Buchla 208r V2.1 (Blue) Build Thread

davebr · Post by **davebr** » Mon May 22, 2017 1:56 pm

Dimitree wrote:this is what I mean,
what do you think?

Flip the diodes around. They need to pull low but float high.

Dave

Dimitree · Post by **Dimitree** » Mon May 22, 2017 2:07 pm

davebr wrote:
Dimitree wrote:this is what I mean,
what do you think?
Flip the diodes around. They need to pull low but float high.

Dave

cool! thanks!
and based on your prev. reply, I guess I need the resistors to be pullup, and not pulldown as I drawn, am I right?

davebr · Post by **davebr** » Mon May 22, 2017 2:13 pm

Dimitree wrote:
davebr wrote:
Dimitree wrote:this is what I mean,
what do you think?
Flip the diodes around. They need to pull low but float high.

Dave
cool! thanks!
and based on your prev. reply, I guess I need the resistors to be pullup, and not pulldown as I drawn, am I right?

I missed that. Yes, pullup to +5V. However, I believe a 74LS03 will work fine with 5 pullups on the two chips. That is a more direct plug-in solution.

Dave

Don T · Post by **Don T** » Mon May 22, 2017 3:08 pm

davebr wrote: However, I believe a 74LS03 will work fine with 5 pullups on the two chips. That is a more direct plug-in solution.

Dave

I've added a few 74LS03 to my next Mouser order, will report results here as soon as available!

Dimitree · Post by **Dimitree** » Mon May 22, 2017 3:34 pm

davebr wrote: However, I believe a 74LS03 will work fine with 5 pullups on the two chips. That is a more direct plug-in solution.

Dave

thanks Dave!
do you know what's the purpose of the pair of diodes on each of the 3 comparators in front of the NANDs network?
I see a pullup resistor to +5V on each output of the comparator, I guess they need it so the output is always 0V to 5V (since the NANDs want 0V to 5V), but why the diodes?

Don T · Post by **Don T** » Mon May 22, 2017 4:04 pm

Dimitree wrote: do you know what's the purpose of the pair of diodes on each of the 3 comparators in front of the NANDs network?
I see a pullup resistor to +5V on each output of the comparator, I guess they need it so the output is always 0V to 5V (since the NANDs want 0V to 5V), but why the diodes?

I would imagine the purpose is to keep the comparators centered between +5V and Ground, so that the output of the comparators can have a full 5V p-p voltage swing without any DC offset putting the range of the waveform out of the 0-5V range needed by the NANDs. That's just a guess though, I could be wrong. It seems like a way to get around not having a rail-to-rail comparator and powering it off the 5V line. Like I said, just a guess...

davebr · Post by **davebr** » Mon May 22, 2017 6:43 pm

Don T wrote:
Dimitree wrote: do you know what's the purpose of the pair of diodes on each of the 3 comparators in front of the NANDs network?
I see a pullup resistor to +5V on each output of the comparator, I guess they need it so the output is always 0V to 5V (since the NANDs want 0V to 5V), but why the diodes?
I would imagine the purpose is to keep the comparators centered between +5V and Ground, so that the output of the comparators can have a full 5V p-p voltage swing without any DC offset putting the range of the waveform out of the 0-5V range needed by the NANDs. That's just a guess though, I could be wrong. It seems like a way to get around not having a rail-to-rail comparator and powering it off the 5V line. Like I said, just a guess...

Correct. This is why you can't substitute a more modern op-amp like the TL071 as it would swing rail to rail. If you do swing rail to rail one typically adds a series resistor (like R29 and R29) to limit the current and then add diodes to +5 and ground. Don didn't add the diodes but often added a current limit resistor and relied on the internal of the chip for the clamp. I never liked that approach.

Dave

Dimitree · Post by **Dimitree** » Thu May 25, 2017 11:43 am

good to know! btw those 3x LM301 are used as simple comparator rights? to me it looks so, since none of them have resistors used in the feedback loop. But one of them (IC7) has got a capacitor (C7) in the loop.
Does this means that there is some feedback and it is not working as comparator?
Or their output is still 0V OR 5V, with no other in-between states?

Don T · Post by **Don T** » Mon Jun 12, 2017 8:45 pm

Sorry it's been awhile, life sometimes intervenes...

