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

Thomas Henry VCO-1 for Eurorack
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author Thomas Henry VCO-1 for Eurorack
dogfaceaudio
I'm currently working on adapting the fantastic VCO-1 design by Thomas Henry to an 8HP Eurorack module. I implemented the resistor changes suggested by Oscar over at the electro-music forum, and changed a few small things. Are there any errors in my schematic or things that should be changed? Thanks!

*edit* Direct link to schematic image:
http://s12.postimg.org/51th1c517/VCO1_rev01.png

executiveBlaster
the image isn't working! I'd love to see.
dogfaceaudio
Added the URL for the schematic image to the original post.
J3RK
How exact are you keeping with the original? I notice that you included reference voltages for the pots for example. If you're open to a little interpretation, I would HIGHLY recommend using the LM13700 instead of the CA3080, as it's still manufactured. Then you have the other half of the LM13700 to either use as a sine shaper, a modulation index VCA, or maybe even a slew for example.

Another thing might be to AC couple the exponential input like the linear input. I would also increase the cap size to allow for lower rate modulations to work better. Something between 1uF and 22uF works well IMO.

You might want to make the integrator op amp a single. I've done it both ways, but settled on using a single op amp after noticing a little bit of odd behavior using dual/quad types.

I absolutely love the VCO-1. In fact, the basic core is where many of my own VCO designs started out. A little while down the road, a little advice from Thomas Henry and Dave Brown (among others) and I'm still using quite a bit of it.

If you're interested a couple of other things I changed are:

Using an LM311 for the core comparator instead of the discrete transistors. The transistors work fine too though.

I would also recommend on the board layout, providing a footprint for a monolithic matched PNP pair for the exponential converter. SSM2220 is nice, and not terribly expensive. (I would also keep the discrete footprints as well. It adds options for the builder.) You could use a THAT 340 and it would take care of the exponential transistors, and also the pair of NPNs in the core comparator at the same time, or better yet the ones in the sine shaper if you keep the NPN based one.

One last item for now. I have a sawtooth shaper that will work with this VCO. If you're interested, I could export a schematic for it tonight.

Anyway, just a few ideas based on my own experience. Your schematic looks good from the short time I looked at it. (at work right now...) Clean and easy to follow.

thumbs up

Happy to chime in again after I look at it more if you like.
nigel
It depends on how much you want to change the original circuit, but I would be tempted to replace the 3080 with one half of a 13700. (You could then use the other half to replace the sine shaper.)

[Edit - I overlapped with the last post!]
forbin
+1 for replacing the 3080 with 1/2 a 13700 -- personally I wouldn't touch the transistors for the schmitt trigger/comparitor or the sine shaper -- i used 2n2222 for the sine shaper in mine and i recon (it might have been luck) but i have got the cleanest sine that i have out of my analog VCO's

The core of this thing is very similar to the OTA incantations of the 258 VCO clones.

Have you asked Thomas about this? just out of politeness...
dogfaceaudio
I'm trying to make a slightly more stable version of the original Thomas Henry design, which is why I added in the reference voltage for the pots. Ideally I want to stay as close to the original with this design, later on I'll experiment with changing things.

The CA3080 is still very easy to find so I just left that as is. I did look for a cheap matched PNP pair to replace the two 2N3906's and found the DMMT3906, but the tiny SOT-363 package would've been too annoying to solder by hand. The SSM2220SZ comes in a nice SOIC-8 package so that looks like a good replacement for the two 2N3906's. Do the two 2N3904's in the sine shaper also need to be matched?

Lastly, is it possible to use a 1.87kOhm thermistor instead of the 2kOhm thermistor? I have a few 1.87kOhm thermistors on hand but no 2kOhm ones.

Right now I'm just doing this design for myself, but if enough people are interested I can ask Thomas Henry if he is okay with me doing a run of PCB's to sell.
forbin
Ideally the sine shapers should be matched but probably not worth a proper matched pair.

You can get away with 1.87K instead of 2K by adjusting the input resistors to the opamp by 1.87/2 = .935 -- so instead of 100K on the 1V/oct use a 91K or work out a parallel or series pair to get you closer... your just trying to keep the gain the same (Rf/Rin). The other input resistors probably don't matter as much as their effect is adjustable anyway... You may well find that the tracking adjustment will get you in anyway...

Oh and since your putting in a voltage reference to make a quiet 10V you probably should use it for the schmitt trigger and R9...
J3RK
Here is the sawtooth shaper that I mentioned. You can get some other interesting shapes out of it too by feeding it a square with a different phase, or possibly even other waveforms. I haven't tried others, but noticed some cool jagged wave shapes while I was experimenting to get it working by taking the square from different places in the circuit. In the UltraFade VCO, I take the square from the non-inverting input of the LM311. You can probably get it from the transistors in the Schmitt trigger of the VCO-1. If I remember correctly, Scott Stites mentioned that in his VCO-1 article.

