How I rescued a Commodore 128

I have got a Commodore 128 for free, when buying some Polski Fiat 126 parts some times before. The C=128 was waiting in his ashes, to rise, since it couldn't be turned on, and got some other issues as well.

When I pulled it out last weekend from the storage, to take a closer look, I instantly realized that the main problem was the power switch. Took one of the spare part C=64 as well, and soldered out the switch and installed it on the C=128. Checked, connections are OK, power is OK. However, there was a small discontinuity on the ground connection for the power LED. Soldering done, and shines as before.

Checking luminance on VIC

Since I didn't have an analogue TV at hand, I used my laptop with AverMedia Hybrid USB TV, to capture the display of the machine. First, I connected it to the RF antenna slot, to try to tune in the C=128 boot up screen. I only got screen pictures for some prompt times, mainly static pictures buffered by the TV card. And it was all black & white. I tried to fine tune, without chance, also tried to fine tune the RF modulator on the Commodore. The best I could get with PAL settings, is a black & white picture scrolling upwards, with great flickers on every second row. Also connected the video output to the composite in of the TV card, with the result of somewhat nice picture... in black & white.

C=128 main board

I quickly opened up the troubleshooting guide (which is for NTSC versions anyways), and followed the steps to narrow down the problem. Fortunately I have an old analogue oscilloscope, with 20 MHz measurement frequency. I quickly checked the different clock lines on the motherboard, and measured their frequencies with a rough estimation, since there is no such things on this 'scope as freq counter. The pin 16 of the VIC chip gives the luminance encoding, and pin 17 the chroma. Both were OK, however, the chroma seemed somewhat low on amplitude.

Donor C=64

So either the chip, or the M1 (the RF modulator) was faulty. Anyways, it seemed unreasonable, that the VIC chip give good signal with bad amplitude, so I turned to the RF. The modulator also contains an amplifier for the luminance and chroma output for the video connector. I measured with 'scope, and got a pretty distorted, low level output on it. So, the problematic part should be the RF modulator's chroma amplifier.

C=128 & C=64 RF modulator

I quickly soldered out the modulator from the C=64 in under 2.5 hours (since practically it is the same as the C=128 has). The procedure was pretty frustrating, not having the right tools to do it. What I missed for this action, is some kind of an IC pin heater head for my soldering-iron, and a solder sucker pump. Anyways, in 4 hours, the "new" RF modulator was in place in the C=128. Check with the oscilloscope: higher amplitudes on VIC chip, less distorted and higher level chroma on the video connection.

When I connected to the TV card again, with composite, from time to time, colors appeared and disappeared, wasn't usable. On the forums I quickly searched through, somebody mentioned that the frequency of the PAL oscillator crystal is very important, and should be very accurate. I didn't have a chance to measure frequency as accurate as 8 digits on an analogue oscilloscope, unless the 'scope can do XY plot option, and have an accurate freq generator. I lacked the later one, so I went blind. Exchanged the xtal and the variable capacitor from the C=64, and fortunately enough, it did the trick!  :)


Now I have an (other, also) fully functional Commodore 128, in nice visual condition. Later, I will try to test it also in 80 column mode, when I have a chance.

Also, I have a newcomer, an even older, Commodore CBM 720 to be restored. But this will be an other story.