The best feature of this VCO is the Drive knob, it’s the best sounding overdrive I have, especially when you use it in a subtle way, just adding some fatness/saturation.


Thanks for sharing your findings, rvlt!

It took me some time to get back to it, this is what I found on mine:

SYNC: I get some something that sounds like sync in someway but also a bit like FM. It's not the typical sync sweep sound you would expect but near and playable. Very rich in the harmonic content but less formants. When you apply a low frequency saw to the sync input you definetely get a frequency sweep as you would get from a linear FM input. On the Oscilloscope you see waveforms restarting but it seems not to be the only thing happening there.

SINE: I am not so much concerned about the shape of my sine. It does not seem to be the most perfect sine shape but what begs me more is the occasional spike at the top of the wave making it buzzy (same with the triangle). If your workaround does fix this I would be keen to learn about it and see if it is doable for me. If there is another solution to remove the spike I would be very happy...

HOTNESS/OFFSET: All single outputs go up to 12V pp here and the individual saw outputs are positive, not centered around zero.

There are still some things I could work on so I might try the sine mod too. The v/oct calibration had accidentally led me to mismatching FREQ pots, one PWM CV control is wired backwards and I don't like how all thee of them only use one third of their range/way before maxing out. But I have to procrastinate this until some other oscillator projects are finished as in the current rebuild process of my rack it's the only one that is actually powered and usable.
Lemmy
apoisontree wrote:
How do you solder the tempcos with the BCM847DS ?


Should we be using thermal paste or epoxy, or just sitting the tempcos on top of the BCM847DSs?
PWM
Lemmy wrote:
apoisontree wrote:
How do you solder the tempcos with the BCM847DS ?


Should we be using thermal paste or epoxy, or just sitting the tempcos on top of the BCM847DSs?



You should if you care about V/Oct tracking.
Lemmy
OK, thanks. I'm getting close to the end of the build for this. A couple more questions...

I'm using the most recent board I think, for 12v operation, and have been using this as a guide: http://randomsource.net/haible/lvcos/Haible_LVCOs_2017_values.png

Now I'm wondering about UD4 (IC on the far right) which says OPA2604? in the diagram. I have a spare OPA2134, which is how this IC was marked in the BOM. Would that do the same job?

Also, in my board there are two holes marked JP next to each 50K trimmer. Having read this thread I'm not sure whether to add the trimmers and put a link (JumPer?) in the space next to them, or leave out the trimmers and add the link directly between trimmer points 2 and 3.

Thanks for any suggestions. The board seems really well designed but the documentation needs some overhauling and properly separated out for the different versions eek!
Leverkusen
Lemmy wrote:

Also, in my board there are two holes marked JP next to each 50K trimmer. Having read this thread I'm not sure whether to add the trimmers and put a link (JumPer?) in the space next to them, or leave out the trimmers and add the link directly between trimmer points 2 and 3.



I do not exactly remember all details but I found that part of the documentation unclear and contradictory. The way I worked it out did not work and after a second look at it was shorting to ground somewhere - I then just installed the trimpots...
Lemmy
Leverkusen wrote:


I do not exactly remember all details but I found that part of the documentation unclear and contradictory. The way I worked it out did not work and after a second look at it was shorting to ground somewhere - I then just installed the trimpots...


OK, it will be trial and error then, once I have my final parts to install. I already installed the trimpots, so I will try with and without the jumpers.

Regarding the op amps, it looks from my research that I should be able to use the OPA2134. Though I suppose they must have some different characteristics....?
didahdrieghe
Moderator edit - you'll need to post it in the Buy Sell Trade forum.
clorax hurd
Lemmy wrote:

Regarding the op amps, it looks from my research that I should be able to use the OPA2134. Though I suppose they must have some different characteristics....?


They seem very similar to me. Would be cool if the documentation included info about the all the substitutions compared to original BOM, so we would know what really matters (like the tuning opa2134) and what is less important. I used opa2134 instead of that more expenive one and seems to work ok.

I have to say that I am generally disappointed by the quality of this project.

The documentation lack a lot. Things different from original are not explained properly. You need to dig info from forums and two sets of documentation and put the pieces together + double check against the PCB, instead of getting one true set of documentation containing all the important information. Also schematic - you get link to original web with schematics + few new additional schematics from R*S, but not much info about how those pieces are connected and where exactly. And while most parts have the same names (original vs r*s), some does not. Quite confusing.

Also not much info about the product. No details about how the VCA is supposed to work. I thought that a mix of the three oscillators would be normalled into the VCA-IN jack. Would feel natural, but seems isn't not the case.
Or what is the the amplifier section / which problem of the original does it solve? I only know it's there because it's listed among features and there is schematic for it.

The pcbs also not so great - the thermal decoupling between ground pads and ground planes is not very good. Pads are also quite small. Small annular rings and small holes. Both things made unsoldering bit difficult. And there was lot of unsoldering, because:
1) Some of the specified values were not chosen sensibly. I had to add one resistor and fiddle with 3 others to tame the vibrato depth pot. This was hell, because half of those resistors had one pin connected to ground...And I had to try more than 2 values for each of them, before it was kind of OK.

2) For most pots it's not mentioned (in docu) what is "reverse" and what is "normal". It is mentioned only for the "linear detune" pot. For rest of the pots you have to just guess if "reverse" means opposite of how it was on original (and you don't know how that was) or if it is opposite of how things commonly are on most other modules. If you guesses are wrong (like mine was), then more unsoldering pots... and as some of the pots have one pin connected to ground then once again... I had to throw some pots (and hours of my life) into thrash while building this.