J3RK
Oh, and also you can add reset sync to this VCO very easily. Put a FET across the integrator capacitor, and run a comparator into the gate. You can use a transistor based comparator, op amp, or an actual comparator IC. As long as it switches the FET to short the cap you're set. This will allow you to completely stop the core as long as the comparator is high. If you want a more typical sync, stick a small cap in line, (or maybe a more fully featured gate-to-trigger circuit,) to make the pulses short.

You can also add an LFO mode, by putting a switch and a larger cap (100nF works well) across the integrator cap. When the switch is on, the larger cap will take effect since it's in parallel to the normal one.

I believe I mentioned this above, but a THAT 340 IC is around $8, and gives you a matched PNP pair and a matched NPN pair in a single DIP-14 package. This would take care of your exponential converter, and your sine shaper. They also have an SO14 package, but that one always seems to be out of stock at Mouser.
J3RK
forbin wrote:
Oh and since your putting in a voltage reference to make a quiet 10V you probably should use it for the schmitt trigger and R9...


I second this.
dogfaceaudio
I'm assuming I'd have to change R9 to 1.0M if I change the input to the +10V reference?

The resistors around Q3 and Q4 would probably also have to be changed, though I'm not entirely sure how to go about recalculating those values.
J3RK
dogfaceaudio wrote:
I'm assuming I'd have to change R9 to 1.0M if I change the input to the +10V reference?


That's what I've done on my newer cores. As far as the comparator goes, I use the LM311, so it's a bit different. You may need to make a few tweaks in the discrete one.
J3RK
Here's my current revision triangle core VCO. I just built it up from scratch because my other schematics were a bit on the illegible side. I added voltage references (my older VCOs just used the rails, which also made the comparator section a bit more complicated than it is with the references.) Attached a PDF as well, as it's easier to read.

J3RK
Also, an OPA134 might be preferrable in the integrator.
dogfaceaudio
Here's the schematic with some of the suggestions added in. I ended up not using the 10V reference for the Schmitt trigger as I couldn't find a good resistor combination in LTspice to get 10Vpp output, I could only get 8Vpp. I kept the transistors for the exponential converter and Schmitt trigger to stay as close to the original Thomas Henry design as possible for this board. I'll probably end up experimenting with a matched pair on another board design. Are there any issues with this schematic, or does it look good enough to start doing the PCB layout?

http://s22.postimg.org/8fkooskv3/VCO1_rev02.png

nigel
I'm fairly certain that Q2 is back to front (C and E should be swapped).
J3RK
nigel wrote:
I'm fairly certain that Q2 is back to front (C and E should be swapped).


Yes, the emitters should be tied together there. Nice catch!
dogfaceaudio
Thanks for pointing out the issue with Q2, nigel! That was a huge mistake on my part. Here's revision 0.3 of the schematic. Hopefully I haven't screwed anything else up:

http://s15.postimg.org/nh6jp8atl/VCO1_rev03.png

dogfaceaudio
Did a quick and messy PCB layout to test this schematic. I sent the design to the awesome people at OSHpark and should have boards in about 2-3 weeks.

Front Panel Board:
http://s16.postimg.org/y3on4qj37/thvco1_panel.png



Core Board:
http://s16.postimg.org/nscac2rdv/thvco1_core.png

calaveras
I was kind of interested in this project if it became a PCB/panel type of party.
But honestly I'm not big on SMT. d'oh!

just my 2 cents seriously, i just don't get it



Ps I know folks say SMT isn't that bad, but I have really shitty eyesight and already have a hard time soldering!
dogfaceaudio
I went with SMD on this one as I wanted to fit the whole thing in 8HP. It wouldn't be that difficult to do a through hole version, but it'd most likely be a larger module.
calaveras
no worries, I just love to whine about my poor eyesight and SMT.
J3RK
calaveras wrote:
no worries, I just love to whine about my poor eyesight and SMT.


I hadn't planned on doing a run, and I'm not quite sure when I'll have time, but I have a similar, small VCO PCB (basically a PicoVCO with added features and improvements in a smaller space) all laid out. I definitely don't have time to do a fully fleshed out project, but I could probably squeeze in a run with BOM and schematic in the somewhat near future if there was interest. It's through-hole. Actually the SSM2220 is SMT, but the rest is through-hole. lol I've also started lining up all board IO so it will work with a SIL header. (which facilitates control PCBs, something I didn't used to do.)