The only card I haven't presented yet, Card 6:

Card 6 builds up and runs as expected with no surprises. On this card, the 2N4339 JFET need not be selected, so you have a place to use at least one of the ones that weren't suitable for the Sine wave shaper!

A couple of notes on this particular card though -

The uA726 has been replaced by yet another prototype board designed by jhulk. Some part values have been tweaked since the last one I built to improve performance. These are now in production and available as bare PCBs. Also in the photo is one I built on a factory-made board to write an updated build doc. These boards are currently available at Modular Addict, and I believe will be carried elsewhere soon.

The other item of note on this board are the Molex connectors. The part number called for in the BOM has recently been marked "end of life". There are still many in stock at various places, but I went ahead and searched for an alternative. The part used on this board is Molex part number 09-48-2061.

http://www.mouser.com/Search/ProductDet ... 09-48-2061

I found the part to be a suitable alternative. The main differences between this part and the previously called for part are: The new part has plated brass contacts as opposed to plated phosphor-broze contacts, and the housing is slightly different, as it is molded to guide the pins into the connector instead of being wide open. I find this new part easier to plug in because of this.

Here's a close up showing the difference. The old part is the translucent one, while the new one is the solid white one:

I hope to get the motherboard going soon!

durwin · Post by **durwin** » Thu Jun 22, 2017 9:51 pm

Very green to this party - where can I source the board/panel/kit from?

captnapalm · Post by **captnapalm** » Thu Jun 22, 2017 11:09 pm

durwin wrote:Very green to this party - where can I source the board/panel/kit from?

https://electricmusicstore.com/products ... 208-rev2-1

http://siddarthianinnovations.bigcartel ... ront-panel

Sammus · Post by **Sammus** » Wed Jun 28, 2017 7:27 pm

With Excelitas virtually unobtainable now (I have 4 for the LPG at least!) I've been looking for alternatives.

There is the Xvive 5C1 recommended by Don T, but I've also found the Macron series on smallbear. Has anyone compared to Xvive for Excelitas?

Cheers

livefreela · Post by **livefreela** » Thu Jun 29, 2017 6:51 am

Sammus wrote:With Excelitas virtually unobtainable now (I have 4 for the LPG at least!) I've been looking for alternatives.

There is the Xvive 5C1 recommended by Don T, but I've also found the Macron series on smallbear. Has anyone compared to Xvive for Excelitas?

Cheers

the short answer, yes, xvive 5c3s can work - my response to another wiggler in a different vactrol-related thread:

i’m one of the opinion (admittedly probably the minority) that the xvive devices can be used with splendid results in buchla modules - however, there is a caveat… a sizeable percentage of them are bullshit.

when i was building my 208r rev 2 I ordered a lot of 30 or so xvive 5c3s and put them on a test rig similar to the one dave brown details on his website. about 25 of them responded very similarly to the excelitas samples I had on hand. however, of the remainder, a few measured almost 2x the on-resistance of the excelitas devices, and the others had this really strange delayed off - taking almost 1.5 to 2 seconds to close after current was applied. were one to plug these into the lpg circuit in the 208 or a 292 i have no doubt they’d report the “not-quite-right” behavior some experienced buchla builders have reported with the xvives. no “snap”, sloppy response, etc...

all that said - I took the of speediest samples (notated during testing), paired them appropriately and set them aside for the “mission-critical” parts of the 208, namely the lgs and the balanced-modulator. I can’t speak to the “authenticity” of my build as I don’t have much experience with genuine buchla, but i’m really happy with how the instrument turned out. Gryphon, I think you’d be fine using the xvives, just make sure to buy a few extra to account for the wonky ones & be sure to test then before installing"

masterofstuff124 · Post by **masterofstuff124** » Fri Jun 30, 2017 12:23 pm

can you link to Dave Browns test rig details? I combed through his website(which is amazing btw) and couldn't find it. I am planning on buying a bunch of 5c3/c1 s and would like to test before putting into modules.

http://modularsynthesis.com

thanks

cygmu · Post by **cygmu** » Fri Jun 30, 2017 2:24 pm

http://www.modularsynthesis.com/roman/b ... 92qlpg.htm

I matched the vactrols so the channels would be near equal. I powered the vactrol LED from +15 volts through a 2K2 resistor and measured the output resistance. The resistance slightly changes as the vactrol warms with the LED current so I used the initial resistance and selected four that were close in value.

