When 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!).
It 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.
The “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).
Only 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 woodworking 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.
My customer Aaron, was fed up with the poorly back-lit display on his black re-release SP-1200, and wanted to have the Noritake CU16025-UW6J display from Digi-Key fitted. They are expensive, and being glass, fragile; but they are bright and clear so a good upgrade.
Unfortunately the display is smaller than the aperture in the circuit board, and an attempt at notching out mountings was made to no avail. The display is also thicker than the original and spacers are needed to move it down. After several attempts I decided to make 3mm deep adapters on the 3D printer, and this gave me a stable mounting that I could squarely align the display on. You also have to make an adapter cable as the pins are reversed.
The next issue was that the display character area was smaller than the original, and it did not have a bezel. This looked bad, and the glass was exposed. Thankfully I had an 80×2 line lcd filter lens in my parts bin and could cut that to size. It covered the ugliness nicely, and protected the display. Aaron was happy! Final job was to open all switches and drum pads to clean them and restore playability.
My customer won this working unit on Ebay, and agreed to pick it up from the seller. Between the end of the sale, and it being picked up the seller decided to get his tech to replace the batteries as they were leaking. Following this it would partly boot but hang, not good! Thankfully a heavy discount was negotiated to cover repairs, and in it came to my shop.
I have not worked on one before, and I was amazed to find it is based on the first IBM PC, ISA bus and all. This one had a lot of Forat mods, including the SRAM256 card which I have read is a favorite for folks to blow up by inserting backwards. I checked the batteries that had been fitted, and was only getting 2.4v, and one read shorted. I purchased new lithium batteries and a charger which confirmed that one was indeed shorted. New batteries were fitted. Powering on found the screen blank, and this was traced to the CPU being stuck in reset. The transistor between the batteries and this line was bad, and replacing it released the reset. Checking the CPU clock found it running less than 1/10 correct speed. The osc chip SN74LS04 and its socket were badly corroded by the battery acid; parts replaced and clock working correctly.
We were back to it hanging, but things were now solid as the battery/acid damage was out of the way. Address line A6 was found to be low amplitude, and the SRAM256 card was dragging it low. Removing the card put the 9000 into a mode where it was looking for the SRAM 256 card (Forat firmware and card), so it would not come up. From the schematic, -12v would go to A6 on the SRAM card if it were reversed (surprise!). I traced the shorted A6 line to a PAL on the SRAM256 card, and socketed the chip. Lifting the PALs leg enabled the card to be plugged back in. Bruce Forat was contacted for parts, but by the time he came back with a price I was well into reverse engineering the card fitted detection in the PAL using a logic analyzer and 8080 disassembler. Clearly the address decoding part of the PAL was still good as after zapping the code that looks for the card everything came up. The customer wanted to go this route as he wanted to see what else was bad on the 9000 (i.e. was it a write-off) before spending more on it. As the PAL is socketed he can eventually order the PAL from Bruce, and fit it along with the original firmware eprom.
Testing showed the floppy controller was bad, and a chip change fixed that (did the cable get reversed at some point?). Last thing was to replace the screen with a more modern high contrast led back-lit one. Customer sent me a YouTube video of it in action that night, he is delighted!
Ace Tone Rhythm Ace
I have an interest in old, analog, drum machines, and take them in for service. Sheldon brought to of these model FR-1 units in for service; one was in pieces, the other had gummed up switches, and was generally sad. The biggest issue I has was a lack of schematic, so if anyone has one to share I would appreciate it. They also wired the boards into these units in a way that makes it extremely difficult to work on them. A full service was carried out on both, and both were fully re-cap’d as one had low volume, and the other excessive hum.
There was also a severe crackle on one unit that was traced to a noisy germanium transistor, that thankfully NTE had a replacement for (NTE103A). You had to adjust the bias afterward, but the pot fell apart on touching it, so that was replaced and all was well.
This was purchased by the customer some time ago, and he had stopped using it due to boot problems. It had sat in storage for some time and gradually degraded. It was a missing large touch-pad button, plus the boot diskette had been lost. Without the diskette there was no way of finding out what else is required, and the customer went off to find one.
A boot diskette was finally sourced, and after many attempts it finally booted. The boot issue needed to be addressed, and it showed a need for pad button switch cleaning. The LCD back-light was also out. Continue reading
This E-Mu Drumulator was blowing fuses due to a shorted tantalum capacitor in the power supply. The power supply held the CPU in reset following this, and a CA3086 transistor array was replaced to resolve this, along with the related FCO from EMU. One switch was missing and replaced, the switches were cleaned, OS 3.0 was installed.
via This Old Synth – Vintage Analog Synthesizer Repair Portfolio Page.
Three of these have come in recently with missing sounds or crackling sounds. Two components seem to be failing with time, M5218AL’s and 2sc2603 transistors. The 2sc2603 transistors are a common crackling sound cause (use something like a BC547 but watch out for different pin-outs).
Dead sounds tend to be the M5218AL IC’s. These are SIP’s and getting rare. I have had great success with mounting a 4558 DIP on a SIP module as can be seen on the picture here.
via This Old Synth – Vintage Analog Synthesizer Repair Portfolio Page.