Anyway, let's not pollute this thread with it though. I'll see what my schedule looks like for the next couple of weeks, and then make a separate thread.
dogfaceaudio
I'm certainly interested in this small VCO design, would it work in a Eurorack system? How many HP does it use?
J3RK
Currently the board is 3" x 4". (so 15-16 HP) However, I'm trying to squeeze it all into a 2"x4" PCB right now for fun. I'm not sure if I'll be successful, but we'll see. grin There are 12 ICs in this VCO and a lot of passives. Dead Banana

Edit: I've got the 2" x 4" version almost routed. A few areas are being a little less than... ...shall we say... ...cooperative. I think it will work though. I also had to do a couple of things that I'm not super-fond of. (like put a pulse output near a sine trim...) I'm going to see if I can resolve a few of these items, clean it up to acceptable levels, and we may have a 2x4 VCO. All IO is in-line, so a control PCB could be doable. More soon...
emmaker
Don't think the J3RK VCO will work as shown and needs to be verified and tweaked.

R4 is 100 ohms which would suck ~0.2 amps relative to -10V ground at +10V output. Can't find a spec on the output current relative to 'ground' (which is -10V) but for the positive/negative rail the max specs are 7.5ma/5.0ma.

???

J3RK VCO:
J3RK
You're correct. That needs to be changed. I'll post an update asap.

Edit: 2.2K for R10, 8.2K for R11, 10K for R4, might do the trick. I'm at work, so I'm unable to test this.
J3RK
I ordered some prototypes of this VCO. (actually more than the schematic posted as I included more wave shapers, sync, control/IO, FM, etc.) I've done some simulation, and done a bit more calculation around the comparator, and think I've come up with the right values. I'll give them a whirl as soon as the boards show up. Also, I did manage to squeeze it all into 2" x 4". w00t Sorry for the confusion around the comparator. That was based on some quick simulation while at work. Mr. Green On the Pico and Ultra VCOs I used the LM311 quite differently because I didn't provide reference voltages for the core in those ones. This was a bit new. Anyway, it will be tested in about a week or so.
Tombola
Great minds thinking alike...

I've been doing the same thing...

Mine is built almost exactly as the Thomas Henry Schematic, with a couple of small value changes. So far it seems to work beautifully - lovely sine, square is a bit wide range - dead patches on either end. I got about 5 octaves as close as my Dixie, didn't persist calibration beyond that.
10HP
Parallel boards - one for pots + a few resistors & one cap - one for everything else.
Through Hole
LM13700 + THAT 340 for the matched pairs

Happy to share the eagle files if anyone wants them. I haven't been able to contact Thomas Henry, not really planning anything with this.

VIDEO: https://www.instagram.com/p/_etQT6DmdM/

geecen
That looks great Tom!

Looking forward to seeing the J3RK boards too. cool
J3RK
Tombola wrote:
Great minds thinking alike...

I've been doing the same thing...

Mine is built almost exactly as the Thomas Henry Schematic, with a couple of small value changes. So far it seems to work beautifully - lovely sine, square is a bit wide range - dead patches on either end. I got about 5 octaves as close as my Dixie, didn't persist calibration beyond that.
10HP
Parallel boards - one for pots + a few resistors & one cap - one for everything else.
Through Hole
LM13700 + THAT 340 for the matched pairs

Happy to share the eagle files if anyone wants them. I haven't been able to contact Thomas Henry, not really planning anything with this.

VIDEO: https://www.instagram.com/p/_etQT6DmdM/



Nice!

You should be able to get rid of those dead spots on the PW controls by tweaking the summing resistor values. (R31 / R41) If I'm not mistaken the two will interact a bit, so you'll need to bounce back and forth a little bit.
Isaiah
J3RK
So, is this effectively like your UFO minus the VC-fader?
Will the boards have mounting holes (0.15" from the PCB edges)?
Will a control panel PCB be necessary?
J3RK
Isaiah wrote:
J3RK
So, is this effectively like your UFO minus the VC-fader?
Will the boards have mounting holes (0.15" from the PCB edges)?
Will a control panel PCB be necessary?


It's basically like the UFO minus VC-Fader as you've mentioned, but with the addition of reference voltages for the core and controls (something I've always thought about doing, but then this thread kinda pushed me into implementing.) I've lined up all IO in lines around the edges, so while I wouldn't say a control PCB is NECESSARY, it would be very easy to do. I haven't made anything with Euro-friendliness in mind in a couple of years now, so this seemed like a fun little project with that in mind. Also, I'm kind of going away from doing board runs now, so I can focus on my custom synth work, but something like this would be easy enough to do at some point. After I test the prototype, and make sure everything can stay where it is, I'll make a control PCB for it, and then if people want some, I can place an order for them. w00t Oh, and yes, the mounting hole spacing is the same as all my boards. .15" in from the edges, X.7" center to center. (so if it's a 2" x 4" board, the mounting holes are 1.7" C2C and 3.7" C2C.)
emmaker
J3RK
Quote:
It's basically like the UFO minus VC-Fader as you've mentioned


I'd like to see pins for a fine freq pot and expo/lin FM.