And then there are small things like polarity of the keyed power IDC connector is other way around. But yes, shit sometimes happens... but it would be better if it was not mentioned only on the webshop page, but also on the 12V PCB documentation file with values (which is what most people look at while building).

Also the Linear Detune range is bit small. I would suggest chaging the R16ABC from 10K to 15K to make the linear detune range bit bigger.
(I understand that the Linear Detune supposed to be subtle, just to allow beating phenomenon, but with 15K the resolution is still very fine for tuning it, while it also allows also bit faster beating between oscillators.)

Regarding the vibrato depth, I Used a 100K pot to save power and IIRC
I adjusted R150 and R149 to around 3x (one should lower how much current gets from LFO into the OTA, the other should lower the control current going into OTA. (both should make the maximum depth lower).
Adjusting the pot's parallel resistor even to very low values was still giving me depth pot doing "nothing" for big part of the range, and then doing too much. So I also lowered the R157 to 5K6 or 6K2 which made the "nothing" (or too shallow to be audible) part in the begining much smaller.
I was perhaps bit too cautious, but wanted to make sure, that it will really start with doing nothing. (having a pot with switch there would be fancy. when it clicks to OFF then no modulation and when it's ON it could start with some tiny modulation depth and whole range of the pot could be used)...

nordlead wrote:
Hey guys,
Lemmy
My Living VCOs are Dead VCOs

Any help with troubleshooting appreciated, as I am a relative novice with circuit analysis.

I get faint white noise hiss from all triangle and sine outputs, but nothing at all from pulse and saw.

I can hear the vibrato working on the hissy output.

I've reflowed everything that looks like it needs it. Where would be a good place to start checking? I have an oscilloscope so I can check specific points in the circuit.

It's quite possible that I have missed something simple, as I found the documentation confusing in places. I have had the 50K trimmers installed from the start, and now have added the jumpers next to them. Same problem.
Lemmy
Oh. I had the three BC560Cs on the right the wrong way round. Turned them round and still have the same problem... Maybe I killed something.
Lemmy
Well I had quite a few transistors the wrong way round. Moral of the story - just match the emitter position and ignore the other helpful markings on the board. (I misinterpreted the semi-circle on the board linking C to E. Thought that showed the shape of the transistor body.)

Anyway, now I have sound, which is great. But my pulse wave isn't right, and PWM isn't doing anything.



This seems similar to what others have mentioned, but I followed the most recent schematic: http://randomsource.net/haible/lvcos/Haible_LVCOs_2017_main.png, which matches my board, and the updated BOM for 12V values. I also tried with resistor values from http://randomsource.net/haible/lvcos/Living_VCOs_PWM.png (330R and 12K instead of 300R and 10K), and I have tried with the jumper next to the 50K trimmer.

Is it possible I could have damaged something when I had my transistors installed in reverse, e.g. Q6A and Q7A?
Lemmy
I wonder if anyone would be kind enough to help me troubleshoot this.

The pulse wave, as shown in my last post above, is not right.

I had my BC550C and BC560C transistors the wrong way round, but I have now installed replacements the right way round throughout the board.

I also replaced the bottom SMT BCM847DS on the back of the board, in one of the VCOs. The pulse output remains the same in spite of the changes. (On all three VCOs.)

Ralf at random source tells me that the pulse section is a transistor circuit - here's the schematic from the original Haible circuit: http://jhaible.com/legacy/living_vcos/living_vcos_sch_1of4.pdf

The oscilloscope image in the post above was from the point marked in red here:



Here's the back of the board - I've marked the replaced BCM847DS:



If anyone has suggestions they would be greatly appreciated - or if anyone has a working Living VCO board it would be useful to compare oscilloscope output to try to narrow this down. I don't want to replace every component, but it would be a shame not to get this thing working properly!

I'm presuming that the driver section isn't relevant?

It's quite possible that I have done something stupid which may be obvious to a more experienced eye

Thanks!
the bad producer
Hi, is it just the pulse which is weird? ie saw etc is all good?
Lemmy
Hi, yes the rest are OK, though triangle and sine have very nasty spikes in them. I'll add some more pics later this evening.
the bad producer
Cool, if it's all working apart from the Pulse it should be fairly easy (easier!) to troubleshoot, you've replaced all the transistors around the pulse circuits here:

http://jhaible.com/legacy/living_vcos/living_vcos_sch_1of4.pdf

And you've checked soldering in that area? And the diodes are in the correct orientation and the resistors are the correct values? Have you measured the end of R29 and can see +vb1, measured the end of R28 and can see -vb1 etc? I mean checked the voltages as marked on schematic are indeed correct? Another thing I find with these RS boards is they require a bit more heat than 'normal' PCB's, so might be good to check your soldering, especially on GND pins as these wick the most heat away and and give bad connections. Have you the exact same problem on all 3 VCO's?

I have an original here, I just put it on the scope to check my waveforms and they are indeed square, so it's not a 'feature' of this VCO (you never know!)

Nasty spikes? Shouldn't really be there, odd...
Leverkusen
Lemmy wrote:
Hi, yes the rest are OK, though triangle and sine have very nasty spikes in them. I'll add some more pics later this evening.


I have those nasty spikes too. If you find a way to get rid of them pleas share it.

Regarding the pulse - I had problems with it too and then finally found that I had wired the sync wrong, i.e. I shorted it to ground. After desoldering one of the two pads, that seem to be labeled sync the pulse was right. I see that you have headers on both of them, this might be right cause you probably built the euro version w/ panel pcb but I would check this anyway.

