Guts of a Korg MS2000

I bought an MS2000 with an issue, described below, because I enjoy a challenge:

1) On power up, the LCD shows “IPL s.p.u [ ]”, as if the synth is in update mode

2) If powered up using e.g. the “Mod Seq On/Off” and “1” buttons, the LCD shows “KORG MS2000” for a short while, then comes up with the A0 Stab Saw patch loaded.

3) When connected over midi, the synth is sending midi data – e.g. the keyboard and the controls send midi messages.

4) Patches can be changed using the numbered buttons, and the various LEDs on the synth change accordingly when patches change, however there is no sound from the headphone jack when the builtin keyboard is played, or an external midi keyboard is played.

5) If I attempt to load the OS 1.07 system update with the synth in “IPL” mode, using the Korg updater, then the LCD display changes to “IPL s.p.u [hrd]” and the update utility shows “Receive Error Timeout MS2000 did not respond”. The Seq 3 and Amp LEDs flash slowly when in this mode.

6) The midi channel the synth is working on appears to be 1, as this is the channel that keyboard messages and control changes get sent out on from the synth.

I’ve been through the synth internally, cleaned all connector contacts, and reseated everything. Nothing looked amiss on either of the main boards (no odd looking solder joints or discoloured components). I’ve checked the voltages on the supply, 3.3V, 5V and 7V and they are fine.

If I send SysEx messages over midi, the synth responds with two or three bytes, the values of which vary.

Here are some photos of the guts of the synth:

After some voltage measurements on the PCBs, it was evident that one of the TTL chips on the keypad circuit was faulty – the chip in question a 74HC138. After replacing that chip, everything improved! All the buttons and LEDs began working, and the synth booted up in the standard mode.

However, it produced no sound. After running a couple of the diagnostic checks, the following dreaded information appeared on the LCD:


This indicates the DSP chip is probably toast. A replacement DSPB56362AG120 has been ordered, but this is a 144 pin surface mount component, and may be tricky to deal with.

Update: I removed the existing DSP using Quik Chip:


After soldering in the replacement, and checking for shorts and continuity from each pin to the PCB, I found that the synth still showed the DSP error. I then went around the circuit checking voltages and signals – the DSP appeared to be receiving signals from the CPU, but doing nothing with them. My guess is that either the new DSP was a dud, or I somehow managed to destroy it in the process of soldering it in. So, I ordered a second replacement, de-soldered the “dud”, and replaced it with the fresh one – and success! – everything worked, the MPUtoDSP error was no more, and the synth produced sound.



Replacing the LCD display on a Roland MC-505

Here’s my MC-505 as it arrived from the seller on Reverb:


The LCD display (the orange rectangle top centre) shows the common problem that these units suffer: many pixels are inoperative. A closer look:


The first thing to try as a fix is to re-heat the LCD ribbon cable, to re-make the electrical contacts. We need to extract the LCD from the unit first. Start by removing all knobs from the front of the unit (they just pull off) and opening the case:

Photo Jun 09, 13 59 22

The white ribbon cable  attached to the upper main board is for the LCD – it just pulls out. To extract the LCD itself we need to get access to the screws which attach it: these are underneath the plastic screen at the front of the MC-505. Some people use a knife and pry it off from the front: a bit risky, as the screen could bend and break. Instead, remove the mainboard, which involves detaching all the ribbon cables, all the jack sockets, and several screws. Take photos of everything as you gut the machine:

Photo Jun 09, 14 01 37

Cut any cable ties as necessary, such as the one in the photo above (and remember to replace them on final reassembly). Here we see the interior after removal of the main board;


We can see the rear of the LCD display top centre. Now remove the large board that holds all the keys and switches. No need to detach the ribbon cables from it. There are many small screws that attach this board to the case. Here’s the key switch board removed:

Photo Jun 09, 16 22 41

If you are going to clean the switches etc., remove them from the board – the plastic piano key switches have small latches that hook them to the board, but are not hard to remove. Be very careful, as they are quite flimsy. Here’s a view of the case, without the boards, but with the various switches still in position. Remove them for cleaning if desired. With access to the rear of the plastic screen, it can be gently pushed outwards by applying firm pressure through the aperture – it’s only held in by double sided tape, so little effort is required.

Photo Jun 09, 15 57 01

With the screen removed, the LCD display unit can be accessed and removed from the front of the unit – it’s held in place by four small screws. Here’s the case with all switches, knobs etc. and the LCD display removed.

Photo Jun 09, 16 15 49

With the LCD display removed, re-assemble the unit’s circuit boards, including the power supply board, but leave the back of the unit’s case off so that the innards can be accessed. From the front, feed the LCD display’s ribbon cable through the aperture and attach the ribbon cable to the main board. Now the machine can be powered up again.