One thing that I was wondering if using the linearization diodes and boosting the voltages in the oscillator loop would make the oscillator better. Anyone done any looking into this?

Thanks
J3RK
emmaker wrote:
J3RK
Quote:
It's basically like the UFO minus VC-Fader as you've mentioned


I'd like to see pins for a fine freq pot and expo/lin FM.

One thing that I was wondering if using the linearization diodes and boosting the voltages in the oscillator loop would make the oscillator better. Anyone done any looking into this?

Thanks


Those are included. The final version will also have initial frequency trim, a switch for mor traditional hard sync as well as the indefinite reset sync thst the Pico and UFO have.

I kind of doubt the linearization diodes would help much. I do use them in my VCF and in the fader portion of the UFO.
J3RK
Ok, here is the final version. Putting in a not so readable image just for a quick visual. Attaching a PDF for proper viewing. The comparator is now completely correct with proper hysteresis. The prototype PCBs were ordered, and don't need any changes, just a few value tweaks (which are present in this schematic.) Feel free to use this however you like. The core obviously resembles the EN triangle core with some updates and improvements. Some of the suggestions for improvements came from Dave Brown (in the form of the Pico VCO and UltraFade) Thomas Henry, and a relative of mine helped tweak a few items as well. The sine shaper is OTA based. 10V reference voltages are fully implemented, so this can be powered by 12 or 15 volts, and will also be quite a bit more precise with tracking and stability. The matched PNP pair is an SMT footprint for SSM2220 or LS352.

Exponential and Linear FM are present. Initial frequency trim is in. Typical hard sync and an indefinite reset/hold sync switch is in place.

This should do the trick for a full featured, accurate and small VCO. (not that there aren't a million out there already hihi ) Still though, it was fun to do, and smashing things into 2" x 4" was interesting. (we need a smiley in a vice)

Oh, one other "SMT compromise" is that the bypass caps for the ICs are 0805. Otherwise all through-hole. While I'm waiting for the prototype I will try to come up with a control PCB. If a few people end up wanting some of these, I'm happy to do a small run of them. I know I could squash a few of them into a few projects.

More soon. I will open up another thread for that after the prototyping is done.



I'm making a few improvements to the board layout, traces, etc. Here's the prototype version:



sammy123
Very cool. I'd definitely be interested in a couple.
Grumskiz
Since I don't have a triangle core VCO in my european racks yet, this is relevant to my interests wink
sammy123
These will come after the dual mirror core. SlayerBadger!
J3RK
sammy123 wrote:
These will come after the dual mirror core. SlayerBadger!


You could actually use these with the dual mirror core control PCB instead of the MCVCOs if you wanted to. You'd get additional things like sync and exp FM, and a slightly easier build.
sammy123
eek! that's amazing. Maybe a Stroh Dual Mirror Core Special Edition in the future? wink

I am committed to finishing the DMC first. Been sitting on it way too long, just waiting on some parts.

edit: of course If I could get a second control board then anything is possible I guess. I'd probably need some help with the new wiring though.

J3RK wrote:
sammy123 wrote:
These will come after the dual mirror core. SlayerBadger!


You could actually use these with the dual mirror core control PCB instead of the MCVCOs if you wanted to. You'd get additional things like sync and exp FM, and a slightly easier build.
J3RK
Actually, I should just lay out a new, smaller control PCB to go with these. I've got better versions of all of the circuits now anyway. For example, for the phase detector, I'd just use an AD633 now instead of a CMOS XOR, as it would be WAY better and more flexible for detection. I'd also use the voltage controlled integrator/slew as it's better than the simple RC network. I think I could also make a smaller/better clock divider that still did half-steps maybe using a cheap readily available MCU. One control PCB for the whole set. Etc. Etc. That will need to wait a bit, but is definitely doable. I need to finish building my blue synth first though. (or I'll never actually get around to it lol )
Isaiah
Apologies in advance for going off topic...

J3RK
This looks fantastic. I'm interested.
While it's small enough for Euro and Frac, I'm glad you kept the integer inch dimensions and CGS mounting hole convention.

I very much hope that the new elements included in your most recent full panel project find themselves available in CGS/Serge-format, or this ?" X 4" size.
Seems a little more flexible than the (mostly) 3"x3" (?) Frac PCBs for use outside that format, maybe.
Maybe just ignore me, I always get a little giddy whenever I see you're releasing something new!
I'm just waiting for enough of your circuits to be available in one format so I can make a dedicated Stroh Modular voice.
J3RK
Isaiah wrote:

I'm just waiting for enough of your circuits to be available in one format so I can make a dedicated Stroh Modular voice.


