PAiA Proteus I restoration

Hello All!

I’m new to the forums, but I think this will be of interest to all.

I am restoring a PAiA Proteus I. I built one in 1982 and gave it away in 1987. Wish I still had the one I built, but oh well. Anyway I found a rough one on eBay. Most people would stay clear. But as the Proteus was a DIY back in the day, I knew it would be easy to disassemble and work on. When I got it, the first think I noted was the all the key bushings were either bad or gone. The keys would rattle around in their sockets. Got on the web and found I could buy replacement bushings. So, I took the keyboard apart and cleaned all the keys. Now that I have all the parts cleaned up and ready to go, I’m looking at what I’ve got.

As you can see the key-bed is a mess. It’s rusty and the paint is flaking off. I don’t think I should disassemble the key-bed completely because the switches are 40 year old plastic, I’m afraid I may break something that cannot be replaced. My plan is to use some Naval jelly to clean off the rust, scrape of all the loose paint then repaint the key-bed. What do you think?

The circuit board isn’t in bad shape. There is one broken trace and several nicks. My plan is to clean the corrosion off with a wire brush then “tin” the traces to protect them from further corrosion then clear coat the trace side of the board with an acrylic spray. Again, what do you think?

I have copper tape that I’ll use to fix the bad trace. It looks like they used a 100w soldering gun on the board. The trace was obliterated. I’ve tested this board and it works 100%. Only four chips, but it was a good first test. The key-scan circuit works perfectly. I hope the key switches all work, but I’ll deal with that later.

I’ve also included a few other pictures of the internals. It doesn’t look too bad. The key-bed got the worst of it from spills and general exposure, but the voice card was tucked up under the control panel and didn’t get too much corrosion. The power supply / backplane needs a good clean up (like the keyboard scanning board) and the same with the controls board.

My biggest problem is just taking my time with it. I want it working now! But it’s going to be a long haul and I don’t want to break something by going too fast.

Good wishes to you all,

Hey! I stored mine in a closet since built, so it fared a lot better over the years.

I haven’t pulled the keys off of my keybed, but from the pictures I’d say that the naval jelly sounds like a good idea.

Acrylic spray on the circuit board makes me a bit nervous. We used a non-conductive clear coat for production equipment. I would just clear a bit of area on the existing trace and solder a small wire (like a resistor lead) across the gap. As long as the traces are conductive, the appearance will be fine.

The boards look great. I’d recommend following the manual’s steps for testing each them. Do you have a copy of them?

The biggest issue I’ve had so far was that I’d forgotten that the knobs/switches don’t affect the sound unless you are in program mode.


Thanks for the reply. I was gettin’ a bit lonely. Just tonight I finished the keyboard scanning board. I completed the “tinning” and cleaned it all up. I agree with you about the clear coat. Not necessary and may cause unintended issues. And it would be difficult to reverse. I am going to clean the frame as best I can, then try a rust reformer, then spray paint the frame. I’m going to mask the switches and the key “springs”. I’m hoping by this weekend to be able to test the keyboard.

I have a low res copy of the manual / test procedures. Adequate but not pretty. I fixed the broken trace more or less how you described. It looks fine and I’m sure it’s working perfectly.

The rest of the boards don’t look too bad. I’m going to clean all the old flux off, do a general clean, check the caps and power them up. If the keyboard works, then I’ll do the backplane, the “controller” then the Proteus I synth itself. It’s all a process and I’m struggling to be patient.

Anyway making slow but true progress.


PS you wouldn’t want to sell your Proteus I would you???

I have the original manuals, so just say the word if you have any trouble reading your copy.

I broke mine out to see if I could fix an issue that I had with an oscillator, but several problems have cropped up on the way. (Like a broken connector for the transformer!) I’m just slightly ahead of your restoration.

In case you’re interested, I found a free VST/Standalone emulator based on the Proteus:

I’ve been wondering how you’re getting along on restoring that keyboard.

The preset buttons on mine are just too flaky. Sometimes you can advance, and sometimes not. Sometimes it advances several presets at a time. I’m convinced that the buttons are just too cheap. Ordered a few replacements and hope to have them soon. I’m in a hold until then.

Thanks for checking in! I’m not going very fast right now. I have taken the entire device apart. The keyboard itself is completely torn apart. I need to do rust protection / paint on the bottom tray and the keyboard frame. Then I’ll be ready to reassemble the keyboard. If that works (there are plastic switches for they key contacts). Then I’m confident I can get the rest of it going.

The power supply will be next. I have the transformer out and disconnected from the power supply back plane. I want to test it first, then connect it back to the power supply and move on from there.

