Case Studies in Retro Computer Repair (#1): Diagnosing and Fixing a Commodore 64 with a Bad MOS 7708 Multiplexer Resulting in a Black Screen

Disclosure: Note that as an eBay Partner Network Affiliate, I may earn from qualifying purchases.

Introduction

I purchased the following Commodore 64 on eBay for $90. The seller said it was not working and had tried to repair it with a GAL PLA. He further said it booted to a black screen which means the VIC-II was likely good. I liked this auction because he showed a picture of the motherboard. I could see that many of the chips including the SID were not socketed suggesting they had not been replaced with faulty ones. Also, the case and board looked really clean and in great condition.

Case Study

I first confirmed that the unit did indeed book to a black screen. I then tried the Dead Test to see if I could get some clues. The Dead Test would not boot.

I then checked the voltages as shown below. The 5V and 12V lines looked ok. Further, none of the chips on the board were hot to the touch which would suggest a short. I also checked the front and back of the board for any damage or broken traces. Nothing.

I then used the oscilloscope (I bought a Siglent) to probe all the pins on the 6510 CPU. I always start there and make sure all the signals look like they should. Indeed, I could not find any evidence of a bad signal.

I then examined the RAM and did not see any Micron Technologies (MT) RAM that is unreliable.

Next, I examined all the logic chips and multiplexers for MOS-branding that is also unreliable and prone to failure. I found two MOS 7708 multiplexers that I desoldered and tested using my chip programmer. Both chips tested fine multiple times. Since I had the chips out, I installed sockets and replaced them with working counterpart 74LS257s. This fixed the computer an reaffirms the faulty nature of MOS support chips and the unreliability of the chip programmer at times.

I finished up this board by adding an RF replacement board and putting heatsinks on the VIC-II that gets pretty hot.

Comments

This repair ended up being a common fault with MOS-branded multiplexers that are prone to failure. These were likely interfering with the RAM addressing thus not allowing the Dead Test to run. It is always good practice to remove MOS logic chips and multiplexers as well as any MT RAM present.

Note that the previous owner installed the GAL PLA shown in the above photo (top left) in a repair attempt.