Actually, just looking at the handful of things that are approximately this size that I've got whipped up right now, this would be quite possible. The main issue is that of time. All the circuits are designed, PCBs are all laid out in multiple sizes and shapes (I get bored late at night sometimes hihi ) etc. The only thing is, in no way am I in a position to pack and ship PCBs, and I can't really ask the distributors to order a whole line of PCBs at a time.

One possible solution would be to build a Voice PCB (like the one I did for the blue synth,) but maybe to make it accessible to more people make it a size that would fit inside a standard 84HP euro-chunk. Then it could really go behind any panel in any format, and I wouldn't have to worry about a ton of separate boards that all have to be shipped separately in multiple quantities. I'll start thinking about that a bit. I think I have everything needed to do a pretty good voice this size/shape.

Now we're way off topic though Mr. Green I'll work on the idea, and post a new thread.
emmaker
J2RK
Just a personal preference but I'd rather solder 0805 or 1206 resistors than vertical mount thru hole.

Thanks for the soap box.
dogfaceaudio
J3RK - That looks awesome, I'm sure if you went fully SMD you could squeeze it down to 8HP. What schematic capture/board layout software do you use?

Tombola - I would definitely be interested in seeing your VCO-1 schematics! I'm also working on an all OTA triangle core VCO based around the LM13700. I'll be starting another thread for that though.
Isaiah
J3RK
Is it possible to implement Linear Detuning with your design?
Could it be done by sending a small variable offset to pin 6 of op-amp U6B?
Tombola
dogfaceaudio wrote:

Tombola - I would definitely be interested in seeing your VCO-1 schematics! I'm also working on an all OTA triangle core VCO based around the LM13700. I'll be starting another thread for that though.


Email me and I'll send them over.
calaveras
emmaker wrote:
J2RK
Just a personal preference but I'd rather solder 0805 or 1206 resistors than vertical mount thru hole.

Thanks for the soap box.

I'd rather solder through hole because my peepers are shot. If I have to start using a magnifying glass to solder it's just not fun anymore.
Isaiah
Isaiah wrote:
J3RK
Is it possible to implement Linear Detuning with your design?
Could it be done by sending a small variable offset to pin 6 of op-amp U6B?


Sorry, I meant pin 3 of the SSM2220.
Would that work?
J3RK
Isaiah wrote:
J3RK
Is it possible to implement Linear Detuning with your design?
Could it be done by sending a small variable offset to pin 6 of op-amp U6B?


Already there. w00t
Isaiah
Great news!
Is it on the schematic? I couldn't see it anywhere... hmmm.....
J3RK
calaveras wrote:
emmaker wrote:
J2RK
Just a personal preference but I'd rather solder 0805 or 1206 resistors than vertical mount thru hole.

Thanks for the soap box.

I'd rather solder through hole because my peepers are shot. If I have to start using a magnifying glass to solder it's just not fun anymore.


Yeah, I hear you. Right now the only SMT parts are the SSM2220 and a small handful of 0805 bypass capacitors. Everything else on the board is through-hole. I did use a few vertically oriented resistors here and there to squeeze things in, but most things are mounted flat.

Vertical resistors are a bit more of a pain to install, but they aren't any taller than most through-hole capacitors, and are solid enough once soldered in. I did try to keep them to a minimum. Typically I don't use them, but they don't hurt anything, and I have a couple of boards that use them without any issues. Before I started trying to squeeze this board down to its current 2" x 4" size, it was quite a bit larger, and not nearly as Euro friendly. I'm not even sure how many people would be interested in something like this, but I will have the protos back in less than a week, so if anyone does want some, it will be possible.

As far as doing an all SMT version, anyone is welcome to take the schematic I posted and lay one out. I'm actually working on another project right now, so I don't have the time to do this myself.
J3RK
Isaiah wrote:
Great news!
Is it on the schematic? I couldn't see it anywhere... hmmm.....


Oh, did you mean a manual control? If so, no that's not on there. I don't see why one couldn't be summed in at U6. I don't think I would go directly into the SSM2220 though.
Isaiah
J3RK wrote:
Isaiah wrote:
Great news!
Is it on the schematic? I couldn't see it anywhere... hmmm.....


Oh, did you mean a manual control? If so, no that's not on there. I don't see why one couldn't be summed in at U6. I don't think I would go directly into the SSM2220 though.


