Monday, January 16, 2012


wow. this is definitely the longest project i have worked on to date. i began this commission shortly after the "Hartman Mod Job" last february. i was contacted by a guy who was interested in having me build a synthesizer for him that looked like one of Dan McPharlin's miniatures. it sounded like fun and there were some new applications that i was anxious to try out so i agreed. the initial idea was to build a mini synthesizer with a detachable keyboard and as many controls and effects as i could fit. i decided to be courageous and model the toy synth on the notoriously stubborn yamaha pss-30. 

i quickly got to work reverse engineering the main board and taking some dimensions. i made a few rough sketches and headed over to my stepdad's place. he has a wood shop and agreed to help me build the cabinet for the synth. i brought some fancy  hobby wood from home depot that he laughed at. he quickly whipped out a set of cabinets for the synth that i was extremely excited about. i guess i figured that this project was going to be easy... the idea was to build wooden cabinets for the electronics and upholster them in some powder blue tolex i had ordered. i had never applied tolex before but i was confident that i knew what i was doing after reading tutorials and online forums on the subject. however, nothing could prepare me for the giant sticky mess that is tolex. i followed all of the instructions from the tutorial and the directions on the back of the glue and that stuff is wrong! i wont go in to detail on what works because that's what all those other people did, and i thought that their knowledge would work for me but it didn't. basically, if you are going to try tolex, the only thing you need to know is that its going to be a sticky mess the first time, so don't wear your church clothes. it did work eventually though and i got the tolex to stick to the wood. the only problem was that when i trimmed the tolex around the edges, there was a pretty noticeable seam. the was due in part to the fact that the tolex i ordered was powder blue, which is very revealing of imperfections. i guess that's why it was on sale. i figured i would just find something to cover it up since everyone i showed it to seemed to think it was too noticeable too. but i would cross that bridge when i got to it. 

the yamaha pss-30 has quite the reputation for being extremely difficult to modify because it's circuitry is so well consolidated. it is nearly impossible to interrupt any of the keyboards functionality without rendering it useless... this little keyboard caused me a lot of grief but in the end i was able to tame the little bastard. some things to consider when modifying a pss-30; first of all there is a battery alarm tone that goes off something like every 90 seconds if the keyboard is not being played. overcoming this problem easily consumed a month of my time. though now that the issue is resolved i feel like the resolve was pretty obvious. in the end i had to devise a circuit that would continuously pulse a redundant function while the keyboard was not being played in order to reset the battery alarm timer. in doing this, i had to sacrifice the record and playback functions that the keyboard had, and whats worst, i had to sacrifice the rhythms too. i did have a plan to create a different circuit. in which case i would have been able to utilize the rhythms and be able to disable the drum circuit. however, i soon found that the little bastard keyboard had another trick in store for me. the keyboard's chip has two analogue audio outputs. one for the melody and one for the rhythms. however, the two separate channels have bleed through, and the signals can be heard through the others' respective channel. this was not acceptable because like most of my synth circuits these days, the melody voice runs through a frequency divider to generate sub-octaves. when the sound of snare hiss runs through the divider, the output is something like a crackling pop sound and it sucks! i had planned to add a step sequencer to the keyboard anyway, and since the melody section and the rhythm section are on the same clock, it would have been hard to make any sense out of the rhythms constantly pitching up and down. i decided to add a separate drum circuit. i had a couple of these cheap drum machine key-chains that had some pretty nice drum samples in them. i figured i would build a crazy rhythm generator that triggered the sounds, and mix that with the sounds of the pss-30. i found this schematic somewhere that used the cd4089 binary rate multiplier as a pattern generator. i figured i would use those to trigger the drum sounds through a web of dividers and synced to the step sequencer's clock. i built it all and it sounded great. i have to say though, it was kind of odd to hear clean realistic drum sounds against the shredding sub-octaves from the pss-30. i went to work on the filter section of the pss-30. rather than using the usual noisy op-amp filters i usually use, i decided to give the lm13700 VCF a try. it was a little more work but i must say, the sound is way better then any other filter ive built. i figured out a sneaky way to steal a little current from the key matrix of the pss-30 and turn it into a trigger without disturbing the functionality of the keyboard. i used that signal to trigger the envelope generator that modulates the cutt-off of the filter. now that the synth is finished, i do regret not putting in a more sophisticated envelope generator. it works but its range is low. oh well, its great at what it does. next i built the step sequencer. in the past i have always used the cd4017 chip for sequencers but this time i wanted to have the ability to run the sequencer up or down, so i drew something up using the 74hc193 counter and the a cd4051 analogue switch. it turned out great! i could run the sequencer in both directions and i was able to do other things too. i added a feature that would hold the last step in the sequence until the the keyboard is inactive, and i added another feature that gates the sequencer from the first step every time a key is pressed. i even added a rotary encoder circuit to be able to jog up and down in "program" mode to individually tune each step or to just transpose the keyboard quickly. i gave the sequencer its own cd4089 pattern generator and moved on to incorporating the melody and rhythm circuits together. ok, so reading back a bit i realize that i have not mentioned how long i have been working on this project and how many shitty and/or distracting things have happened to me. by this time in the project i am going to say that it is probably around july... anyway, so i am going back to the drum circuit in all its glory and i want to put all these circuits together. unfortunately the drum circuit doesnt work any more. after spending a week or two trying to figure out why it suddenly doesnt work any more and replacing everything including the drum circuit its self, i came to the conclusion that it was totally fucked. i was left with a big hunk of circuits and two fried drum toys. not a problem. by this point im so fucking jaded that it doesnt even bother me. and for some reason the dude who i am building this for is not bothered that i have spent six months on this shit and there is no end in sight! sweet... all the while i am thinking of all the other stuff i'd rather be building. so i take a little break and do a rebuild of one of my older projects that i sold to another guy some years ago. more details in the previous post. 

