Help me build Buchla/Serge style Capacitance Touch Keyboard?

[akrearke]

I was told by a most credible source that in fact 2N3904/2N3906 GP trannies should work fine. I'm currently building an ambitious 25 note version. I tried the japanese version with very little luck. that's not to say that it won't work, but mine did not. i think there may be something with the constant current source. It is said that any number of "steps" can be made but i have some doubt about that. I'm it's one of those things that should have added to it "within reason". As soon as I get a working model of this size complete enough to suit me, I am very particular about these things, I'll post the layouts i made. however, I should state they will be from MS paint drawings and will need to be scaled. My goal is clone a Buchla 200 system for myself. any successes will be published here...

[TekniK]

Great news akrearke!

[ericcoleridge]

[akrearke]

I use a standard copperclad board. not so pretty but the keys are lined up right. I used two separate pieces of press n' peel, then lined them up. Now, this may or may not work, there are some issues having to do with the size and thickness of the traces, being a condition of capacitance. After reviewing what photos there are of this body contact to ground type, I made mine to what seemed the norm. i will post some photos when I can get my hands on a camera. This finger-board was then lightly coated with an light laquer. when tested with the japanese version of the buchla circuit I could get some not to trigger, or rather switch, but no output. i think there may have been just enough current to light the LED's some, but not enough actually send a pulse. So I have stripped down the circuit to just 12 notes with two other keys for control, and when money is available I will give that a try. But, this will not be the one of focus, that is the synapse version. I see much more potential from this circuit than with any other. But, it like most things will need some tweaking I suspect. I will keep everyone on this thread up to date of any success in methods.

[scriptstyle]

I'm wondering how the synapse circuit(with some tweaking) matchs up against the paia circuit(with some tweaking)?

[akrearke]

I have not tried the PAiA version yet but have a foil for that one ready as well. The thing about that one is it says in the article that the pads are of a relatively small size, and that changing the size of the pads may affect the circuit functionality. I want something bigger that i don't have to be as accurate with. A touch switch that turns off or on is fairly easy, they are used in lamps, and there many circuits out there for that. I have two articles in an 1982 electronics magazine that detail one type using CA3140 that produces a +5v pulse and one that uses a 2N3904 and a MPF102, also producing a +5v pulse. I am not an engineer, so most of what I do know is from trial and error or really detailed study and and asking questions. What I do know is that capacitance is determined by the size of the conductive plate, the distance between them. changing this plate size will to some degree or another affect the operation. Okay reviewing the article it state a change in the plate size may require a change in the time constant. The time constant is equal to RxC. It further states that the "C4-R7 time constant is long enough that C4 can discharge between pulses." So, this is what your shooting for. Now whether to change the "C4" value or that of the resistor in parallel is the question. Sorry for the long, and possibly unnecessary details. It helps me to think. I'll try to experiment with this later. The coolest TKB to me would be single plates with no surrounding ground plane fitted to a sheet of plexi-glass and epoxied over. This is what I'm going for, personally. But, whatever works you know...

[akrearke]

I feel I should explain myself here. There were a lot of questions raised in this thread early on that I ran into when building the japanese version. It seem that there two things being the main questions raised, (perhaps just out of curiousity). One is the keys or touch switches. I think the material and dimensions will have a great affect on the opporation of the circuit. The second being, the number of keys that can be used (maybe this is just my question), but the point of there be "no theoretical limit to the number of stages" is important in my own case because while i love experimental an uncertain methods of music making, I also enjoy the familiar piano style activation of cv change and pulse output. So, if i talk/type too much tell me shut-up.
Has anyone successfully built a large version of any one of these circuits?
thanks everyone.

[ericcoleridge]

[bugbrand]

Sorry not to have jumped in recently... but....

I've done a compact version with just four keys, mainly for use in my banana systems.

I use PCB front panels for all my modules, though this is the only one currently that has actual circuitry presented on the front. PCB front panels very much lend themselves to these designs -- on this one there is a ground plane all across the front, though it is covered by solder mask.

I get panels (and boards) done by www.pcbcart.com - good price and service! Panels are the 0.126" thickness for zero flex, great strength. I do the designs in FrontDesigner (Abacom -- not FrontPanelDesigner) and then use EagleHPGL to transfer into Eagle.

Quickly reading the thread above -- I think the keys lend themselves best to groups of two or four due to the use of the LM3900. Actually, I only used one LM3900 for the four key version -- one part for each key 'cos I replaced the gate circuitry and everything else down the line.