The boards have a lot of rosin flux on them. Over time they have become gummed up with dust, dirt, insects, etc. I’m going to clean the boards to get all that goop off of them.

Really not a lot going on. But it’s still on the bench, my plan is to finish it.

On the flaky patch select switches, I had the same problem with my original Proteus. After about two years, it was difficult to select the patch. I thought that maybe the problem was switch chatter but there was also a de-bounce filter and some circuitry for the patch advance when you hold the button down. The thing is all switches chatter. It may be better to adjust the timing of the de-bounce filter (R4 / C1 on figure 35 in the Assembly Instructions). Tuning that should prevent any chatter from getting through.

I’m finally getting back around to working on my Proteus.
You got me to thinking about R4/C1. With the values chosen, the time constant is very short. Increasing the size of the capacitor would take longer to charge up and could eliminate the bounce. I’ll have to dig up a few caps and see what works best.

Before going too far, I patched in a replacement button across the old one to see if it would make a difference. It does. It’s recognized each time I press it, and I only get a bounce occasionally. I don’t think they’ve ever worked this well. Those original buttons always did seem a bit flaky. It does mean altering the Proteus from stock, but the replacement buttons look very close.

Looks like I’ll have to do implement both modificaions. As I recall, those buttons are soldered on to the board well. I followed those instructions to the letter - including wrapping the wire around each lug. The cap, at least, is easy to remove - or I could just solder another cap in parallel.

I haven’t done a thing. My job has completely overwhelmed me. No time for anything but work right now.
I really need to get some rust protection on the metal parts.

Okay, I’ve got those buttons replaced and I’ve been testing it. Everything seems to be working, but I’ve run into a few things…

The keyboard behaves quite different from what I’ve gotten used to. Since the circuit stops scanning the keys when it finds one pressed, it will not respond to another key unless you release that key first. The highest key seems intermittent, but I believe it has to do with the patch and not the keybed.

I didn’t hear anything from the LFO on step 37, page 48, of the assembly manual. It wasn’t until I turned the ADSR Sustain down that the LFO had any effect. I’m not sure what’s going on yet, but I’m hoping that this is something that will be cured by calibration.

I’m getting some noise through to the amplifier - mostly some noise, but I can affect it by changing the LFO frequency. I suspect bleed through in the mixing circuit.

The great thing about this old gear is that all the circuits are on 10th in center spaced chips and the board is completely open. Really easy to follow the traces. You’ve got a great theory of operation and schematics.

That said, I know what a pain it is to have to chase issues down like this.

Good luck brother,


For me, this goes beyond a typical repair or restoration. My build wasn’t 100% complete: I set it aside when I had an issue with the oscillators that I decided were design flaws - (the output was a result of switching between two waveforms.) I’m looking at my work from long ago with the skills I’ve learned over those years with the objective to finish the job I started… Any issue I have might be a mistake I made when building.

I look at those wires flying across the voice card and wonder… Plenty of ways to introduce noise.

I’ve had some success!

It’s been frustrating. I can read the sustain voltage at the vco, but can’t read anything at all off of the pot. (buildup on the surface of the pins?) I eventually decided to examine the voice card and see if I had any dodgy solder joints. Everything looked fine, but I reflowed much of them anyway and washed much of the flux residue from the board.

Along the way I found a bent pin on IC18. It wasn’t obvious. I had backed the chip out in order to reseat it, and pin 14 didn’t go back into the socket. Serves me right for not doing it the right way. Pin 14 is Vcc, so it’s a wonder that it did anything at all.

I left it on all day. The noise has not returned. Each knob alters the sound in the correct ways. I still need to test the envelope controls so that they are delivering the proper range.

Still a few issues. I dialed in a few sounds according to patches from the Tethys emulator. Not the same, but that’s expected. But the filter doesn’t seem to have as much effect as I would expect. I have to crank up on the mix, and the filter isn’t going into oscillation.

The keybed seems flaky. I don’t hear anything from the highest key, and some keys seem to work intermittently. Could just be the matrix decoding.

Okay. I now understand why this is described as an “ultra-rare synth:” The design is rather odd, and it’s a wonder it works. It’s easy to see why someone could have such trouble getting it to work.

The +/-12v supply employs 7805/7905 5v regulators offset with voltage dividers rather than using 7812/7912 12v regulators. This makes it susceptible to drift, and I’m getting some fluctuation in the +12v supply that may be 60Hz bleeding past one of the capacitors. While I don’t expect precision from it, it is in effect another LFO that can affect anything based on +12v - such as both VCOs.

I think I’ve (finally) gotten mine working!

I recapped the backplane and got a decent +12v DC, but ripple returned once I connected all boards together. Went back to my old analog equipment and narrowed it down to a flaky C8 on the control board.