If you are lucky, you can fix the missing LCD pixels by heating. There is a wide ribbon cable that connects the LCD display at the front to the LCD circuit board at the back. This ribbon cable is poorly connected. Using a soldering iron, and while the machine is powered on, so you can see the results, run the tip of the iron slowly backwards and forwards along the ribbon cable where it attaches to the circuit board. You should see pixels reappear as the connections get re-made.

In my case, I could get most of the display back using this method, but parts were still missing no matter how many times I applied heat.

I ordered the following:

Arducam 1602 16×2 LCD Display Module Based on HD44780 Controller Character White on Blue with Backlight for Arduino


for $5.99 on Amazon. I then removed the ribbon cable from the old LCD display by de-soldering it, opened up the individual wires in the cable, and soldered it to the pinouts on the Arducam unit, referring to the Arducam datasheet and the MC-505_groovebox_SM .

Photo Jun 11, 17 17 44

I added a 680 Ohm resistor between the VDD (+V) pin and the A pin,  and a connection between the K pin and VSS (ground) which adjusted the brightness of the backlight on the LCD to a good level. The display worked:

Photo Jun 11, 17 17 38

Now all that remained was to position the new display in the aperture, drill a couple of new holes at the left for mounting purposes, and fix it into position.

Photo Jun 11, 17 52 13

Finally, I reattached the plastic cover using some double sided tape. The LCD unit stands a little higher than the original, so the cover sits a bit proud of the unit – not a big deal. It certainly looks a lot better than the original, and is fully readable!

Photo Jun 11, 18 00 08


Mitsubishi LT-70 Linear Tracking Turntable

This is an LT-70 that at one time was part of the Mitsubishi “Audio Intelligent System Model DA-L70/LT-70”. I am trying to get it to work standalone.


On the back panel, there are phono outputs for the cartridge, and an 8-pin DIN plug that connected to the rest of the system.

DIN plug

The wiring is as follows:

  1. Pin 1: Yellow -12V
  2. Pin 2:  N/C
  3. Pin 3: White (“AF”?)
  4. Pin 4: Orange +12V
  5. Pin 5: Black (“SYNC”?)
  6. Pin 6: Braid 0V
  7. Pin 7: Red (“STP”?)
  8. Pin 8: N/C

This plug is marked “TO CACEIVER” J106 on the following schematic for the turntable:

LT-70 schematic (part)

The same plug is marked “PL CONT” on the schematic for the main unit:

DA-L70 Schematic (part)

Working from the schematics, I saw that the turntable uses a -12 0 +12 power supply, which I’ve duly connected to pins 1-6-4 on the plug (29,28,27 on the turntable schematic). Sure enough, the unit powers up, the track indicator 7seg LED lights up, and I can operate the tray using the “Open” button on it, the fwd/rev buttons that move the cartridge left/right, and the various programming buttons on the tray (track select, program, etc.)

However, the turntable never spins, the “Start” button has no effect. I think this is because I am not providing the correct signals on the pins marked “SYNC”, “AF” and “STP” (shown on the main unit schematic).

Two of these three pins (24 & 25) are connected to IC151, a BA612 quad driver, on the turntable circuit board: 24 seems to be an input signal to the turntable (since it goes to a BA612 driver input, pin 2). The other, 25, looks like an output signal from the turntable to the main unit, and measures +12V when the turntable is powered up. Finally, pin 26 also looks like an output, some sort of current source from Q221?

EDIT (April 2016): I don’t know what I was doing wrong above, but I just checked three years later, and the unit works fine using the connections specified above. I was able to play and listen to an LP without problem.

Summary: open the DIN plug connector on the cable coming out of the unit, and connect the following wires to a 12-0-12 power supply.

  • Yellow to -ve 12V
  • Orange to +ve 12V
  • Braid to Ground

Connect the RCA plugs (left and right) to your amplifier.

The red wire, if touched to +12V, will stop playback, but it’s not required.

Tube Adapter for the Tektronix 575 Transistor Curve Tester

The Tek 575 Transistor Curve Tester can be used to measure tubes like the twin triode 12AX7 (which has a 12V heater requirement) and the 6DJ8 (which uses 6V at the heater). The 575 can sweep collector (plate) voltage between 0 and 200 Volts, which is a useful range for these tubes. All the 575 is missing is a heater supply, and a suitable tube socket.

I constructed a tube adapter that can plug in to the banana plug sockets on my 575’s DUT panel, using a panel from a parts 575 I acquired years ago (to replace the bad CRT tube in my original 575). The angled faces on the panels combine to result in a horizontal surface, which I used to mount the tube socket (B9A), a pair of terminals for an external heater supply (I used a 6VDC wall wart adapter rated at 300mA), and a single pole double throw (SPDT) switch to select between 12V and 6V tube types.

Here is the completed adapter, measuring the transfer curves for a 12AU7A dual triode.