we_squirm · Post by **we_squirm** » Fri Jun 30, 2017 3:33 pm

Slightly off topic but has anyone heard from Roman recently? I ordered a 208 V2.1 and 218 PCB set three weeks ago and haven't even gotten a shipping notice. Dying to start my build. I have all the other parts ready to go....

livefreela · Post by **livefreela** » Fri Jun 30, 2017 10:15 pm

this is what i did. i just flicked the power on for a second or so and quickly off - the strange, anomalous xvives have a very pronounced / extended decay upon powering off.

cygmu wrote:http://www.modularsynthesis.com/roman/b ... 92qlpg.htm

I matched the vactrols so the channels would be near equal. I powered the vactrol LED from +15 volts through a 2K2 resistor and measured the output resistance. The resistance slightly changes as the vactrol warms with the LED current so I used the initial resistance and selected four that were close in value.

benadryl · Post by **benadryl** » Sat Jul 01, 2017 1:41 am

I want to build one of these more than anything but have no DIY experience

Starting out with a bastro quattro figaro kit and I've learned a lot so far. Hopefully I'll be able to work my way up to something like this soon.

davebr · Post by **davebr** » Sat Jul 01, 2017 5:20 pm

benadryl wrote:I want to build one of these more than anything but have no DIY experience

Starting out with a bastro quattro figaro kit and I've learned a lot so far. Hopefully I'll be able to work my way up to something like this soon.

Practice up very well. The 208 is in no way a beginner's kit. I get a lot of emails of 208s that once assembled don't operate. The ones that we can't get running via email sometimes get sent to me and I have been sent a lot of them. I have two of them on the bench right now.

There is no silk screen and the resistor lead spacing is different between 1% and 5%. It is easy to put a resistor or capacitor into a feed through hole instead of the pad. Diodes can go in backwards. If you don't solder the Molex connectors on EXTREMELY tight to the PCB then it is a bear getting them to plug in because they flex every which way. With no solder mask it is easy to get bridges. That is what I usually find on the cards.

The motherboard has a lot of components and a lot of parallel runs. Solder bridges on the motherboard are a real problem as that is a multi-hour project to separate, repair, and reassemble. There are a lot of LEDs which need to be aligned. I solder mine in place once assembled, but when you have to tear down, you can't move them at all or they won't realign. People get them in backwards and the legs can bend and touch.

One of the biggest issues I see is poor assembly. The switches need the backing nut. Otherwise when you tighten the top nut you just smash the panel down and then it will never stay tight. Wires don't need an extra 1/4" of insulation removed when soldered to the PCB, especially the power leads. Any shielded wire needs to be small and flexible. It doesn't help to mount the reverb on springs and then use a stiff coax that holds it firm or worse rubs on the springs.

The program extender card doesn't need solder covering the entire pad. It need to be straight and clean so it doesn't short when inserted. I've spent a lot of time chasing shorts that were from misalignment. The first thing I do now is remove that card before troubleshooting.

Do NOT save money by buying cheap sockets or switches. These are the most unreliable part of the module and you want quality components. Replacing a broken switch is not much fun. It's even worse if the banana wires are soldered straight with no slack. With slack you can sometimes lift the panel up high enough to slide a new switch in.

It is not unusual for me to put up to 8 hours in doing a repair, full verification, and calibration on a 208. It is not a beginners project but with enough time they all work and work well.

Just my experience.

Dave

jersupereq · Post by **jersupereq** » Sat Jul 01, 2017 9:02 pm

Has anyone built this with Roman's uA726 replacement boards? I'm wondering if there's any bom for them.

davebr · Post by **davebr** » Sat Jul 01, 2017 11:14 pm

jersupereq wrote:Has anyone built this with Roman's uA726 replacement boards? I'm wondering if there's any bom for them.

I don't know if this is Roman's uA726 board or not but the 208 I have on the bench for repairs has these uA206 replacements. They are just a matched SMT pair and a tempco. Note the tempco is actually wired between the two bases so could be wired to the daughter board instead of the card.