EDIT: Also it is every pulse, right? This might speak against bad soldering cause it has to be something you did wrong in all circuits in the same way.
the bad producer
Does the SAW have this spike? If not it may be to do with the waveshapers, which were added by RS from other designers... Are there trimmers for this? I designed/built my own so have not this problem, sorry!
the bad producer
Leverkusen wrote:

EDIT: Also it is every pulse, right? This might speak against bad soldering cause it has to be something you did wrong in all circuits in the same way.


I thought the same, made me wonder if wrong resistor or diode wrong way round across all 3...
Lemmy
Yes, all three are exactly the same, so I think I've either wired something wrong in all three, or all three sustained the same damage when I had the transistors the wrong way round...?

I replaced all the non-SMT transistors, but only the one BCM847DS which if I remember is a replacement for Q4 and Q5 in the Haible schematic. Perhaps I should try replacing the other SMT transistors.

I'll put up some pics of the other waveforms.
Lemmy
Leverkusen wrote:


Regarding the pulse - I had problems with it too and then finally found that I had wired the sync wrong, i.e. I shorted it to ground. After desoldering one of the two pads, that seem to be labeled sync the pulse was right. I see that you have headers on both of them, this might be right cause you probably built the euro version w/ panel pcb but I would check this anyway..


Yep, mine is Euro version but I'll take a look at that.
the bad producer
What is 'JP' for? I was looking at the build docs and a bit confused!
Lemmy
OK, these are taken from cable output from the front panel, so after the driver section, I guess. The previous pic of the pulse wave was direct from the component board.

Saw:


Pulse:


Triangle:


Sine:
Lemmy
the bad producer wrote:
What is 'JP' for? I was looking at the build docs and a bit confused!


Hm, yes it's unclear, but I think it's jumper. It bypasses the HF trimmer. It's a Random Source mod for improving tracking. I tried mine with the jumper in place and without, but it didn't fix my issue.
the bad producer
I've just found my PCB for this, it's the 2015 version and it doesn't have the 'JP' thing. Comparing the 2 it looks like it's connected to R37, ie in the pulse circuit, not the HF trim part. Anyway, it's not you problem I think!
the bad producer
Have you got a better pic of one VCO section - also what have you put at R32A, B, C?
the bad producer
I can't really see from your pic, and the different V's of the PCB are slightly confusing me, but have you swapped R30 and R32? 1M vs 6K2...
Lemmy
That seems to be OK - I have 1M on the left (R30) and 6.2K on the right (R32). I'll upload a close-up pic.
the bad producer
Should be:
R32 = 1M on right
R30 = 6K2 on left
Lemmy
Ah! You're a genius! My print out is of the 2016 board with the resistors reversed. d'oh!
Though I did match the resistors in the lower part of the board to the 2017 update for the HF trimmer bypass.

I will do some desoldering and see what happens.
Lemmy
Well it's looking a lot more like a square wave now, albeit with some interesting peaks on one side. I am guessing there may be more changes between the 2016 and 2017 layouts which I missed. I'll check and report back here when I've gone through them.

Thanks for the help much appreciated! we're not worthy Mr. Green
the bad producer
thumbs up Hopefully that's it, I guess the JP thing is to do with this:

http://randomsource.net/haible/lvcos/Living_VCOs_PWM.png
Leverkusen
Wow, great that you could solve this! applause

Now I wonder if we can find out what all of us who get those spikes in their waveforms have done wrong which all others have done right? hmmm.....
Lemmy
Yes, I wonder if anyone with the newer board with the triangle and sin addons doesn't have the spikes?

If so it would be useful to compare setups.
Lemmy
In the meantime this is turning into a frustrating learning experience for me... very frustrating

I connected my component board to my front panel and powered up. I then realised that my power supply (Doepfer A-100) wasn't working. The external power supply is dead.

So I found another one, 12V AC instead of 9V AC, but when I plug it in the two LEDs for +12 and -12 on the A-100 light up. So it seems to be working OK.

But, when I power up the Living VCOs - the driver section is fine but when I connect the VCOs the +12 LED on the power supply doesn't light. And at the power connector on my board for the VCO section, I only get 12V at the top two contacts, on my meter, where I was getting 24V before.

So something on the Living VCOs board is now broken - what? And was I just unlucky with a breaking power supply, or was it because I connected the component panel to the front panel? I've done it numerous times before without problems.

Aargh. Dead Banana
sduck
No spikes or weirdness with my pulse waves - although I do have very thin spikes on the tri and sine waves.

Lemmy
That's a thing of beauty love

Which version of the board did you use?
sduck
The very first r*s version, one of the prototypes actually. Pictures of it a few pages back.
Lemmy
For anyone else building this module - my power problem was due to installing both the PWM trimmer AND the jumper next to it on each VCO (2017 board).

I thought the jumper would bypass the trimmer, but it seems that you should install one or the other, but not both.
the bad producer
Didn't spot that as not on mine, but is it all cool now?
Lemmy
Yep, all looking and sounding good. Just calibrated it. Some minor issues I might come back to later - I still have spikes in triangle and sin, and sin levels are slightly lower than the others. But I'm so pleased to have this thing up and running I'm not so bothered... Sounds great anyway. w00t
ebardie
This on a 2017 PCB:
ebardie
Zoomed in at a higher frequency:
ebardie
Appears to be a problem in the triangle shaper. If I raise the triangle level, the glitch moves up out of the sine wave. No idea which bit of the triangle waveshaper circuit to mess with though... Help anyone?

