And now for some pictures of home built synthesizers
Page Last Updated 04-27-2003


This was the second synthesizer (that worked well) that I ever built.  It was built in about 1978.  It has 5 VCO's (one was used as an LFO, but it was identical), 4 ADSR generators, a 4 input voltage controlled mixer (4 VCA's going to one output), a biquad VCF, 4 glides and 8 control voltage outputs.  It was built on 4.5" x 6.5" vector cards that have a 22 pin edge connector.  Some of the cards were converted to PCB's, some are still vector boards.   This thing is not operational at the moment.  One of my previous cats got inside and chewed up all the wiring (a siamese cat).  It has a lot of work to do, plus there was an expansion panel above that needs some serious work done to it.  I also hope to be able to add more VCO's.

This is a vocoder I built.  I built it without ever having seen a vocoder.  I heard one at an electronic music concert at UCSB around 1980, and did my best.  Hopefully, I will soon rework this beast to make it work a little better.  It is constructed all on PCB's, far too many.  The boards were all hand taped, and not as dense as they could have been.

    Latest Update...I have it working again, in fact, just today (4-27-2003).  Here is an MP3 Sample.

        JP I

    This was my first successful synthesizer.  Although, you see it repackaged.  The analog electronics was originally built in 1977.  It consisted of a VCO, VCF, ADSR, Balanced Modulator and VCA.  Latter, I added a Quadrature VCO, and when I repackaged it around 1984, I added another ADSR, a Glide and a Midi to CV interface.  The Midi interface was needed because I gave away the keyboard I had built in favor of using Midi Keyboards.  There is a thumb wheel switch near the power switch to select the mdid channel.  The thing is missing quite a few knobs, they were stolen for another projects, but I soon hope to replace them.


    This was my third synthesizer.  I think I started it around 1986.  It was never finished.  It was to be all controlled by firmware.  It was going to be five voices.  It has 5 VCO cards, based on a CEM3340.  It is on a 4.5 inch by 9 inch card with a 22 pin edge connector.  This plugs into a backplane.  There are VCA's on board for controlling the various modulation inputs, wave form shapers (sine) and other things.  These were the last PCB's that I ever hand taped.  I probably could have got two VCO chips on the card if I had been using cad.  The VCF card has two state variable filters (CEM3350), a VCA (CEM3330), two ADSR's (CEM3310), various control attenuators (two CEM3360), dacs and other stuff.  It is also on a 4.5 inch by 9 inch card with edge connector.  The DCO's were never built.  The MIXER was never built.  The Shepherd Modulator was also never built.  The LFO was never built (this was originally going to be done in firmware).  And the sequencer (also in firmware) was never built.  The Midi interface, which was my first 680xx project (a 68008) handles the midi and talks to the cards.  The firmware was just starting to be debugged when I lost interest in D.I.Y. synthesizers (this was about 1988).  Plans are to soon start back in and finish this.  Hopefully, I will be able to do this.  I have only enough CEM chips to finish this (Except for the CEM3340, I seem to have plenty of those).

This board is one of  the VCO boards for the JP III synthesizer.  It talks to a uP bus that is 6500/6800 compatible (it was originally intended to be used with a 6502 uP, but later, decided to switch to a 68008).  There are five DACs, one 12 bit DAC and 4 8 bit dacs.  The 12 bit dac is for setting the pitch of the of the VCO.  Another 8 bit dac is used for tuning the scale of the 12 bit DAC, if that is needed.  Another DAC is used for setting the Pulse Width, and the other two dacs control control voltage attenuators.  There is also an 8 bit latch for controlling various functions, such as control voltage inversion, wave form select, etc.  The VCO is a CEM3340.  A CA3280 is used to shape the triangle into a sine wave.

    This is the VCF/VCA board from the JP III synthesizer.  It also has a bus that is compatible with the 6502/6800 uP.  This board features two state variable filters implemented with a CEM3350.  The filter can be connected either in series or in parallel.  The two ADSRs are implemented with a pair of CEM3310's, the VCA is a CEM3330 and the four control voltage attenuators are implemented with a pair of CEM3360s.  There is an 8 bit latch for controlling various functions on the board, such as connection of the VCF, control voltage inversion, etc.  Both of these boards were done around 1984.

    This was going to be my midi merger.  It was started in about 1987.  It was never finished.  Mostly because I did not yet have some of the skills I really needed to do so.  I have plans now to finish this.  It is based on a 68008.  One of the reasons it failed was because I still had not purchase a good compiler (that would be in 1988).  I was attempting to use the Mark Williams C compiler for the Atari ST to write the code.  That proved less than successful.  I lost interest in this before the above synth.  But at least I did get the above synth ported to the good compiler before I gave up.  When I start up the Midi merger again, I plan on making the UARTs in a Xilinx Gate array and using a real time operating system that I wrote.

    This is the MOSFET power amp I designed and built.  While it still has negative feedback, it only has 20dB of negative feedback.  It also has a 1MHz bandwidth (treble anyone?).  It puts out about 80 watt's per channel.  It has a dual JFET input (U235), a mosfet second stage (sorry, I don't remember which one, but it is a P-MOSFET).  The output stage is made up of of two IRF120 's and a IRFD9120 and Two IRF9130's and a IRFD110 connected up so that the output stage has a gain of 5, if I remember correctly.  Temperature compensation is a real problem with Mosfets.  This amplifier has a slight thermal oscillation but it is not noticeable (by me).  A newer circuit I am working on doesn't seem to have this problem and soon hope to start implementing it.

    This is a very early work.  It dates to about 1975.  I was still in collage at the time, and it was built out of whatever happened to be around.  What is it?  Why, it is a sequencer.  I don't remember exactly how it works, but, it was based around 4 256x1 ram chips (those were very pricey back in 1975).  The data was stored as 3 nibbles, I think.  The first two made up an 8 bit word that went to a DAC, and the third was loaded into a counter, to set the tempo.  If you notice from the picture above, there is a board with a bunch of 1% resistors.  This was an R-2R ladder (Dacs were very expensive).  The boards were all etched copper clad, this back when I still did that kind of stuff.  It was cumbersome to use, but I had a lot of fun with it.

    This is not a very good picture.  But, this is a mono synth I started to build at about the time when I gave up DIY back around 1990.  It is half done.  The only thing left to be built are the two VCO's.  This thing was based on the CEM chips, 2xCEM3340, CEM3350, CEM3310, and CEM3330.


This is a Rogers Glockenspiel that I converted over to midi.  It used to be part of a Rogers Organ, but, the organ was scrapped, but I managed to get this.

    Link to technical data for Glockenspiel

    As of January 11, 2003, the current status of the Glockenspiel is that all the note now work.  There were two that did not fire because they were stuck in the decaying foam that Rogers used to keep the plungers from making noise when they came back down (they just hit wood).  A third note did not work because of a broken wire.  Taking this thing apart was not at all fun.  I needed to replace that decaying foam because it was pretty noisey every time you played a not.  I replaced the foam with felt, which while it is a whole lot better, still makes kind of a thump when the note is played.  Also, to replace the foam, I had to remove the circuit board that held the solinoids.  To keep manufacturing cost down, it was just stapeled to the frame...grrr.  I had to very carefully pry the stapels out of the phenolic board.  The phenolic was pretty old and brittle.  But hey, now you can play tunes.