Sorry, yes. I should be a little more thorough with my questions. hihi
Linear Detuning as found in J├╝rgen Haible's Living VCO design.
http://jhaible.com/legacy/living_vcos/jh_living_vcos.html
Shown on schematic page 1, R14,15,16.
I think that arrangement would be wired to pin 3 of the SSSM2220.
Or, maybe summing a DC offset at U6 is the way to do it.
I really don't know if either of these would give control over Linear Detuning though, I'm just guessing... seriously, i just don't get it
J3RK
Some good info in this thread:

https://www.muffwiggler.com/forum/viewtopic.php?t=96124&sid=23d1f27c4a 17e63ac437eda3b7ba585c

Basically you could do it by summing in a control or CV input without an AC coupling cap. I've always AC coupled my linear FM inputs, but jumpering over it would achieve what you want.

That said, I'm not sure there would be that much to gain under most normal circumstances. Most of my VCOs have an initial frequency trim these days, so you can get them pretty spot on with each other via the exponential side anyway.

I think where maybe a DC coupled linear adjustment might make the most sense would be a through-zero VCO. Maybe someone has some other ideas though?
Isaiah
Thanks for the information and the link.

My main thought was being able to maintain a constant beat-rate between VCOs as they track the same pitch CV.
J3RK
So, a little side note. The test PCBs for these have arrived today. I'll get one built sometime soon. This will perform two functions. Test the new core for the FlexWave VCO, and also confirm that these little ones work. I'd be happy to make more of these too if all is right with them. (I believe it is, but will have it verified shortly.)
J3RK
Yikes! I'm aware of the dimensions of 2"x4" but was actually surprised at how small the PCB actually was for this VCO. I think maybe I just got used to seeing it on the screen while I was laying it out. I just need to cut one out (ordered other prototypes on the same PCB) and will have it built up shortly.) I'm thinking this evening.
J3RK
FINALLY cut apart the big prototype PCB for this and a few other small-PCB projects. I'll be building one this week for testing. It's using the same basic circuit as part of the FlexWave which is now tested and working great, so I don't expect any surprises here. I'll post some results as soon as I've got it built. It will either be tomorrow or Thursday, as those seem to be my "synth days" hihi
Phetus
Any news?
J3RK
I do in fact have some news. First, the basic version of this works perfectly. No changes needed. So anyone wanting to etch, or use the schematic I posted, it will work just fine. (I would however swap out the LT1013 after the voltage reference with an LF353.)

There is another version of this in the works now though. I added phase modulation to it. I just need to get permission to use one small part of the circuit before I'm happy posting it. It's a little different than other implementations, so it should be fine, but I'd feel better asking anyway.

The prototype PCB for the simple version works great, so that is also possible to get made, and is a little smaller. I'm hoping to make both available.

More soon!
J3RK
Ordered the prototype for the phase modulation version. w00t

Sin_Phi
Gave this VCO a try, so far it is semi functional. Triangle output looks pretty decent, but is offset about 1v negative. The same for the square wave. All other wave sharpers seem to not be working, they are offset either pos or neg with very low amplitude.Trimmers have minimal effect. Had to make a part substitution for the MPF102, went with the 2N5458, but I don't think that should be causing issues. Used a 8.06k instead of the 8.02k resistor, that may be giving me my offset issues with the triangle. Hopefully I can figure out what is going on with this guy, have not had much luck with DIY modules this week.


J3RK
The wave shapers will be fairly sensitive to offsets. (sawtooth and sine particularly)

When I've built these, I've used a 500R trimmer instead of the 267R resistor. It allows a decent amount of offset and scaling in one trimmer.

8.06K should actually be ok. That's actually what I use because I have a million of them. 8.02K was just the exact calculated value based on 267R and 3.01K around the comparator.

I would just try putting a 500R (or even 1K if that's easier) in place of the 267R and see if you can level that out.

You could also lift one end of any resistor going into a shaper, and make sure none of those sections are pulling the comparator down somehow.

I've built several different oscillators using this exact design now with no issues.

OH!!!!!!!!!! One thing I just thought of. Are you using an LT1013 to buffer the voltage reference? I initially picked this for DC accuracy and low-drift, but it has trouble supplying current to the whole VCO. I updated this on my other schematics, but maybe I left it out on this one. Swap it out for something a bit more robust like an LF353. That may just solve everything for you.

Beautiful layout by the way!
Sin_Phi
Already have the 500R trimmer, might actually need a 1k there, but I don't really trust the offset on my little diy scope. Replaced the LT1013 with the LF353. Also was not powering to ICs from the negative, just a schematic mistake where I linked them but didn't put the power symbol =/. Getting all the waveforms now on two builds. The pulse maxes out at about 80% duty cycle with the first quarter of a turn with 0%. Getting odd sine and saw. Saw trimmers really seem to do nothing at all. Might hit it again with fresh eyes another day as that helped.