[edit 1:]
Looking at http://electro-music.com/forum/phpbb-files/thumbs/t_new_saw_to_tri_192 .jpg I wonder if is worth experimenting with replacing RT6[ABC] with a 20k trimmer. So I do that and...

... put the power cable on backwards.

Arse!

Now none of the oscillators do anything at all! eek!

Who wants to play guess-the-blown-components?

[edit 2:]
And the answer is: "just" the OPA2134s.

Oh, and RT6[ABC] controls the triangle symmetry, but doesn't affect the spike.
303ish
Hey guys! I finished my build and calibrated it today, everything was fine except for the pulse wave, which was a thin pulse at best (trimpot and pw knob both full clockwise)

So I swapped R30 and R32 and changed R34 to 10k to match the updated values from R*S page. Now my pulse looks like a slewed square (pic attached), it sounds ok but either the trimpot, pw knob and pwm cv have no effect on the width…

Can someone help me figure out whats wrong? My pcb version is 2016.

clorax hurd
sduck wrote:
The VCA drive knob is weird, and may be something that works differently in a 12v system - I haven't experimented with it yet. What I'm experiencing is that you can use that part of the circuit as a VCA, or a drive circuit, but not both. It works well as a basic VCA, but once you start dialing in the drive, the cv you're applying stops working as a vca, and turns into more of a modulation sort of thing. Not how I was thinking it was going to work, but still fun and useful.


I was just checking this too and it seems that that's exactly what it should be doing.

found words from Jurgen about what the circuit should be doing:

- With "Drive" at minimum, it's a simple, noninverting VCA
- Without a CV, with different "Drive" settings, it's an inverter or inverting Clipper/Overdrive
- Combination of VCA and Inverter: CV reduces gain at certain settings of "Drive"

here http://www.jhaible.info/tonline_stuff/hjvcaod.gif

By the way, anybody has some calibration method for this?
The only sources about this VCA I found:
- On R*S page there is a schematic of it in the addons PDF
- on Jurgen's site you can find it (don't look for "vca/distortion/overdrive" - that's probably a typo, but vca/inverter/overdrive.

but neither of those two sources contains calibration info.

I was (and still am) disappointed that the panel PCB doesn't have mixed sines or mixed triangles normalled into the VCA IN.
(or even better, using a 2 position switch, it could allow both. it would definitely fit on the driver part of panel the same way as "track" switch on VCO parts of panel.)
That way, the VCA would make more sense as a part of the whole module. Now the module feels confusing. There is just 1 channel of VCA for 3 channels of VCOs and no way to mix it.

I think i seen somebody mention, that a mixer circuit is already on the board anyway, just unused?
If that's true, then I could and should hack that normalled connection. Fortunately, it seems that VCA IN normalling pin is not connected to ground or anything else, so it should be possible using few wires...
sduck
Yes, there's a mixer on the pcb. Have you seen my panel for this (pictured a few pages back)? There's a switch which sends either the square or saw wave from each vco to pots leading into the mixer - the mixer out is normalled to the vca in.

edit: so I looked for it, and it looks like I never actually posted my panel here. It looks almost identical to the synthcube version, which is modeled after it:

Jurgen Haible - Random*Source Living VCOs front by Stephen Drake, on Flickr
Starspawn
Lemmy wrote:
Ah! You're a genius! My print out is of the 2016 board with the resistors reversed. d'oh!
Though I did match the resistors in the lower part of the board to the 2017 update for the HF trimmer bypass.

I will do some desoldering and see what happens.


Im not so sure about this. I have the 2016 version, and although Ive yet to do the panel wiring, I can clearly see that the righmost resistor is R30 going straight from the transistor, and R32 is the leftmost.
This is also correct from the updated designator and values from 2015 PCB (2016 mark was just 2nd run of the 2015 ones as far as I can tell.)

So 6.2K right and 1M left should be right going from Jurgens Schematic, but your non 10K R34 might have been the culprit if youre running on 12V
303ish
Thanks man, I swapped R30 and R32 again and got a perfect square! So in the end I just changed R34 to 10K and kept the trimmers (they are adjusted fully ccw tho), everything else is matching the 2016 instructions.

PW knob and PWM CV now go full range but never fade away, exactly as it should be.

Everything calibrated and running, just the oct up/down trimmers seems to do nothing...
Starspawn
Thats good smile
R140 was also changed and is relevant to the octave switch, that part isnt complicated, the 100K trimmers is just to finetune the octave jumps.
If main tune works its just a few parts to check.
If they do nothing Id think a change of R140 to 150K should help.

Edit: Another clarification for anyone soon to power up theirs. BOTH power connectors have negative bottom and positive top. I see theres been some confusion about that since the picture is of the leftmost.
prune
HI, I just got my living vco's setup, I have a few problems

My PWM section isn't working - on all three of the VCOs. Ive been trying to debug it, but the original schematic on Jurgen Haible's site doesn't seem to match the board from R*S - for example, the pulse output (I have found through digging around on the PCB) comes from an op amp, not straight off the diodes. The schematic also doesn't match the designators on my board

it's quite hard to debug when neither the schematic is the same nor the designators. Does anyone have an up-to-date schematic? Did I miss something? I feel like there should be another piece of information here somewhere

EDIT: Just read the thread and changed the resistors 30 and 32 and have a good square wave. However I am also trying to resolve the spikes..
Starspawn
While Im here.
Just finished wiring up mine and all works as it should.
However the vibrato LFO is waay to slow for my taste, has anyone experimented with changing the 47nf capacitor?
sduck
Is your Vib speed knob not working? What range are you getting from it?
Starspawn
Sloooow to some vibrato, nothing like what I would consider fast, not even 200ms Id think.

