Vintage Synth Lab

12/15/2025

This Fender Blues Deluxe has been previously worked on. While previous service is not unusual, the amount of work performed, including rescue of PCB traces to the degree that this amp has incurred, was absolutely unusual.

I suspect the ribbon cables between the PCB‘s got destroyed during that precious service attempt, as they were soldered and not done very well. Usually these cables don’t need to be soldered if you’re careful and not removing the board a bunch. But in this case, it just wasn’t worth fighting fraying wires. So here we are again upgrading cables to solid Molex pluggable type. In this case, the main board side has the connector, whereas the cable is permanently soldered to the tube board.

12/05/2025

Fig 1. The Roland Juno-106 power supply. This J106 came to me with a flavor of intermittent behavior that I had to put this one aside for several days at a time. The synth would often lock up, not responding to any of the panel controls. Some LEDs on and also frozen in their state of logic. But this frozen state would only last a several seconds, or a minute if you are lucky. However, last night I finally solved the mystery.

Fig 2. Power Supply turn-on, rise-time, ringing, shown on the oscilloscope. These look okay, all settling within 60 ms or so.

Fig 3. Zoomed out view of the HSE clock (crystal) for the main CPU. This just shows the clock turning on. The next image (Fig 4) shows the clock itself.

Fig 4. The main clock at the clock input in of the CPU. This is a good crystal oscillator clock signal.

Next, here is where I found the issue: while the buttons and LEDs were in a frozen state, I noticed the switch matrix was not cycling / polling. All states were frozen HIGH or LOW. So then I traced back to the peripheral interface ICs between the CPU and the switch matrix. Those also showed frozen signals. So then I verified the same for the GPIOs of the CPU. Next, I looked at the RESET pin of the CPU. And here was the big hint: RESET (active LOW), was being held LOW. The first place I looked was making sure the pull-up resistor on the CPU board was okay, including the + supply side.

Fig 5. The schematic of the main CPU and its RESET net/connection. 2 RESET connections to the Module Board, and 1 to the Power Supply board.

Next, I followed the RESET signal to the PS board (not the Module board)... See Fig 6.

Fig 6. On the PS Board, we have what is also known as an old fashion "Power Good" signal generator.... or "POR" (Power On Reset). Once C18 charges up AND the +5V regulator rises above about +3V, Tr4 turns on after D4 zener conducts at 6.8V minimum. Tr4 collector should go LOW (0 V). This would turn Tr3 "off", and Tr3 collector goes HIGH (+5V). This then sets RESET to HIGH, which allows the CPU to operate.

The problem: broken solder at several points on the PS Board!

11/29/2025

The Linn Electronics “Linndrum” LM2

ORIGINAL ISSUES: Some of the drum sounds don’t sound right, such as RIDEs, CABASSAs.

The issue with the RIDE sound was resolved by reseated all of the ICs that are socketed. But the CABASSA sound remains problematic.

Experiment 1: Since the Decoding, Strobing, and Data Lines for the BASS, TAMBOURINE, and CABASA are all shared, this means that it would be fair to assume that we can swap any of the chips of these voices for the purposes of troubleshooting / debugging.

A quick swap of the TAMB and CBSA sound PROMs reveals that the problem stayed with the circuit, not the PROM. This means that both TAMB and CBSA chips are fine. The issue needs further investigation perhaps starting with U26.

Since the addresses for each PROM memory location of the drum sound IC is accessed by a ripple counter, we can examine the signals there. But first, from the service manual: Sequential Addressing of PROMs: “Bass, Tambourine, and Cabasa: use a 13 bit ripple counter to address the PROMs. On the second pulse of the drum strobe, the cout nets are enabled and begin sequentially counting through the PROM’s addresses. When he count reaches 4096, the 13th bit goes high, clearing the latches and disabling the counters.”

ACTION: Compare known drum sound circuit strobe counter with Cabasa circuit:
So if we compare the 13th bit for the BASS vs the CBSA circuits, we find that the BASS drum sound process does invoke this pulse on U51.11(A2), precisely at the end of the sample extraction sequencing. However, when triggering a CABASA sound (which continues to repeat in a rapid short sample sound artifact) we probe U45.11(A2) and we never see this pulse… So now what?…

OBSERVATION: If we look back at U46, which is the LSB side of the ripple counter (A0 through A7), we see our ripple counter waveform (or clocks, and respective frequencies) as expected. However, when we look back to U45, which is the MSB side of the ripple counter (A8 through A11), we find a problem:
U45 (LS393)
U45.1 (CK1) = good
U45.3 (A1) = good
U45.4 (B1) = missing

11/25/2025

This is a Hoffman DIY build Princeton Reverb amp which needed an output transformer. The output transformer is a perfect choice for this amp

11/24/2025

11/22/2025

Part three of three videos demonstrating a repaired 1975 Yamaha CS-50 analog synthesizer.

11/17/2025

Part two of three videos demonstrating a repaired 1975 Yamaha CS-50 analog synthesizer.

11/16/2025

1975 Yamaha CS-50 analog synthesizer recently serviced!

This CS50 came to me not working and with many problems, including a damaged power supply capacitor, two cracked circuit boards, many broken wires, oxidized calibration trim-pots, and 2 defective VCA chips. It had been sitting in someone else’s shop for nearly 2 years. I was able to turn it around in 6 weeks.

11/13/2025

Digging deep into a Roland Jupiter-6. This one has a lot service ahead, including many fader replacements, switches, and 4 dead voices.

11/02/2025

Sometimes you need to make a modification that is necessary. This mid 1990s Fender BXR-100 bass amplifier came in for the typical upgrade to the input jacks, soldering, and cleaning. However, when reassembling the unit, the terrible and inappropriate cabling scheme between the two boards broke. I worked on these at the old Sunn/Fender factory, then in the field under warranty, then out of warranty for years to follow. I eventually acquired some tools to build custom cables. And I happen to have five pen, headers and connectors for a keyed type Molex “Spox” line connector. The most difficult part is stripping the wires and crimping the pins just right, but I am a professional at that as I have done tens of thousands of wire crimps for a former start up I worked for—in a different lifetime. In this case, I soldered the wires to the power amp board and put the connector on the preamp board since the preamp board most commonly gets removed on these units for service.