New! Roland TD-20 Backup NG error hardware fix

My drum kit worked fine for 20 years. At some point, it started requiring a battery replacement. However, after the 3V CR2032 battery replacement, it refused to save data to memory and returned an "backup ng" error message.

Replacing the firmware didn't help, and a search of the forums led to the conclusion that this was a hardware problem and required replacing a part, although they didn't specify which part needed replacing. I had to figure it out by myself.

Begin by removing the screws and top case. Then pull out the connectors to release the motherboard.

I had a schematic for this device (I'll attach a copy to this article). IC3 is responsible for saving temporary data. However, the manual didn't provide a solution for this problem. I assumed IC3 was faulty and replaced it. The chip installed was a CY62146DV30LL, but I replaced it with a similar CY62146EV30LL. You don't have to do this.

Before that, I measured the voltage on capacitor C6 and found a strong voltage drop there. The 3V from the battery was reduced to 1V, so I assumed the microcircuit had failed due to its power supply. I couldn't have guessed that the fault would be located elsewhere entirely.

The 20K resistance didn't disappear after replacing the chip. It couldn't operate in that mode. I examined the circuit to see if there were other places this trace could connect to.

According to the schematic, this power line also connected to capacitor C435, located on the back of the board. There were obvious signs of corrosion underneath it, which I hadn't noticed initially. Replacing it didn't solve the problem either.

I was about to give up, but I decided to check the last trace, which led to the auxiliary processor IC2. I thought the processor had failed and the setup couldn't be repaired. Just for fun, I cut off the trace, but it turned out the resistance wasn't on the processor side. You don't need to cut the path in that case.

I also started to go around all the nodes of this power line and measure the resistance in order to find out at what point it would be minimal, using Kirchhoff's law. If the resistance at the next point was higher than at the previous one, then we were moving away from the short circuit.

Following this path led me to the RA14 47R resistor assembly located near IC2. The resistance was nearly 10K to the GND on this assembly.

It turned out that the CS line also runs through this trace, which is why the data wasn't being saved. It was in this component that the 20K resistance was discovered.

Indeed, the chances of the problem being in this component were extremely slim. Apparently, it had degraded over time due to the constant power supply on its line.

I replaced this assembly with an assembly of 8x33R resistors, which I found on a similar board.

After that, I measured the voltage on the chip. It rose to the required level (the drop was no more than 0.3 volts from the input).

The "backup ng" error has been fixed.

My board later stopped detecting the TDX-20 module. Also, unfortunately, after some time of use, the display on my device failed due to age. I had to order a new one; it's still in transit.

UPDATE From 09.12.2025: to fix the "TDX-20 module not found" I had to reflash the older (1.10) firmware version instead of version 2. It's unknown, why the board was dead, maybe I corrupted the flash earlier.

The high-cost LCD display from China is working well. But the display area is smaller than original, also it has a different background color. But the device is still usable at current moment.