I still had a bit of fluctuation affecting the +12v line whenever I pressed a key. I already had some problems with the sustain level so I pulled IC12: 3310 envelope generator. No fluctuations noted. Reinstalled the envelope generator and getting a nice, clean sound. LFO is working properly! I need further testing, but it’s looking good.

Looks like I need to rebuild my keyboard now. I had one key that began to stick under one of the other keys. I played around with it for a minute and it suddenly released higher than the others. New set of bushings on order… Once I get them in, I can look into why some of the keys don’t register every time.

How are you getting along with yours?

Just in case someone else runs into this problem:

My success was short-lived. All noise returned, and I could see a 1v p-p fluctuation on the +12v line. Not quite an oscillation - which should have been a clue. The voltage appears to be dipping and then ramping back up. This affects everything! It causes a change in the CV as much as an octave, kills off the LFO, and confuses the envelope generator

The 78L05 +5v regulator used in the +12v supply is held up with a +7v offset and is most likely folding and recovering. After I reduced the voltage to +11v-+11.5v, the keyboard worked perfectly for awhile, then the fluctuation returned. All caps in the power supply circuit have been replaced.

At this time, I will either replace the 78L05 or modify the circuit slightly to replace it with a 78L12.

I think I’m getting close. At last!

I noticed a problem on the display. I could step through the presets, but it would only display even numbers - and I don’t think all of them. Theorizing that this was causing trouble with the memory and DAC, I decided to look into this.

IC12 looked okay. I was seeing +5v and ground levels on Q0-Q3, but I saw about +1v at the display driver IC1! I removed the display driver thinking that it might be defective. The synth sounded normal, but the address lines were still low. The only other components on this bus are the memory (which appears to be functioning) and IC10. After I pulled IC10, the display shows all digits as I progress through the presets, and I’m registering +5v on the display driver pins.

IC10 is used to drive all lines to ground in order to immediately jump to preset 0. Something about this small circuit is dragging my address lines down all of the time and apparently affecting the +5v supply causing many issues. I’ve still got some intermittent ripple on the +12v supply, but at least it sounding more like a keyboard.

This problem has been with my proteus since the first time it was built and I believed it to be a problem with the LFO. This could be a defective part, construction error, or related to the micro-traces I found on the PCBs when I started building it.

And then there’s the keybed… I replaced the bushings, but the white lithium grease recommended doesn’t work well. The keys began to stick after an hour. And more keys are intermittent. Something about the placement.

I was so close a month ago… I should have stayed with it. Sustain was low whenever it sounded okay. Whenever sustain reached the halfway point the keyboard would make a horrible screeching sound. The only reason it seemed to work was that it powered up to a random patch where the sustain level was low.

I thought back to Sam’s videos in the Synth Wizards series: start at the oscillators and work your way through the signal flow. I found that my envelope was peaking at 7v when the specs say that the output of the CEM3310 will be from 5v maximum. Attack wasn’t looking right, either - instead of slowly rising, it snapped to sustain level after a delay. Replaced the chip with an AS3310 I had on-hand and it is working great. Probably presenting too much load on the +12v line as well. I couldn’t see much change in the filter section, either, so swapped out the CEM3320 for an AS3320 as well. That’s sounding good now too.

I’ll take another look at the #0/Norm switch - Probably a bad 4066. Then I just have to finish rebuilding the keybed and tune it up.

I have declared this keyboard fixed so many times, only to have that ripple return to the +12v line each and every time… Finally found it!

There isn’t much circuitry on the +12v line itself to present enough load on the supply. Load is coming from the +5v supply which shares the same smoothing cap as the +12v supply. With sufficient load, the ripple at this junction will increase. In my Proteus, this ripple voltage was just enough to dip below the threshold needed for the 7805 in the +12v supply to function on the peaks – which causes the +12v supply to decrease.

I reduced some of the load by replacing the heatsink on the 7805 in the +5v supply with one that made better contact. That solves an issue when the regulator gets too hot.

I tracked down the main culprit to the AC line! A load from elsewhere in the house was causing my AC to drop down to 108-110VAC, causing the voltage at the secondary of the transformer to drop a few volts which put it too close to the threshold needed for the regulator.

Still have a few issues with intermittent. The mylar looks okay. I suspect contact issues with the connectors, but I noticed a bounce on a few keys which make me suspect the scan circuit. Especially since some keys worked when moving upscale but not when going down.

I’ve finally got the keybed back together and working good mechanically. I did break a key in the process – nothing a bit of epoxy can’t fix. Really slow progress, but I’m looking forward to getting it going.

Congratulations on working out the power supply issue. Makes me wonder if I shouldn’t put in a switcher with a lot of bypass caps.