Tag: repair

SNES capacitor replacement

Measuring SNES capacitors

I had intermittent problems powering on the Super Nintendo, it would just black screen screen sometimes, turns out the electrolytic capacitors were bad. The yellow wires and PIC chip as seen on board is my SuperCIC install giving NTSC 60Hz and PAL 50Hz switchable using reset button, that modification is not related to this repair. There were no visible leaks from the electrolytic capacitors but several of them measured bad using the ESR meter so it was high time to recap.

SNES PCB and capacitor kit

I ordered a Super Nintendo PAL Capacitor Kit from Retro Upgrades via eBay, they shipped fast and also included a new voltage regulator which was nice. There are some good capacitors on the board, but I replace all of them since those are likely to go bad soon as well.

Removing capacitor from SNES main board

I like to use the “push down and turn” method using pliers to remove SMD type capacitors, some prefer using SMD rework station to desolder with hot air. In my experience bad capacitors sometimes explode when using hot air which can damage traces and be dangerous due to metal shrapnel flying around, so that’s the main reason I prefer using pliers. Besides, leaky caps sometimes smells like raw fish, especially when heated up (and exploding), so not recommended.

Desoldering leg remnants from bad capacitors

After removing all capacitors by turning in one direction while pushing down towards the board until the capacitor is removed it’s time to get rid of the broken off legs still soldered on the pads.

Capacitors desoldered, cleaned up the board

After removing the leg remnants with the soldering iron I add some flux and a touch of solder for each pad, and then clean up the board with IPA (Isopropyl Alcohol) to remove flux residue using cotton swabs (cotton buds) to make it clean before soldering the new capacitors in place.

New capacitors on the board

The new capacitors are installed, along with the voltage regulator. I added an extra 470µF 16V capacitor hooked up to ground and output of the voltage regulator to fix the potential vertical line issue, not sure if it really was an issue on this console but cleaning up the power rail voltage is probably good anyway.

SNES PAL (with NTSC using SuperCIC)
Pro Fighter X Turbo copier screenshot
Screenshot from the game R-Type

Time to test, moment of truth, seems to work.

Amiga 1200 chip RAM repair

Another adventure with my Amiga 1200, recently when booting I was met with a green solid color boot screen instead of loading AmigaOS 3.1 (workbench) as expected. According to boot error color codes that’s most likely bad chip RAM (the first 2MB of RAM shared with custom chips). I burned Diagrom v1.3 on EPROMs using the MiniPro TL866A programmer and installed those in the motherboard ROM sockets. Diagrom outputs diagnostics via serial port (9600 baud 8n1) on the A1200 which I connected to the COM port on my PC running Linux, the output from Diagrom confirmed that the chip RAM was bad.

The Amiga 1200 motherboard

The chip RAM consists of the four chips located close to the center (upper) part of the motherboard, the two ROM sockets are below. Here’s a close up photo of the area with the RAM chips.

Amiga 1200 chip RAM

Amazingly I managed to find the same RAM chips KM416C256BJ-7 sold on eBay, so bought five of them which arrived a week later. These chips are in a SOJ package, directly soldered onto the motherboard, so I used my SMD rework station to desolder the suspected bad chip using hot air, then cleaned up, and added solder.

One chip removed

I soldered in a new chip, but this was apparently the wrong chip, RAM test still failed, but instead of switching back I just replaced one by one until three chips were replaced and I ended up with address line errors in the Diagrom output. This was turning into a nightmare, and I started to doubt the “new” chips were working.

After going over the schematics again I figured out a way to cross check by measuring zero Ohm (connection) on the data lines and address lines between RAM chips using a multimeter, this made me discover several bad solder joints so had to apply more flux and reheat the chips as needed several times before the solder joints were solid. Admittedly I’m a rookie with the SMD rework station (hot air), much prefer the soldering iron, but things started to improve after upping the temp to 380C.

The Diagrom chip RAM test finally went through OK, so I happily assembled the A1200, put the original ROMs back and AmigaOS (workbench) booted fine, but only 1MB of chip RAM was detected. Frustrated, after looking at the schematics again I noticed that the A1200 3.1 ROM will boot (pass tests) even if only 1MB chip RAM is detected (U16 and U17 required, U18 and U19 optional).

I disassembled the A1200 once more and started measuring the data and address lines on the motherboard, all seemed OK. After some head scratching I decided to measure all pins (not just address and data lines), and found out that on U19, one of the RAM chips, the UCAS line (“Upper Column Address Strobe”, pin 28) was not connecting to the board. After soldering pin 28 on U19 manually using my iron the full 2MB chip RAM was detected, and passing all tests. The repair was finally over.