REPAIR GUIDE: Commodore 64 with a common problem

Recently a friend of the museum sent us his breadbin Commodore 64 that he acquired online with complains that there is garbage on the screen and there is no cursor. As much as we like restoring vintage computers and dealing with newer problems that enrich our knowledge base, we have never actually repaired a breadbin 64. After doing quick online research, we noticed that this is a common problem with the C64. Some solutions included wrongly placed character ROM, faulty CIA or PLA and the common faulty MOS chips. One thing was sure, different machines with the same problem, had different repair procedures. We decided to document our repair in this blogpost to be helpful to all of fellow collectors out there.

Our repair actually consisted of replacing two faulty 4264 RAM chips and some faulty MOS chips. Will this be the solution if you C64 has the same problem? Maybe yes of maybe half way. Read our post through the end as we document chronologically the repair.

One thing very important is the type of C64 you have, because there are different board revisions with different component layout. Our board has an assy number: 250425, which is a rev. A board with part number: 251470-01. The schematic that we used for this board is marked number: 251469 from the Commodore 64 repair manual, located on page 30. The service manual can be downloaded here. For diagnostics we used a dead test cartridge, which was helpful in guiding us to where the problem lied. Also, a RAM tester and TL866 programmer which has the capability to test logic ICs. We also measured some voltages on the board with a multimeter, but everything was in spec. We went through the logic with a scope, to see if the basic things like clock and reset were present. The zone we will focus here is memory and memory management circuitry.

Let’s first look at what is showing on screen. It is garbage all right, but we can see the garbage actually consists of known characters. That means that probably our character ROM is working.

Time to hook up the dead test and see what show on the screen. Immediately we see blinking on the screen. There are two blinks, then a pause and then it repeats. Time to see what that means by looking at the reference table for the dead test.

We can see those two blinks on rev. A/B boards means that there is a problem with U24, which is a 4264 RAM chip, which after removal tested bad on the tester.

Quickly we installed a 16pin socket and a new RAM chip, but the problem still was the same. This time the dead test did not blink, it went to a partially loaded condition. We can see the red lines of the dead test table and some of the white background.

After a quick head scratching, we decided to desolder all of the remaining 4264 RAM chips, so that we can test them and to install some sockets on the board for easy future repair. During that procedure we found out another bad RAM chip that was not detected by the dead test. Now all of the 4264 chips are socketed, the two bad ones are replaced and the problem is still the same…

Looking at the schematic we can see two 74LS257 multiplexers, that deal with the memory. On our board the chips were located on U25 and U13. The chips are MOS 7708, which is an analog of the LS257.

Knowing that MOS chips tend to be defective, we carefully desoldered both of them and BINGO, one was testing bad on the tester.

The problem was still the same, partially loading the dead test…. Time to go back to the schematic. There were three possibly bad ICs. One was U26 (74LS373), which tested OK. The other one was U27 (74LS04), also tested OK. The last one was U14 (74LS258), which was our eureka moment since it was the last remain MOS (7709) chip. That tested bad, we replaced it.

Now the moment of truth. Will the C64 finally work, since we replace most of the chips in the memory section? The answer is YES! the dead test finally loaded and showed OK on all the components.

The C64 also worked without the dead test and the keyboard was working. We now had a moving cursor.

Quick conclusion of the whole repair journey: Will this repair procedure help you with your C64 if it has the same problem? Maybe yes and maybe no. It won’t be a waste of time to tackle it like us by testing individual chips with a logic tester, this eliminates the guessing game. To end it here, just test those MOS chips, because they are known to be defective over time. 🙂