(dead drum) 

after a much needed break from this project, i came back refreshed and decided to use an analogue drum sound generator. why not? it would sound great with  the pss-30. i grabbed a little KoolShades drum machine from the basement and reverse engineered it. i decided that each of the four drum sounds should have its own pattern generator, pattern divider, volume control, and accent control. on top of that there is a pattern reset switch for each of the pattern generators to be able to sync or de-synchronize patterns from each other. there is also one "all" reset switch that resets all of the five patterns at once. it came together really well. by this time in the project i was unstoppable. if an issue came up, i knew what the problem was without even looking at it, and before i knew it, the circuit was built.

now i had to think about the case. after the mess with the cabinet and the tolex i was just not satisfied with the edges of the case. i figured i would find some kind of decorative edging to cover the seams of my amateur upholstery job, but when i hit up google i was left disappointed. now i really must say that this was definitely the most emotional part of the project. i could not for the life of me find anything remotely close to what i was looking for for a whole month! and its all i did for the whole month too! i was consumed. all i wanted was some of that cheap shiny edging that you see everywhere on everything... why couldn't i find it? what is it even called? after a month of wasted hours of google searching and bad dreams about it all, i finally found this thin steel tape that is apparently used for tagging machine parts in big dirty shops. the stuff is crazy durable and polished, and very flexible. this would have to do.

next up; faceplate. designing the faceplate and the whole interface was really fun and pretty easy. i sent my design to pololu to have them cut it like i have before in the past, and began work on designing my graphics. when the faceplate came, it was opposite the direction i needed it so they sent me another one no charge. sweet! now i have two. 

sad to say though i probably wont send any more designs their way since i have a laser of my own now. now that i had the right faceplate i needed to think about how i would apply my graphics. for some reason i thought that vinyl was an option. hahaw. no, vinyl was no option. i then started looking in to companies that offered dry transfer services but that shit is offensively expensive. after wasting more time on dead ends and being offended by returned quotes, i landed on screen printing. i had never screen printed anything before but how hard could it be to screen print 12" by 12" of 8pt text on a smooth surface? well, its pretty fucking hard. i bought all the stuff and did some trials and after being totally frustrated by it all, decided to just go for it. the result was actually way better than i would have expected. there are a lot of fat smudgy parts, but all in all it passes. 

at this point in the project i was pretty distracted with traveling to new cool places with my GF so it was kind of slow to wrap up. i gave the project a few hours of my day each day before work and a little more on my days off. once the wretched holidays were over i put it all together like a puzzle. and so here we are. almost a year later and i'm left with a big beautiful symbol of the last year of my life that i cant remember very well. check out the demo video for more details. as always, click on the pictures for larger sizes. 



  1. it is fantastic!!!
    so it was some order or you gonna sell it now?

  2. yeah, its already sold. believe it or not the guy stuck with me even after a year of excuses and extensions. i guess he really wants it.

  3. yeah, i'd want such beast too...
    how much it costs? if not commercial secret ))

  4. Can you explain :" i figured out a sneaky way to steal a little current from the key matrix of the pss-30 and turn it into a trigger without disturbing the functionality of the keyboard "
    Have you added a parallel diode matrix on a row of existing diodes mixed to a debouncer ?

  5. exactly! if you look at the schematic you'll notice though that one of the diodes is connected through the 4066 switch. this is to keep the filter from being triggered by the battery alarm disable circuit.

  6. I want one too! What would it cost if you built a second one for me?

  7. I have said it before and ill say it again... you are my hero.

  8. yeah, i'd want such beast too...
    I want one too!