I did change out all my OPA op amps for TLs so I wouldn't fry them poking around. Will try a swap again and see if that helps anything.

Thanks for the debug help. Still sounds and tracks great in spite of what I have done to it razz


J3RK
Happy to help! Since both shaper are acting up we should double check the tri and square. On my phone but will reply more later.
J3RK
I could see getting that odd sine wave if the triangle symmetry is off, if there's an offset, or possibly one of the sine trimmers isn't working properly. One of the trimmers can actually cause an asymmetry similar to that, just not as extreme.

Not sure what's happening with that sawtooth. It looks like maybe the square duty cycle is off a bit from the 50% it should be. (which could also happen if the triangle symmetry is off.) I would start at the triangle, make sure it's centered around DC0V, and make sure it is symmetrical in phase/shape.

I'll see if I can come up with some other ideas for you. Did you include the LED driver? I'm trying to think of anything that might pull on the core a bit if something is wrong. If something's pulling a bit too much current I could see the charge cycle going a little funny.

Oh! Just noticed one more thing. It looks like you're using resistors for the power filter. Swap those out for either ferrite beads or even just jumpers as a test. That might clear things up. I've had bad luck using resistors. Sometimes pitch fluctuations (especially while change pulse width on the pulse wave,) some asymmetry, etc. Ferrites have eliminated this when I've seen it.

TL07X op amps should be just fine on this board, so those shouldn't be the problem.

The sawtooth trimmers will be subtle when they're working. A very small DC shift, and the little join glitch at 0V will get slightly bigger or smaller.
Sin_Phi
The square is about 4 time divisions when high 3.5 divisions when low on the scope, so yeah symmetry is off (I was able to make the square a little better by turning ostrim the 500R all the way down). The phase may be a little off on the triangle, but not that much. No LED driver in the circuit at them moment, have had the ic pulled. I measured the negative output of the voltage reference and I was getting -9.65V with the 10ohm resistors, jumpered that and got -9.8V. With the jumpers the module is getting exactly +12V and -12V in. Still getting the same issues with offset and symmetry.

I played with the resistor and trimmer values in the sine waveshapper for a few hours, so I doubt anything is going on there. Might try to get that -10V reference right...

Link to the schematic I had to board made from. Already fixed the negative supply to the pulse wave shaper. Unable to spot any differences from your implementation.
https://dl.dropboxusercontent.com/u/4966230/Synthesizer/STROH_VCO.pdf

*edit just noticed I have C1 as 10nF instead of 1nF, will fix that tonight and see how it goes.
J3RK
C1 could potentially cause some odd behavior, but I wouldn't think it would impact things that badly. I would definitely fix that though.

For the negative reference voltage, you could either tweak the gain a little bit, or try a different op amp. LF442 might work nicely there if you have one, and there's no reason you couldn't try a TL0xx.

I haven't had to tweak the OSTRIM all the way in one direction before with this core. Only a little bit from center. The top of the waveform should stay fairly constant across the trim range at 5V, but the bottom will move pretty dramatically.

Let me know what happens when you swap that cap. I'll see what else I can think of to look at.
J3RK
It's labeled correctly on your schematic, but thought I'd ask. You have 10 Ohm for R9 right, not 10K?
Sin_Phi
Swapped the cap, that helped the square wave get closer to 50%. Changed out R10 from 3.01k to 5.6k and I was able to get amplitude symmetry. R9 is 10ohm.

Tried a TL072 instead of the LF353, no improvement that I could tell. Don't have a LF442 on hand.

Using the LFO cap the duty cycle is 4 time divisions high and 3.8 low, but while in VCO it is 4 high 3.2 low at the same frequency. All the other waveforms of course have the same issue with phase symmetry.

The sine is running with almost twice the amplitude as the other waveforms, but at least it is sine shaped now.

Saw still has the odd bump.
J3RK
Glad to hear you're making progress with it. Very strange about the phase symmetry. The last time I ran into anything like it, it wasn't just a static offset like you seem to have. It was fluctuations due to unwanted current fluctuations. (which is why I mentioned the ferrites/jumpers...)

It seems like something might be affecting the switching points of the comparator. Maybe it's switching perfectly on one edge, and not on the other, so the cap isn't charging/discharging evenly. That would also probably cause the lumpy saw. The thing is, those values work fine on four different implementations of this core that I've recently built.

Nothing immediately jumped out at me when I was looking over your schematic, but when I get a minute, I will compare it to the version in the FlexWave VCO which is now heavily tested, and see if there are any differences. I kind of doubt it though.
cloudberry
Did anything come of this with regards to pcbs? I have found Tom Whitwell's gerber files etc.
Anyone recommend a hobbyist friendly pcb manufacturer in or near UK?
Tombola
cloudberry wrote:
Did anything come of this with regards to pcbs? I have found Tom Whitwell's gerber files etc.
Anyone recommend a hobbyist friendly pcb manufacturer in or near UK?