For my own I simply used tape and reel 2N3904s which are matched good enough and epoxied them together and then epoxied a tempco on top. Details of that with measurements are on my Buchla 208 V2 page.

Dave

tarandfeathers · Post by **tarandfeathers** » Mon Jul 03, 2017 1:46 am

Definitely agree. Building isn't an especially cheap way of getting a music easel. It's a good project if you really enjoy planning, processes, electronic assembly and troubleshooting. I had a relatively trouble free build, did not make any mistakes and did all of the necessary modifications as part of my build but I still have a couple of outstanding bugs, and had to make numerous modifications of my own to get it operating how (I feel) it should. I suspect that many people "complete" a build but leave it with bugs/not properly calibrated because they either don't know/care it isn't working properly, or don't have the means to find the problems they do have (or more likely both).

I really enjoyed doing it but it would have been painful if I did not have a fancy PCB assembly jig, good desoldering station, a couple of good bench power supplies so that I could test with confidence that any problems were not power related, and a good oscilloscope. In fact, working on understanding the envelope generator I found that the four channels of my scope were not really enough and I wish I could justify the money for the logic analyser option at the moment. Even without that, that would be a heavy investment in equipment for someone who isn't already doing a lot of electronics.

Not saying don't do it. Just bear in mind that no build is problem free and there are obstacles like none of the available schematics exactly matching the RF version, interactions between the various cards meaning a problem with the complex oscillator (for example) might actually be on card 5, 7, 8, 9 or 10, or a sequencer issue might actually turn out to be on card 3. You need to be able to follow the diagrams and understand what's going on to hunt these things down so consider your level of experience before dropping a lot of money on something that might end up being an (admittedly very pretty) paperweight.

davebr wrote:
benadryl wrote:I want to build one of these more than anything but have no DIY experience :bang:

Starting out with a bastro quattro figaro kit and I've learned a lot so far. Hopefully I'll be able to work my way up to something like this soon.
Practice up very well. The 208 is in no way a beginner's kit. I get a lot of emails of 208s that once assembled don't operate. The ones that we can't get running via email sometimes get sent to me and I have been sent a lot of them. I have two of them on the bench right now.

There is no silk screen and the resistor lead spacing is different between 1% and 5%. It is easy to put a resistor or capacitor into a feed through hole instead of the pad. Diodes can go in backwards. If you don't solder the Molex connectors on EXTREMELY tight to the PCB then it is a bear getting them to plug in because they flex every which way. With no solder mask it is easy to get bridges. That is what I usually find on the cards.

The motherboard has a lot of components and a lot of parallel runs. Solder bridges on the motherboard are a real problem as that is a multi-hour project to separate, repair, and reassemble. There are a lot of LEDs which need to be aligned. I solder mine in place once assembled, but when you have to tear down, you can't move them at all or they won't realign. People get them in backwards and the legs can bend and touch.

One of the biggest issues I see is poor assembly. The switches need the backing nut. Otherwise when you tighten the top nut you just smash the panel down and then it will never stay tight. Wires don't need an extra 1/4" of insulation removed when soldered to the PCB, especially the power leads. Any shielded wire needs to be small and flexible. It doesn't help to mount the reverb on springs and then use a stiff coax that holds it firm or worse rubs on the springs.

The program extender card doesn't need solder covering the entire pad. It need to be straight and clean so it doesn't short when inserted. I've spent a lot of time chasing shorts that were from misalignment. The first thing I do now is remove that card before troubleshooting.

Do NOT save money by buying cheap sockets or switches. These are the most unreliable part of the module and you want quality components. Replacing a broken switch is not much fun. It's even worse if the banana wires are soldered straight with no slack. With slack you can sometimes lift the panel up high enough to slide a new switch in.

It is not unusual for me to put up to 8 hours in doing a repair, full verification, and calibration on a 208. It is not a beginners project but with enough time they all work and work well.

Just my experience.

Dave

Don T · Post by **Don T** » Mon Jul 03, 2017 2:56 am

tarandfeathers wrote:Definitely agree. Building isn't an especially cheap way of getting a music easel. It's a good project if you really enjoy planning, processes, electronic assembly and troubleshooting. I had a relatively trouble free build, did not make any mistakes and did all of the necessary modifications as part of my build but I still have a couple of outstanding bugs, and had to make numerous modifications of my own to get it operating how (I feel) it should. I suspect that many people "complete" a build but leave it with bugs/not properly calibrated because they either don't know/care it isn't working properly, or don't have the means to find the problems they do have (or more likely both).