Cheers,
Tom

[akrearke]

Bugbrand, That is some really cool stuff and looks it awesome. Oh, my location was named after your location.
I have been thinking on it the past couple of days and the Paia version looks like the route I wan to take but the size of the key/ plate is what concerns me. I think I will build one today to see what changing the size of the key does to the performance of the cicuit. I will post the results ASAP.

[akrearke]

Ok. I redesigned the Japanese/218 TKB layout. I will require 3 boards minimum for a 16 note version. The drivers and buffers are separate from the touch plate circuits, which are in sections of 8. I will post everything as soon as one is built and tested. Sorry I'm so slow on this, too many irons in the fire.

[TekniK]

[ericcoleridge]

Really looking forward to seeing this one!

[Coriolis]

Oh yeah - definately been waiting for something like this!



C

[akrearke]

Ok, I just wan to ask a question here... who would be interested in buying board sets for the japanese/218 style tbk? keeping in mind they will be hand made pcb's with tin plating. Who would be interested purchasing a completed 8 or 16 note tkb? I ask because the foils i design are quite large and have found no suppliers online that sell copper-clad board large enough ( for a reasonable cost.) I can buy 14" x 11" copper-clad board for $15 each. anyone interested? If not I can just post the foils and part layouts for all the tkb's designs done so far... and everyone wanting to build one can scale the images and all themselves.

[BananaPlug]

[akrearke]

the touch switches are printed on a piece of pcb 11" x 2". they are oval shaped. As far as the panel is concerned it is a custom thing. Whatever you want it to look like, but again keep in mind, everything is DIY. The circuit run on 15v single rail. My prototype has only one set of assignable cv outputs, as the second row is internal and preset to chromatic tuning. The minimum space requirement I can manage is a panel of 7 1/4 " x 12" for 16 note capacity. The circuit itself is the one posted on this thread, which is a clone of the buchla 218 I believe, using different part numbers, but the design is the same. Outputs are 2 individual cv's, 1 pressure cv, 1 pulse out. There are four jacks for each output section.NO SEQUENCING OPTION ON THE JAPANESE VERSION. power supply included. The way it's constructed is really up to the buyer. I will try to fit it in whatever space is needed, but the boards are long.
Check out this link and the photo of the 100 system. The keyboard at the bottom is roughly the same design.


matrixsynth.blogspot.com/ 2007/02/buchla-100.html

[ericcoleridge]

[akrearke]

The prototype does look a lot like the photo except there are two rows of 8 knobs. I pushed everything to one side so there is room for the joystick circuit and lag processor, and there are 16 notes. The touch plates can be made to look the ones pictured, but the ones I designed look more like the serge style, with the note number and the preset frequency in the foil. Again, a lot can be changed. The main thing is that the capacitance between the positive and ground is close to the same value.
I have considered having the pcbs manufactured, but when I looked into the cost of doing a small run of say 30 units, it drove the cost of production up and the cost to the buyer to over a $1000. I have all the time i need make them, and don't plan on building many, doing all the work myself i can keep the price below $600. The parts run around $300 (it is possible that after the if there is a bulk order of parts this could be more like $250, this only a rough estimate), the enclosure around $75, I do not cut and assembled the wooden parts. So, i'm looking at somewhere between $175 and $250 profit for atleast 3 days of cutting, ironing, painting, soldering and drilling THOUSANDS of little holes.
Doing it your self would save you some cash, but at the cost of a great deal of headache.
The set of boards would be sold for $75 for the 16 note version and $60 for the 8 note version.
I use a dremel and and ruler to cut pcb's. A large disk grinding wheel is most useful for generating a good edge. when a camera is avaiable for me to use i will post photos and sound/audio clips.

[TekniK]

yes plz post pics!

[303monkey]

Sounds Cool.
16 note KB kit for me.

[fonik]

@303monkey
check your PMs. thanks

[Luka]

yeh im keen
16note
ill just take pcbs and touch plates tho
no enclosure or pcb components

[akrearke]

There is a mistake in the my posts about the buchla equivalent of this tkb.
After going through my schematics archive I found that the japanese version and the buchla 112 are the same, not the buchla 218. Sorry, hope this doesn't change the interest. The 112 is still really cool, it delivers all the control listed in my post above. I began to doubt my memory when when looking at another members post, and it hit me that the 100 system had the single sided supply rail and was all transistor based.

[TekniK]

yes,yes,yes,yes,yes,yes still interested