I used OSHpark for mine. Just looked up my receipt:

VCO1 Pots Board,
VCO1 Components Board
VCO1 panel
$132.40 total for three of each including shipping
cloudberry
Do you have any of those left over? I think it's only the component board I really need.
maltemark
Did the sawtooth waveshaper circuit end up in the final design?
zackH
I'm putting together an order of 10 pcb/panel of the Thomas Henry vco-1 as designed by Tom Whitwell. Anyone interested in one?
cloudberry
Yes please! I might only be after a pcb, but, we'll see.
Any idea of cost?
zackH
Quote:
Any idea of cost?
I'm pricing them out at dirtypcbs.com and it seems relatively cheap, but may take a bit of time. I think I can sell them as low as $12 + shipping for all three different pcbs, the components, pots, and panel boards. Wayyyy cheaper than oshpark, although I love them! If more ppl are interested I guess the price could come down if I get more than 10.
LSuveg1
ZackH -- Count me in for a panel/pcb set please!
zackH
3 left! I just finished the front panel and sent it off to the fab lab
zackH
Here's the component and pots pcbs



still 3 left!
komicone
Zack, I'd be down for a set of PCBs and Panel too.
PM me the details. Thanks!
zackH
komicone wrote:
Zack, I'd be down for a set of PCBs and Panel too.
PM me the details. Thanks!


Aight my man, I will let you know as soon as I get them in. I now have only 1 left!
zackH
komicone
I'm getting these boards from China, I'll let you know when I get them in. They haven't been through the manufacturing process yet so if they're utter garbage I'll let you know that as well... Assuming they're good to go I'll get your email and paypal invoice you.
zackH
The boards are in! They look nice, I'll be PMing those interested shortly, still one set left
zackH
Here's my first one... Thank you Thomas Henry and Tom Whitwell! Sounds good and seems to track well over at least 3 octaves. Now to build the rest of them. I had a spot of trouble with the pots board, odd since it's the simpler panel. For those who are also building this, lmk how it works out for you.

tvainio
Just finished first one of my TH-VCO1s. It's based on Tom Whitwell's design and PCBs were ordered from dirtypcbs.com. Thanks a lot for the open-sourced design files! I like this oscillator a lot and I'm planning to build at least four or them.

I have documented the build process in my blog: https://noiseembedded.blogspot.fi/2017/06/oscillators-more-is-more-th- vco1.html

Has anybody had problems with the Linear FM input? Mine is not modulating smoothly. If I feed triangle wave it acts like it's square wave. Frequency just jumps up and down instead of sliding smoothly. Any ideas how to fix this?
sumedho
zackH do you still have boards left?
zackH
sumedho wrote:
zackH do you still have boards left?


No sry man I'm out... I did redo them slightly to have larger separation between the pads and the ground fill, had a bit of a problem on the top board with the last run. If I do another run I'll let you know. Rockin' Banana!
Stepan Dvoracek
Hi, I would like to change the expo converter into current sink with a matched NPN pair. Should I also modify the wave reset part to reflect this change?
mlatu
Sorry, if this comes along as a dumb question, I'm more of a Software
guy myself, but I wanted to build a VCO according to Mr. Henry's schematics (VCO1).

I'm wondering if I should use linear potentiometers for R37 LINEAR FM (because it's called Linear FM confused ),
R35 COARSE, R36 FINE and R39 INITIAL PULSE WIDTH (Schematic No. 2) or if I should use logarithmic potentiometers everywhere instead.

My suspicion is: linear Pots for R35, R36 and R39, and log for the
others, because those others appear to me to control volume, while the
first ones control different aspects of the sound.

Or am I overthinking this?
Jarno
Linear pots for all control voltages, log pots for audio.
I did a layout for the TH VCO, and I used linear pots for linear FM, can confirm that that works fine.
I built a VERY sparse version of the VCO so mine has a single frequency pot (am going to use multi turn potmeters), a V/Oct input, linear FM with attenuator, and only a triangle out, nothing else (two of these on a single board, without wiring other than the multiturn pots). Also, can confirm that the LM13700 sub works without any adaptations (I used on half of this on each of the VCO's).
mlatu
Well, I've heard about using linear for CV and log for audio, however I'm not sure I can reliably identify which is which in a schematic, that's why I asked.

But it's ok, shipping on mouser is messed up high, so I might as well get 2 full sets, both log and lin and experiment with them to see which pot does what... *shrug*
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 4
Powered by phpBB © phpBB Group