Amiga

An Amiga 1200 fighting me the whole way

On my repair list have been two motherboards from Retro32. An Amiga 600 and Amiga 1200. The Amiga 600 repair was simple, and not worth a blog post, but the Amiga 1200 has had so many weird issues, it is definitely worthy of talking about.

The Amiga 600

The Amiga 600’s main symptom was that the picture had ghosting and other quality effects. The motherboard had been recapped by a known person who has done many recaps (I’m not going to name and shame). I immediately figured the problem was likely with the decoupling capacitor, which stabilises the video power (C408). I flipped over the board to test it, and this is what I saw.

Disregarding the fact that there is lots of flux left on the board, the negative pin of the capacitor wasn’t soldered. As soon as I soldered this, the picture was great!

This board will be thoroughly cleaned before it is returned.

The Amiga 1200

The Amiga 1200, on the other hand, came to me recapped and there had been some repairs made to it.

Retro32 told me it was working, but basically died and no longer boots.

Diagnostics

I popped in DiagROM and booted it, the memory test went fine, and it got to the menu. But in the menu, I couldn’t select anything via the keyboard or mouse. That’s not great.

Booting again, this time with RS232 output, and there is something that indicates where the problem might be.

The right mouse buttons are stuck down. Now, the right mouse buttons are handled by Paula’s analogue side. So, the first thing I checked is that all the voltages for Paula are correct, the analogue ground hooked up, and that the buttons were producing the correct voltages to Paula. All was good on that front.

Paula also handles the interrupt handlers for the computer, my theory was that as soon as it hits the menu, the interrupts were getting pinned and that is why I couldn’t even move the mouse. The more obvious indicator would be that RS232 didn’t work at all either, but that was working. So, if Paula was bad, it was only partially bad.

I desoldered Paula and attempted to boot DiagROM without it. It can detect Paula is missing, it behaves a little differently, but is predictable if you are used to it, I was now able to control DiagROM, which indicates that I could be correct about Paula being the cause.

I then put a socket in its place. This required mounting C305 on its side so that the socket would fit, but it would make replacing Paula much easier. Finally, I popped a new Paula into the socket.

Yep, there are a couple of solder ball left behind from the paste used for the socket, I removed them after this photos.

Testing

With the new chip in, I was able to get much further in DiagROM. I was able to control it with the mouse and keyboard. The next thing to check is the IRQ handler.

Excellent! That is a great sign. The final two major functions of Paula are audio and floppy disk related. So, let’s switch to Kickstart ROMs and Amiga Test Kit in a Gotek.

It boots! Great. Now for the audio test. This also worked, but it didn’t sound quite right.

After switching things on and off in the audio test, I found out that the low pass filter was jammed on for the left channel, which would explain why it sounded odd.

I checked the voltages around the filter circuit, the _LED signal was working great, but Q321 appears to be completely jammed. This is the JFET, which turns on/off the filter for the left side. The voltage is correct on the left side of R326, but right side it is pinned to around 1v.

Time to replace the JFET…

I installed a new MMBF102 in place of the old one (the left one of the three transistors).

This fixed that problem. The audio filter worked, and the sound tests sounded normal again.

Parallel port

Whilst checking for the cause of the first issue, I happened to spot that the CP signal was disconnected for the parallel port. This trace was broken.

I’m not sure exactly where, I would assume the via under E301 got eaten by the corrosion from the capacitor next to it. I wired a fix for this under the motherboard.