CHD MIDI Upgrade for the Roland TR-808

CHD TR-808 MIDI ModWhen writing about the MTG TurboCPU upgrade I promised to put a blog entry up for a really nice and non-intrusive MIDI mod for the Roland TR-808, and here it is. The thing I really like about this implementation is the degree to which they went to make in non-intrusive (i.e. drilling holes!).


CHD TR-808 MIDI ModIt comes complete with a printed manual of step-by-step instructions, and supporting CD.

For the MIDI connector the existing Din-Sync connector is removed, along with its In/Out switch. A new connector that includes MIDI In/Out is fitted (a special adapter cable is supplied. The switch is replaced by a 3 way one, In/MIDI/Out.

CHD TR-808 MIDI ModThe “umbilical cord” for triggers and power is socketed at the controller end, so for service it can be unplugged and the TR-808 will still work (so this is not a CPU upgrade/replacement like TurboCPU).


CHD TR-808 MIDI ModOnly one hole needs to be drilled, and it is for a MIDI active Led and sits in the Start/Stop switch, under the yellow label. Awkwardness here is that they specify a metric drill, and for accuracy a CHD TR-808 MIDI Modwoodworking one is probably best. I’m sure you could get by with a slightly larger US drill and use a band of shrink-wrap tubing or some-such to tighten the fit. Of course you could ignore this step and just know that center switch is MIDI on!

Be sure to keep the parts removed such that the TR-808 can be put back to standard if need be. Here is a link to the CHD TR-808 MIDI kit. I may well contact CHD and offer to distribute this in the USA.

March of the Juno-106

Juno101_1I do not know why, but March saw a stream of Roland Juno-106’s. Voice module failures were a commonJuno101_4 failure item, and I spent a lot of time running modules through the acetone bath. The acetone bath method is not a cure all, as once a chip is damaged by the epoxy coating shorting it, then it is bad of course. The acetone strip method is a Juno101_5good preventative maintenance item though. I have been suggesting that customers look at services such as the synthspa as they say they will replace bad modules as part of their stripping Juno101_8process, and I cannot compete with that as I’m having to purchase reclaimed modules for repair.

Juno101_2One unit had been dropped which resulted in a lot of broken keys, and a distorted panel. Another unit suffered a broken pitch bend lever, and another the plastic push “spring” for VCO modulation (both of these parts are available from Sam at

Juno101_7Loss of patches due to battery failure was seen on 2 units, so I’m recommending replacing the battery now even if the voltage is good. The bad side of this though is that the patch loading tool for the Juno-106 only does a patch at a time (no sysex load/dump capability), so you have to do the tedious patch edit and save for all 128 patches.

Juno101_6Two units were highly touch sensitive, and this was due to multiple bad sliders. One issue leading to this is oxidization due to a bad storage environment, another is the crumbling gasket Roland used falling into the slider. The cure is to remove, open and clean all sliders, remove old gasket material, and make a gasket from stiffened felt. While I have been in this area I have also been replacing intermittent patch etc. switches, and really they should all be done if one fails as dust getting into the switch is the culprit.

A final issue seen is with the bubble contact strips failing and giving intermittent keys. On the Juno I seem to be seeing other factors such as a liquid spill (on the Korg Poly6/61/MonoPoly they just plain fail).

All of this is leading me to making the standard service for the Juno-106 include front panel switch replacement, slider removal, open, clean, and lube, gasket removal and replacement, battery change, and optional module acetone strip. This is all labor intensive so not cheap, but it would get the keyboard running nicely again, and last another 20 years.

Roland System 100 Model 101

System100_101_1  This unit shipped in and per the customers description, had a dead keyboard and lots of slider issues. TSystem100_101_4he sliders in here are rare, and the worst one was a dual slider type that could not be found anywhere. All sliders were found to be gummed up with old grease, dust, hair, and other debris. The unit had to be deep cleaned, and all sliders opened, cleaned, and lubricated.