I do however have my own panel, and my rate pot is B10K like the original schematics. I dont know what designators the pots are in RS frontpanel as I dont have it, but as its not mentioned in any change schematic I can only assume its the same.
sduck
Mine goes from 1.75 hz to 10.6 hz, which seems like a good musical range to me.
DabiDabDab
Just built mine. Lin detune control doesnt change the pitch at all, at all, not a cent, on any of the 3 oscillators.. any idea what i could check? Everything else works fine so it shouldnt be so hard to isolate the issue, i wager!? The original schematics is hardly of any help here sad

Thanx for the input!
CliffordMilk
I’m halfway through my Euro version build. My main board has 2015 printed on it.

Reading through this thread, I’m now a little unsure as to the correct orientation of the transistors. Do I follow the PCB design and line up the curved markings with the curved surface of the component or is there another indicator? If there is, I can’t see it seriously, i just don't get it
Lemmy
I don't know about the 2015 board, but on mine there was an E next to where the emitter pin should be. The curve shape on my PCB was misleading and I installed the transistors the wrong way at first because of that.
2015 board is probably different though - do you have a pic?
CliffordMilk
Thanks - I’ve just seen the ‘E’ is marked on the PDF of the board with the component layout. It looks slightly different to mine but close enough that it should be good.
pix
I got asked by another wiggler if I had solved my issue with with vibrato and it was because I forgot to solder one of the pins of the IC.
CliffordMilk
Just finished my build and all seems good apart from the Drive pot doesn’t seem to affect the sound at all. It may be a stupid question but does a signal need to be patched into the VCA for this to have an effect?

If not, I have some debugging to do. Any suggestions?
nurbivore
CliffordMilk wrote:
Just finished my build and all seems good apart from the Drive pot doesn’t seem to affect the sound at all. It may be a stupid question but does a signal need to be patched into the VCA for this to have an effect?


Yeah, the drive knob only affects the signal coming out of the VCA.
schwittersplayspop
Here's my almost but not really working build using the 2017 PCB:




Right now on all three oscillators I've got fairly unsatisfactory waveforms. The pulse and saw waves are about 0.25V pp, while the triangle and sine are a bit better, at 3V pp. The pulse and saw waveforms look roughly OK, but the triangle and sine look more like square waves than triangle/sine. They're both almost the same waveform, though to the ear, the sine is slightly softer.

(Although I think I had the Q7 NPNs the wrong way around initially, changing their orientation didn't change the sound. The picture shows them after the change of orientation.)

I've played with the 2x triangle and 2x sine trimmers but haven't found them capable of getting me away from this basically square-ish waveform.
schwittersplayspop
Here are the outputs of one oscillator (the other two are the same as this on all four outputs: very weak saw and pulse, stronger triangle and sine that look square-ish).

If anyone has ideas about what could be shared across all three oscillators causing these problems I'd be grateful. Thanks!

Sine:



Triangle:



Saw:



Pulse:

Leverkusen
I find your soldering and the way the components are placed looking a bit casual und would check for components that might accidentally touch each other with their legs first, then go over the joints again.

It sounds as if it is the same fault on all three oscillators so chances are you have one or more components placed wrongly on all three while populating. So you might check every component.

When I built mine I used the schematics to look for points where I could grab the signal with a scope to narrow down the parts of the circuit I had to check more sorrowly.

I had some problems with leaving out the trimpot you bridged too, but cannot really remember. It got better though when I installed the trimmer.

Also I don't think that I put a wire link to what looks like R397.
schwittersplayspop
Thanks for your reply, Leverkusen. My soldering technique needs some improvement, though I can say I've managed to build dozens of modules successfully. Although certainly sometimes a bit of reflow has been necessary!

Given the systematic issue across all three oscillators, I agree with your prediction that one or more components is wrongly placed. Or perhaps just as likely in this case, misreading the current BOM PDF has led to me using an incorrect component.

I went for 12v values as described in the BOM, and used the Mouser cart posted early in this thread. That required some supplementary ordering as certain values seem to be for older versions of the board.

I also followed the instructions for PWM here:

http://randomsource.net/haible/lvcos/Living_VCOs_PWM.png

Which is why I have a link across R39X, R39Y and R39Z.

The trimpot bridging is currently recommended as well.

I'll begin the per-component checks.
CliffordMilk
Could someone please tell me which way round R30 and R32 go on a 2015 board?

Everything is working well but I’m getting micro spikes on all waveforms on all three oscillators.

Perhaps someone else has used boards marked 2015 and solved this another way?

Any help would be much appreciated.
Starspawn
schwittersplayspop wrote:


Given the systematic issue across all three oscillators, I agree with your prediction that one or more components is wrongly placed. Or perhaps just as likely in this case, misreading the current BOM PDF has led to me using an incorrect component.


I think it is more likely that one of the ICs are bad and used in all sections, did you source any of the opamps or ICs on Ebay?
schwittersplayspop
Starspawn wrote:
schwittersplayspop wrote:


Given the systematic issue across all three oscillators, I agree with your prediction that one or more components is wrongly placed. Or perhaps just as likely in this case, misreading the current BOM PDF has led to me using an incorrect component.


I think it is more likely that one of the ICs are bad and used in all sections, did you source any of the opamps or ICs on Ebay?


Everything is from Mouser, and was installed into a pre-soldered socket.

I am not using as many OPA2134s as is currently suggested, though. I have some TL072s because of the older cart I used. In the oscillator sections I have OP2134 in IC1A, IC1B, and IC1C and TL072 in IC7A, IC7B, and iC7C. But am I right that OP2134 and TL072 are equivalent enough to not be the issue here?