Tektronix 575 Tube Adapter

One convenient aspect of the wiring in the adapter is that the tube’s two triodes can be compared easily by flicking the 575’s transistor A/B select switch. This allows an instant comparison of the transfer curves, and shows how well (or badly) they match.

The SECO Model 250 Transistor and Tunnel Diode Tester

This is a quite rare device I picked up on Ebay some years ago, for testing transistors and diodes, including tunnel diodes.

SECO Model 250 Transistor Tester

 SECO 250 Transistor Tester

The instructions are on a label pasted to the inside of the cover, and are unfortunately in a sorry state.

SECO 250 Instructions

Here’s a better image from a recent (2012) Ebay auction for the same type of tester:

Instructions (found elsewhere)
Instructions (found elsewhere)

A Modular Analogue Sound Synthesizer

This project was begun in October 2009. After completing my Minisonic 2 synthesizer, I slowly came to realise that I would like to build a modular synth that allowed me to experiment with various ideas I had for analogue sound processing. Over several years I accumulated parts and components that I thought might be useful. At some point it came time to actually start building the synth.

Design Goals and Details

  1.  “Cheap and cheerful”: no fancy and expensive enclosure, keyboard, minimal cost, maximal use of components on hand
  2. Modular: allowing existing modules to be easily removed and replaced by different modules
  3. Simple PSU: the use of a commercial pre-built PSU (e.g. for a PC) to avoid the boring and frustrating aspects of building a decent one myself. (The PSU is a critical component, and using a commercial unit is therefore preferred)
  4. Components: an eclectic mix of discrete transistors, linear ICs, CMOS, latest generation of linear ICs (e.g. via the free sample programmes of Analog Devices and Texas Instruments),  vacuum tubes and moving coil meter(s)
  5. Module interconnects via banana plugs and sockets
  6. Module face plate graphics designed using PowerPoint

Anticipated Modules

  • Two or three VCOs. These may be based on the Minisonic VCOs, or use purpose ICs such as the Analog Devices VFC32
  • At least two VCFs. Probably Sallen Key or Moog ladder varieties. VCFs are the most important modules IMHO.
  • A 10 step Sequencer allowing variable clock rate stepping, manual stepping, or the input of a clock signal
  • One Ring Modulator based on the Minisonic design (I have one of the SG3002 chips spare)
  • BBD delay, based on the MN3208 chip
  • One dual channel audio amplifier module, also with Line Out, including a single decent size speaker
  • One noise source, using the Minisonic design around the Z5J noise diode (which I have a spare of)

Dual VCO

This is the dual VCO I designed. It has a single power rail and uses a single LM358 twin opamp for each VCO. A third LM358 is used to set the output level for a switch selected triangle/ramp or squarewave. The “Shape” controls vary the ramp from being left handed through triangular through right handed. Unfortunately the Shape control affects the frequency of the VCO.

The “Sync” control allows VCO1 to modify the charge cycle of VCO2’s capacitor, and hence to synchronize VCO2 to a multiple or integer dividend of VCO1.

The inputs are linear only. A separate log voltage control will be available in a separate module.

With zero volts at VCin the “Freq” controls can be used to swing each VCO between sub-Hz and around 8kHz.

Here is the front panel of the Dual VCO

Ladder Voltage Controlled Filter


The cabinet is constructed from an old drawer. The base of the drawer, 1/4″ particleboard, was removed and cut into a set of module face plates: six of size approximately 5 3/4″ x 5″ and three of 5 3/4″ x 10″ (for larger modules e.g. the sequencer).

The frame of the drawer was strengthened by two struts, which will also act as attachment points for the module faceplates. Each module will be attached using a pair of easily removed small screws at one or both vertical edges.

Here is the frame and module faceplates after cutting.


The sequencer allows up to 10 voltage levels to be repeatedly sent to another module. Each voltage level is set by a pot labelled “0” to “9” on the front panel. The sequence length is controlled by a rotary switch that selects between 1 and 10. An internal clock determines the rate at which each of the voltage levels is presented to the output. The clock rate can be controlled by the “Clock Rate” pot on the front panel. The clock can also be turned off (using the “Clock/Stop/Ext” switch) to allow either single stepping through the levels (using the “Step” pushbutton) or an external clock input. When the internal or external clock is operating, an LED next to the “Clock Rate” control flashes in time with it.

As each step is selected, a large Nixie tube display at the top centre of the faceplate indicates the step number.

Power is supplied to the sequences using the Power toggle switch on the faceplate: power is on when the nearby LED is illuminated.

Here is the schematic:

Here is the layout for the Sequencer front panel (Powerpoint file here):


Based on the Intersil 8038. A general purpose oscillator covering the full audio range and offering sine, triangle and square wave adjustable outputs.



Useful Links

HV Nixie supply:


Schmitz VCO:


ICL8038 Waveform generator

LED Scope