System100_101_6The keyboard was giving a large negative CV out, so off the audible scale. The customer said it worked until he tried to clean the contacts due to a few bad keys. What he had done was to overSystem100_101_2 aggressively clean it such that a lot of the CV contacts were dis-formed and resting on the gate bus-bar. These were straightened out, but the majority of keys were then intermittent and badly triggering. Closer investigation revealed that the contact wire that is welded onto the key spring was either missing, or broken at one end such that a consistent on and off point for the trigger and gate could not be achieved. Contact sites were rare and expensive,System100_101_3 so it was a case of using the good trigger contacts from some of the contact assemblies to make complete good ones, then salvaging assemblies from my spares RS-09 (fewer keys) to fully populate the 101. This was a LOT of work, but the keyboard is fully functional now and the owner is delighted.

System100_101_5One other note here is that if someone uses the wrong tool on a ferrite oscillator coil slug as a previous repairer did on this one it will crack and not be adjustable in the future.

Roland TR-808 Step Switches

Roland_step_1I have found a source of the correct original type of step switch used in the Roland TR-808, and Roland_step_2put them in the Parts section of my main site. You have to be really careful on how you remove the switch actuator so as not to break it, and the ribbon cable to the panel is very fragile leading to intermittent/no step Roland_step_3button operation. The originals get gummed up and dusty inside, and they cannot be cleaned. I stock 2 types, the original, and a sealed variant (which has a different action). If you replace a switch I would suggest doing them all.

Rack-mount Synths

MKS_70_1I often get asked to work on rack-mounted synths. Unless they are very early units I have to turn them away as without factory extension cables (which I do not have for MKS_70any model) you can’t get the crammed in cards out on the bench in order to work on them. I this post you can see the amount of stuff that has to come out in order to perform a simple battery change on a Roland MKS-70.

Roland VP-330 Service

Roland VP-330

Roland VP-330

This Roland VP-330 came in for general service. VP330_5As is common with Roland synths, the gasket material had crumbled and a lot of debris was in the slider controls and needed to be removed.  Its amazing that something that was designed to stop dust entering the sliders ends up doing damage (this makes me shudder because of the hours spent cleaning off the foam turned to goo on the Moog Opus 3, Rogue, and MG-1).

I cleaned out the sliders, disassembling a number of them, and lubricated them. The old gasket material was removed, and VP330_2I used double-sided tape to attach stiffened black felt (from Michaels), which I cut slots in for the shafts. The end result is nice, and certainly improves over the look of the gaping slider slots. The felt will shed a bit of course, but the cure is better than leaving it the way it was.

VP330_4I have applied felt to a number of Juno-106’s using variations of the attach method (double-sided tape, glue, attaching to underside of case, or to the top of the sliders). I would be interested to hear of any better materials (neoprene?), along with supplier.

Roland Juno 60

Roland Juno 60This unit needed a deep cleaning session and service.It was disassembled and cleaned-up, with fluff etc vacuumed away.Juno 60 Key contact fluff The sliders and volume control had switch cleaner applied to cure the crackles, and key contacts were cleaned. One bank switch was not working, so I removed, opened it, and cleaned it. The battery was replaced, and FCO applied for dropping into manual mode (I have never had one come in that had Juno 60 FCO at Power Supply endthis dealer applied; it addresses the fact that the power to the front panels goes through a couple of small connector pins, and this was insufficient, so direct power connections are added). Patches re-loaded, and unit put on burn-in test for a week.

Roland Space Echo RE-201

RE-201 with Tape Jammed

RE-201 with Tape Jam issues and no control over motor speed

I have had many requests to service Space Echo’s in the past and have repaired a few. I had decided not to do any more as I needed to purchase a number of service items to service them properly, and I was finding a dire lack of parts such as heads, bearings, and motors. One of my regulars managed to squeeze one in, and then I got a request to service one that was in dire straights after it had been to another repairer that specializes in them.

My next move was to scout and purchase a head de-magnetizer, splicer, splicing tape, and the correct type of 1/4″ tape for these units (must be lubricated tape). The first unit was simple enough, just needing a full service. The one that had been to another repairer was a different story.

The capstan was barely moving, and the motor shaft had a Old and new drive motor upper bearingssignificant amount of play in it such that you had to push down on the shaft with your finger to get things moving. I removed the motor, and did a lot of research on line re replacing motor bearings. After a struggle I got the motor shaft out and measured the upper bearing race, eventually finding a replacement type. The lower bearing could also have done with being replaced, but this is a press-fit sleeve bearing and I could only find one on Ebay, shipped from Australia.