Still, I am happy to follow up any line of thinking at this point. I'm curious why you think the opamps and ICs would be the most likely issue? Thanks.
mcbinc
I haven't built a living vco yet (have built other JH designs) but... assembly is an issue. Trim the leads - from the shadows, it looks like some are half as long as the header pins. Then clean the flux and post some higher resolution and better lit photos here. I'd swear there was a short at the top of the middle VCO on the component side and a lot of possibles elsewhere.
schwittersplayspop
Current work in progress pics posted. No appreciable change in the waveforms after a lot of re-flow work. I haven't got anything around to clean flux, but I'm still going to post pics as I feel I've done a further round of work here with several sessions of re-flow and lead trimming (I didn't find any shorts in that process).

I still have the dubious re-placed Q7 NPNs (you can see them looming up tall from the board), which I had reversed at first (before I made my first post). I realize this is dodgy but I don't want to re-work that part of the board too much more before being sure it is the likely issue, as I'm worried about damaging it.

Given the consistency of the problem across all three oscs I'm still inclined to think it's bad chips or a bad/wrong component placed consistently.





I'll get something to clean the PCB in places if I need to post pics again. But in the meantime I'll probably order some more chips to test, and possibly also try to do a better job on the Q7 NPNs.
woodster
Have you checked for continuity between the top and bottom pcb's via the headers, to make sure the pin header soldering is OK ?

I know that some of the pcb interconnect pads for some Random Source PCB's can be very hard to get decent solder wetting without using additional flux, and bumping up the soldering iron temperature a bit.
schwittersplayspop
Thanks woodster, I did check this again on your suggestion. I get the expected continuity everywhere.
CliffordMilk
Isn't there supposed to be a jumper on all three of those trimmers at the bottom in the pulsewidth mod section?

EDIT : Maybe not - I have an older board with out the additional 'JP' holes.
peloazul
Maybe JP in the 2017 version is the same as jumping pads 2,3 of the trimmers next to it? I built the 2016 version myself.

I don't see anything super obvious.
- keyed power connectors are correct.
- ICs are oriented correctly
- diodes oriented correctly
- transistors oriented correctly (supposing they are the type specified in the bom)

You could always try checking the power rails at the opamps just to see nothing weird is happening before they get there? Maybe try replacing the TL072 with OPA2134?

In any case, good luck!
schwittersplayspop
CliffordMilk and peloazul -- thanks, yeah, the newer board seems to promote using JP only.

I'll keep at it and try your suggestion about the opamps.

I haven't tried 2x OPA2134 in a single osc -- maybe worth a shot!
babysealclub
I built two of these, one had spikes. I licked my finger and poked around. I found a spot that cleared up the spikes. I want to say it was in a wave shaper section. I added a 10nf or 100nf cap to ground and it cleared up. I don't have my notes with me and it was about a year ago, but I think that is what I did. I don't have the modules anymore. If anyone is still having spike issues let me know and I'll see if I can find my notes.
Leverkusen
schwittersplayspop wrote:
Current work in progress pics posted. No appreciable change in the waveforms after a lot of re-flow work. I haven't got anything around to clean flux, but I'm still going to post pics as I feel I've done a further round of work here with several sessions of re-flow and lead trimming (I didn't find any shorts in that process).

I still have the dubious re-placed Q7 NPNs (you can see them looming up tall from the board), which I had reversed at first (before I made my first post). I realize this is dodgy but I don't want to re-work that part of the board too much more before being sure it is the likely issue, as I'm worried about damaging it.

Given the consistency of the problem across all three oscs I'm still inclined to think it's bad chips or a bad/wrong component placed consistently.

I'll get something to clean the PCB in places if I need to post pics again. But in the meantime I'll probably order some more chips to test, and possibly also try to do a better job on the Q7 NPNs.


Have you then checked all the resistors? I was quite embarrassed when I finally brought my tripple wave folder to life after of course replacing all ICs and transistors cause it was easier to do so - I accidentally picked 440R instead of 440K on every position...it's just so cumbersome to recheck resistors on a board. d'oh!
pix
in Q7 / Q9 are there any advantages to the SOT23 SMT transistors?
peloazul
Look pretty similar to me:
http://www.onsemi.com/pub/Collateral/BC550C-D.PDF
http://www.onsemi.com/pub/Collateral/BC846AWT1-D.PDF
pix
just finished mine (2017 version). All is working well but I have spike in the sines and PW only goes from 20-65% range even with all the mods and with 22k on R33. Would like to have a bit more range...any suggestions of some values I can try?

Thanks!
pix
btw I used 2N3904/06 instead of the BC series I wonder if that's why I'm getting different PW results.

edit: just installed the 50k trimmer and it seems that adjusting decreases the range of the PW...and I'd like to make it go the other way...
Lemmy
babysealclub wrote:
I built two of these, one had spikes. I licked my finger and poked around. I found a spot that cleared up the spikes. I want to say it was in a wave shaper section. I added a 10nf or 100nf cap to ground and it cleared up. I don't have my notes with me and it was about a year ago, but I think that is what I did. I don't have the modules anymore. If anyone is still having spike issues let me know and I'll see if I can find my notes.


If you can share your notes about this that would be great!
pix
still tweaking the PW... very frustrating

I replaced R33 with a 50k pot and have been playing with different values and it seems that decreasing the value of R33 results in a nice increase in overall PW range but it only changes the left (positive) side of the waveform.

With very low R values I end up with a nice thin pulse with PW pot fully CCW, but when going CW the range ends up in the same spot as with 22k (around 60% of width).