I really enjoyed doing it but it would have been painful if I did not have a fancy PCB assembly jig, good desoldering station, a couple of good bench power supplies so that I could test with confidence that any problems were not power related, and a good oscilloscope. In fact, working on understanding the envelope generator I found that the four channels of my scope were not really enough and I wish I could justify the money for the logic analyser option at the moment. Even without that, that would be a heavy investment in equipment for someone who isn't already doing a lot of electronics.

Not saying don't do it. Just bear in mind that no build is problem free and there are obstacles like none of the available schematics exactly matching the RF version, interactions between the various cards meaning a problem with the complex oscillator (for example) might actually be on card 5, 7, 8, 9 or 10, or a sequencer issue might actually turn out to be on card 3. You need to be able to follow the diagrams and understand what's going on to hunt these things down so consider your level of experience before dropping a lot of money on something that might end up being an (admittedly very pretty) paperweight.

davebr wrote:
benadryl wrote:I want to build one of these more than anything but have no DIY experience

Starting out with a bastro quattro figaro kit and I've learned a lot so far. Hopefully I'll be able to work my way up to something like this soon.
Practice up very well. The 208 is in no way a beginner's kit. I get a lot of emails of 208s that once assembled don't operate. The ones that we can't get running via email sometimes get sent to me and I have been sent a lot of them. I have two of them on the bench right now.

There is no silk screen and the resistor lead spacing is different between 1% and 5%. It is easy to put a resistor or capacitor into a feed through hole instead of the pad. Diodes can go in backwards. If you don't solder the Molex connectors on EXTREMELY tight to the PCB then it is a bear getting them to plug in because they flex every which way. With no solder mask it is easy to get bridges. That is what I usually find on the cards.

The motherboard has a lot of components and a lot of parallel runs. Solder bridges on the motherboard are a real problem as that is a multi-hour project to separate, repair, and reassemble. There are a lot of LEDs which need to be aligned. I solder mine in place once assembled, but when you have to tear down, you can't move them at all or they won't realign. People get them in backwards and the legs can bend and touch.

One of the biggest issues I see is poor assembly. The switches need the backing nut. Otherwise when you tighten the top nut you just smash the panel down and then it will never stay tight. Wires don't need an extra 1/4" of insulation removed when soldered to the PCB, especially the power leads. Any shielded wire needs to be small and flexible. It doesn't help to mount the reverb on springs and then use a stiff coax that holds it firm or worse rubs on the springs.

The program extender card doesn't need solder covering the entire pad. It need to be straight and clean so it doesn't short when inserted. I've spent a lot of time chasing shorts that were from misalignment. The first thing I do now is remove that card before troubleshooting.

Do NOT save money by buying cheap sockets or switches. These are the most unreliable part of the module and you want quality components. Replacing a broken switch is not much fun. It's even worse if the banana wires are soldered straight with no slack. With slack you can sometimes lift the panel up high enough to slide a new switch in.

It is not unusual for me to put up to 8 hours in doing a repair, full verification, and calibration on a 208. It is not a beginners project but with enough time they all work and work well.

Just my experience.

Dave

Excellent advice from both davebr and tarandfeathers! Perhaps I should update the first post with a list of not-so-obvious tools needed, such as an oscilloscope, and/or a frequency counter or tuner?

I promise I'll get to the motherboard soon, just need to figure out what sliders I want to use.

we_squirm · Post by **we_squirm** » Mon Jul 03, 2017 2:40 pm

Where did y'all source the coloured caps for the sliders? I found Mammoth has ones for the switches but can't find the slider guys anywhere!

Don T · Post by **Don T** » Tue Jul 04, 2017 12:33 am

we_squirm wrote:Where did y'all source the coloured caps for the sliders? I found Mammoth has ones for the switches but can't find the slider guys anywhere!

They are one and the same! Seriously. Warm the caps up (I heat mine up in a strainer over boiling water) and slide them down on top of the sliders.

MUFF WIGGLER

MUFF WIGGLER

Buchla 208r V2.1 (Blue) Build Thread