REI <==> Reverse Engineering, Inc.

Old Crow's Synth Shop: Korg Mono/Poly S/H Upgrade

DISCLAIMER: The information contained herein is provided as-is. No warranties of any kind are expressed or implied; use of this information is solely the responsibility of the persons accessing these pages. The reader is responsible for any modifications or repairs. In other words, I cannot be held responsible if you break something. This is a project page intended primarily for hardware engineers. While I try to make the information as easy to follow as I can, this isn't really a project for beginners. Considerable hands-on experience is *required* in areas such as soldering and circuit-board trace modification. You have been warned.

What is an S/H, Anyway?

S/H stands for "Sample/Hold" or "Sample and Hold". S/H circuits are traditionally found just in front of the input to an analog-to-digital converter (ADC) or attached to the output of a digital-to-analog converter (DAC), perhaps along with a channel selector (MUX/DeMUX). The S/H has the task of "remembering" a voltage and holding it steady while something is done with that voltage. In the case of an ADC, the S/H holds a voltage sample so that the ADC circuitry does not get confused as would be the case if a constantly-varying voltage was trying to be converted to a digital number. In the case of the DAC, S/Hes are commonly used to remember a set of voltages, one per S/H, to set the operating parameters of a voltage-controlled device. Many programmable synthesizers, such as the Polysix, have many S/H circuits in them for both ADC and DAC purposes. The Mono/Poly has an S/H for each VCO pitch control voltage, incidentally. This is so the microcontroller that performs the "poly" part of the Mono/Poly's key assigning can control the 4 VCOs individually.

Another common S/H application for synthesizers is an LFO-clocked "random" S/H, which results in a steady stream of random voltages that change at the LFO rate. Many synthesizers have this as standard equipment, such as the Korg MS-20 (though you must use patch cords to make use of it) or the Moog Rogue, etc. The LFO S/H uses white noise as the input to be sampled. Since white noise is essentially all possible values of voltage between the supply rails, at any given instant the value of that voltage is random. The S/H samples a narrow period of the noise (about 5 microseconds wide) and holds each voltage on its output until the next LFO clock pulse causes a new sample to be taken. If this output is routed to the control the pitch or filter cutoff, the classic random S/H effect is heard.

Mono/Poly S/H: The Quick and Dirty Version

Until I get the digital camera pictures done, here is a short howto on adding the S/H to the Mono/Poly.

1. The first thing I decided upon was how to get the S/H enabled and what to use as the LFO clock. Since there were two LFOs, it was worth adding a switch to select between them for clocking the S/H. To enable the S/H, another switch was used to select between the stock MG1 output and the S/H output. The selected output was sent to the mod wheel/intensity knob.

2. After opening the Mono/Poly (power cord unplugged, of course), I removed the pitch/mod wheel platform by disconnecting the cable from the KLM-353 board and removing the single upper and two lower screws. Next, I carefully removed the pitch and mod wheel assembly from the platform (4 screws underneath) and set the assembly aside.

3. I measured and drilled two holes in the black plastic platform for the two new switches. I had to take care to make sure that 1) the front panel wouldn't conflict with the switches when re-installed and 2) that the switch mounting holes were aligned such that they looked good. I then secured two miniature SPDT switches into the mounting holes and made sure everything was squared up.

4. I placed the mod/pitch wheel assembly back in place, but did not screw it down just yet. I estimated how much of the red wire (the wire that attaches to the MOD WHEEL's 'clockwise' lug) would be needed to reach the MG1<>S/H select switch, then clipped the red wire at that point. The mod/pitch wheel assembly was then removed once again.

5. I soldered new wires to the two switches as noted in the schematic above. VIO = violet wire, ORG = orange wire, YEL = yellow wire, RED = red wire. The VIO and ORG wires go to the KLM-353 board to pick up the two LFO clocks, the YEL wire goes to the S/H board and the RED wire connects to the red wire going to the cable plug in the pitch/mod wheel assembly.

6. The other red wire (the short one coming from the mod wheel that was cut in step 4) was then soldered to its place on the MG1<>S/H switch. The pitch/mod wheel assembly was screwed back into position,and the pitch/mod wheel platform was also screwed back into place beside the keyboard. The cable was plugged back in to the KLM-353 board. This leaves the 4 new wires with one end unconnected.

7. I used a 24" length of RG-174 mini-coax and trimmed and tinned one end to attach to the Noise Level pot (VR17 on the KLM-355). The inner conductor goes to the CW lug and the shield braid goes to the CCW (ground) lug. I used coax for this signal wire as the noise circuit is on the other end of the synth from the mod wheel and S/H board location. The coax is tied down (using the wire ties already fastened to the bottom of the case) and routed over to the location of the other new wires. This is the S/H signal input.

8. After assembling the S/H circuit board, I mounted it to the case about 10cm behind the the pitch/mod wheels. Note that there are three power supply wires (red-white = RED-WHT, green-white = GRN-WHT, and black-white = BLK-WHT) that will attach to convenient power locations on the KLM-353 board. I connected the new yellow and red wires to their board locations and also attached the noise signal coax. The ground shield on this end of the coax was soldered to an unconnected place on the S/H board as the shield is grounded at one end only (at the VR17 pot).

9. The last wiring step was to connect the five remaining new wires to the KLM-353 board. This was possible without having to remove the KLM-353 from the front panel as everything is nicely numbered on the back. The violet wire was soldered to the end of R26 that connects to R22 and the orange wire was soldered to the end of R34 that connects to D2. This leaves the three power wires, which were soldered to pads observed as being part of the power traces. See the eventual pictures for further detail.

10. After double-checking all the wiring, I set the new switches to MG1 and MG1, plugged in and powered up the machine, and used a scope to do a quick check of the S/H function. After verifying its proper operation, I flipped the (new) mod wheel selector switch to "S/H" and had the magic random voltage that Korg should have included in the first place.

More to come (pictures!)

Scott Rider --