Now if I increase the value of R33 close to ~45k I get the waveform to go all the way to the negative side and can end up with a very thin pulse, but then the PW knob doesn't have any effect on the waveform when going CCW.

So what other resistors should I tweak to get a full sweep with the PW pot? R34?
pix
OK after testing a lot of combinations here's a set I'm happy with. It sweeps PW from 10-95% in a decent pot range.

If you can test it and find ways to improve it please let me know. These are the changes I made compared to the 2017 suggested values.

R33 - 42k
R36 - 10M
R37 - 1.5M
R39 - 1M
Jop
Thanks for the PW resistor report thumbs up I still have my RS boards unpopulated, so good information for the future.

I did build Logicgates version in 2016 and used that time the resistor values from Benjamin AM, I assume these changes will also work on the RS boards?

Benjamin AM wrote:
Sorry to bring this up so late in the game. I found that the Barcode BOM v1.3 had some values that did not work my build. I feel I should share my experience in case anyone out there has had the same issue. In regards to the PWM section of the circuit, resistors R39 and R36 were troublesome for me. I came to this conclusion because I felt that the PWM wasn't working as advertised because the pulse wave was cutting out even before full CW and CCW. This circuit is not supposed to be like this. The diodes are supposed to protect the pulse from completely shutting off. This means to me that the gain and bias on the transistor at the summing stage of the PWM circuit is to high. So all that I did to change this was return R36's value to the suggested value by Haible(750k opposed to the Barcode BOM 1.5M). Then the PWM mixer resistor(R39) back to the original value 1M, instead of Barcode BOM value(510k). I also found it nice to return one of the biasing resistors(R33) back to original value 20k. Now all is good for me.
R33=20k
R39=1M
R36=750k
pix
Jop wrote:

I did build Logicgates version in 2016 and used that time the resistor values from Benjamin AM, I assume these changes will also work on the RS boards?



Hmm, not sure, the PW on this build was very different than what is described in the build documents. I checked all resistor values and made sure they were all correct, and all 3 osc behaved the same. So it's either something to do with the new boards or maybe due to a different batch of components (I used 2N390X transistors too not sure if that's what caused the difference).

So these values might only work in a subset of builds...I suggest you leave the PW section to the very end and just work on 1 osc first to tweak things right for your build. In my case with the 2017 suggested values I ended up with a PW range of 20-65%...OK but not ideal. If you end up with similar PW behavior then probably the values I suggested will also work for you.
Jop
Good tips, merci!

Did you fix the spikes on your sine waves?
pix
No, triangle and sins still have spikes.
maxluft
Hi all, would really appreciate help with this. I've got everything working properly except no pulse whatsoever out of any of the oscillators.

The thing is I had it working previously out of two of the oscillators (though it was weak in level) when I had the BCM847's installed. On the third I accidentally cut one of the BCM847's so it didn't work properly, and eventually instead of just replacing this I replaced all the 9x BCM847's with 3x MAT14's, which got everything working apart from the pulse disappearing completely (if I amplify it massively there appears to be some crossbleed from the saw but I don't think this is an issue).

I've checked absolutely everything, all the resistor values in the PWM section, orientation of caps, diodes, transistors and ICs. I've tried with and without each of the two jumpers (JP and TRP). There's continuity between the pulse headers on the main board and the output jack so it shouldn't be anything on the control board. I'm at a loss. Could it be something significant has to be changed to comply with the MAT14, e.g. certain cap or resistor values need to be different? It feels like it has to be something like that given that the same thing is happening across all three oscs. Has anyone else here built the MAT14 version?
jimi23
Finished mine yesterday! Everything works as it should, aside from the issues all mentioned in this thread. Ill need to go in and change the Vibr depth pot as it doesnt do anything until 3pm and then is fully on.

My square wave is pretty slewed, which is odd. Ill have to check all 3


I did get the triangle spike, which is fed into the sine shaper as well. I trimmed this visually, as well as using a spectrum analyser. Its not super clean but good enough to make music with hihi
babysealclub
I think the spike issue can be resolved by adding a cap across rt2a. I don't have one anymore but I did resolve the issue on my build. I went back through my notes and never wrote down the fix. I compared the randomsource schematic and Ray Wilson's vco schematic. It looks like c14 on the mfos shaper is missing. Here is Ray's circuit description which describes the purpose of c14.

U4-A's output is fed to inverting unity gain buffer U4-B. The positive excursions of the outputs of U4-A and U4-B are rectified by D2 and D3 and dropped across R21 (20K resistor to -12V) and fed to U4-C inverting buffer with a gain of 2. This results in a triangular waveform at the output of U4-C after the Saw Offset and Tri Offset trims have been properly adjusted. As explained in the drawing this design produces inaudible glitches at the time the integrator is reset by the comparator. This is because of the finite fall time of the integrator's output. Even though the integrator's fall time is about 1 uS the output of U4-A after inversion and gain takes about 3 uS (due to slew rate limiting of the op amp) to go from low to high (remember its inverted). U4-B takes about 3 uS to go low (its the original with x2 gain, again due to slew rate limitations). Thus a glitch of about 6 uS takes place at the peak of the triangle wave. This glitch is so fast that in practice it contains no audible information. C14 reduces the amplitude of the glitch by filtering the majority of it to -V.

Sounds like this might be the fix. Someone try it!
woodster
This little (42hp) beast is the next project on my bench, so the reappearance of this thread is timely.
Will be keeping tabs on any suggested component value changes etc.
nurbivore
babysealclub wrote:
Sounds like this might be the fix. Someone try it!