Felt that had incorrectly been glued in the tape pathCleaning the old grease from the motor and capstan shafts/bearings, and replacing the upper motor bearing restored operation and the speed control now worked. The tape loop was crumpled, so I made a new one and fitted it. I then found that the tape was not flowing smoothly in the cassette area. I cleaned a lot of old oxide and sticky stuff from the cassette chamber, and found that a big piece of felt had been glued in the chamber exit path (checked with the customer and the previous repairer had done this, perhaps trying to address the motor issues (the unit was sent to him due to the motor barely turning)). The felt was removed, along with the glue residue. The tape then rang for 6 hours without jamming at various settings of the speed control.

The final work was to glue the front panel back on where it was lifting, clean the tape path, tighten jack nuts, and clean the controls after which it was working nicely. Another 6 hours of test and I will return it to the customer!


Roland Juno 106

Roland Juno-106 800017a chip removed

Roland Juno-106 800017a board with chip removed

A number of these have come into the shop and for most of them it is a 80017a voice module issue. Roland put 3 chips (including the 3109) and supporting components onto a module, then covered it with a moisture sealing epoxy coating. There is a belief that the coating was required in order to protect surface mounted resistors from humidity change in order to stabilize their value, and/or protect their IP. This coating is thought to become conductive after 20 years or so. The only source of replacement parts is through chip pulls from scrap machines, and they are rare so costly. A typical symptom is a hanging voice, and/or crackling. I recently invested in a de-soldering station in order to get these Juno-106 Main cardchips out cleanly. The keyboard here was purchased on Craigslist by the customer for a good price but when he got it home he found it crackled badly, and a voice was hanging (you can isolate which of the 6 voices it is by putting the keyboard in test mode). For this unit it was not obvious which chip it was but careful testing showed it, and it could be verified with an oscilloscope. A replacement was obtained and fitted; all seemed well. After 3 hours of just sitting warming on the bench another chip started to crackle and had to be replaced. Speaking with the chip vendor this scenario is not unusual. The keyboard was not such a deal now that the cost of two replacement chips is factored in but the customer loves it and will get good use from it.

I have been using the acetone strip method to try and re-use a bad module before Roland Juno-106 Stripped 80017a Modulesscouting a replacement. It takes 3 days of soaking and careful work to get residue out from between the chip pins. I silicone seal the module after washing it in order to seal the above mentioned resistors from humidity. This method has been pretty successful with crackly modules, but has not been good for completely dead ones, or where the VCF is drifting as we are looking at chip failures in those cases. It is a sad fact that all of these 800017a modules will eventually fail, and one unit eventually did have all 6 replaced. I have not tried the aftermarket replacements for the 800017a, but would advise folks to consider that. Also there are places that Roland Juno-106 on the benchyou can send the card and they will acetone strip all 6 modules; I’m not sure what the deal is when the strip does not work on a module though.

The bottom line is that if you are thinking of buying a Juno-106 then be aware of its issues and that they are not cheap to repair.

via This Old Synth – Vintage Analog Synthesizer Repair Portfolio Page.

Roland SH-101

Roland SH-101 in red

Roland SH-101

This is the red version, and this one had the hand-grip. It had several issues, the most common of which was a touchy power switch (causes oscillator drift). I go 2 ways on this repair, either wiring across the switch so it is permanently on (no-go if batteries are being used), or order a replacement from TechnologyTransplant; the customer voted for the latter.

 The switch arrived and was fitted.   Switch cleaner was applied to all pots, switches, and sliders. I then re-soldered tracks around audio out and modulation connector, restoring switch operation and getting rid of crackles, and also re-soldered around the frequency and filter sliders as things were intermittent there.

 The bender was only operating in one direction, the LFO push function did not work, and the unit needed adjustment as putting the VCO amount slider forward caused a frequency change. The last 2 items were cured by setting up the position of the bender on the pot shaft. The big issue was that the LFO switch was bad, and they did not make this repairable. I was unable to find a replacement (although having stripped that area out on a Juno-106 yesterday I see it looks similar). I managed to remove the old switch and modify it so that a micro-switch could be installed, and this works fine.

The unit was finally set-up to spec and sounds great!