I added a 2.2nF cap across RT2. The MFOS docs say 2nF, but I didn't have one. I tried a few others both larger and smaller, and didn't see anything obviously better.

It definitely has an effect, but not huge. I'm not quite confident enough in my oscilloscope usage to say for sure, but at most zoom levels, the spike looks to be much reduced.

That said, I thought the triangle wave actually sounded a little worse, slightly buzzier. The sine, however, sounded WAY better and smoother, without any audible higher frequency bits from the spike. I don't use the triangle wave nearly as much, so this is definitely worth it for me.
Leverkusen
nurbivore wrote:
babysealclub wrote:
Sounds like this might be the fix. Someone try it!


I added a 2.2nF cap across RT2. The MFOS docs say 2nF, but I didn't have one. I tried a few others both larger and smaller, and didn't see anything obviously better.

It definitely has an effect, but not huge. I'm not quite confident enough in my oscilloscope usage to say for sure, but at most zoom levels, the spike looks to be much reduced.

That said, I thought the triangle wave actually sounded a little worse, slightly buzzier. The sine, however, sounded WAY better and smoother, without any audible higher frequency bits from the spike. I don't use the triangle wave nearly as much, so this is definitely worth it for me.


Thanks for trying this out - I will do too and it makes me feel a little more confident beforehand.

On the other hand it seems that while this might smoothen the waveform a bit, which is great for the sine, many builders seem to have the spike also in the saw and square waves. It's not that much of an sound issue there because those waveforms already are rich in harmonics. But it might lead to the assumption that the core problem is not in the added MFOS shaper. I wonder if someone who knows how this oscillator actually works in detail could think of what is going on and why a spike could be expected in an oscillator?
jimi23
I have now done this mod with a 2.2nF cap. Unmodified vs modified scope shots:



Theres a huge reduction in spike. Im getting a large positive spike on the triangle waves regardless of 2.2nF cap:



Zooming in on the triangle spike it hits rail +12v voltage for ~10uS:


The sine wave is much smoother in the peaks:
[/img]
babysealclub
maybe try one end of the cap at the diode / rt2a junction and the other end to ground instead of VSS?
breadman
nurbivore wrote:
babysealclub wrote:
Sounds like this might be the fix. Someone try it!


I added a 2.2nF cap across RT2. The MFOS docs say 2nF, but I didn't have one. I tried a few others both larger and smaller, and didn't see anything obviously better.

It definitely has an effect, but not huge. I'm not quite confident enough in my oscilloscope usage to say for sure, but at most zoom levels, the spike looks to be much reduced.

That said, I thought the triangle wave actually sounded a little worse, slightly buzzier. The sine, however, sounded WAY better and smoother, without any audible higher frequency bits from the spike. I don't use the triangle wave nearly as much, so this is definitely worth it for me.


Gotta chime in here, as I've built a few of these lately and had been sort of frustrated by the voltage spikes, even if I love me some harmonic content every which way. Happily had a few 2nF styros on hand so I added them to my own unit and presto, no visible spikes, greatly improved sine and somewhat improved tri as well. This fix works so well I wish I could get the units out the door back to slap these caps on! I will be including them in any future builds, that's for sure.

So, for anyone skimming:

Add a 2nF cap parallel to RT2A, RT2B, and RT2C!!!
consumer
I have boards with copyright 2016 that I'm looking to build up... my backlog is pitiful.

As I read through this entire thread, what i took from it was that R*S integrated all VALUE adjustments into their 2017 documentation, and (after comparing) the only components that were moved were R30 & R32(A,B & C) -- That is, those components were swapped/flipped (other components were shifted, but it doesn't affect the comparison)

The only other change I found is to add a 2nF or 2.2nF cap in parallel with the RT2(n)

If anyone knows of any other components that were shuffled between the 2016 version of the board and the 2017 version of the (board &) documentation, please let me know so that I can keep an eye on them.

That said, I didn't verify that R*S integrated all the changes suggested in this thread into their 2017 documentation: I can go through the exercise of "redlining" the 2016 documentation to integrate value suggestions if you guys recommend it... I'll happily share my findings, in case someone else has sat on some 2016 boards like me. or something.
Synthbuilder
Does anyone know the value of the three identical surface mount components on the back of the main board that connect the sync input to pin 3 of the LM311? I think it's a surface mount resistor probably 4K7 or something like that.

Tony
woodster
It's been a while since I built mine, but I think they are fine to ignore. Seem to remember them being spaces for alternatives that are already pre populated on the other side of the PCB.
Synthbuilder
woodster wrote:
It's been a while since I built mine, but I think they are fine to ignore. Seem to remember them being spaces for alternatives that are already pre populated on the other side of the PCB.

On my PCB, it's a 2015 version, without any resistor in place the sync input doesn't go anywhere. So they are definitely needed if you want the sync to work. Sadly, the schematic doesn't have any detail on the sync input.

I removed whatever was there on mine when I built it because I modified mine to have a different sync circuit - similar to the Roland System 100 one - which gives the LVCO a more traditional sync sound. But I didn't do the mod on VCO3 because I repurposed the VCO3 sync input to be a second sawtooth output. This allows you to use VCO3 as the master VCO when syncing one of the others without having to use a mult or a stackable.

But I'm now going to sell it, and I thought I'd change the VCO3 sync input back to how it was. I'll keep VCO1 and 2 with the new sync though.

Tony
Synthbuilder
Just an update. I heard back from Ralf at R*S. Those components are resistors and all three are 10K.

Tony
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 15
Powered by phpBB © phpBB Group