VIC-20 Repair?

[areeve]

I broke out my VIC-20 for the first time as I work through testing out my collection and boot up looks odd to me (picture attached from my phone camera). As you can see the characters are interspersed with blue and red ones and they should be all the same color (preferably blue characters on a white background). It's seemingly random as to what color a given character will be (e.g. whether a character is blue or red varies every time I start the system). I was hoping someone with more VIC-20 knowledge that I have might be able to suggest what might be wrong. My inclination is video memory, but it'd be nice to get a second (more knowledgeable) opinion.

 

- Thanks, Alan

 

P.S. It was having problems in that sometimes it wouldn't boot (just a black screen), but after a recap that seems to be resolved.

[hofster]

It looks like a colour RAM issue. I'm no VIC-20 expert but it looks like one of the 2114 SRAM chips is dedicated for color so that one is a suspect. Could also be a 4066 logic chip from what I understand. The names of the ICs might vary depending on what board version you have.
You could check Ray Carlsen's troubleshooting tips. Especially vic20.txt over at https://portcommodore.com/rcarlsen/cbm/vic20/

Skickat från min SM-G996B via Tapatalk

[eightbit]

And, if you cannot figure this out yourself I would highly recommend just contacting Ray Carlsen and ask him if he can repair it (which I am sure he can). I can't tell you how many machines he has serviced for me over the years. Highly recommended.

[areeve]

Thanks hofster for guiding me to Ray's guide. I ended up removing the CD4066 and 74LS04 and adding sockets to make future maintenance easier. The 74LS04 tested as good so I kept that, but the CD4066 was defective. After doing that and verifying by swapping the bad one and a good one and seeing the difference I now had another problem... the characters were scrambled albeit in a mapped manner meaning every space was now a @, and * was J and so forth with some also being correct such as B was still B. Before all the characters mapped correctly... just the colors were off. I ended up getting a Penultimate+ cartridge for its Diagnostic tools (as I had planned to get one anyways once I had a working VIC 20) and then the machine starting booting to just a black screen... and then it magically started working. The Penultimate+ cartridge Diagnostic software verified the memory and ROMs as all fully functional. I'm a bit puzzled about all of this and it was consistent for many power cycles so how it just cleaned itself up is a ?. I can only wonder if when I was testing and I had the board just sitting in the plastic case vs on the paper/metal shield if it was in contact with something, but that seems farfetched. Anyhow, it's all back together now and fully functional. Yay!

 

On to the next project...

[hofster]

Thanks hofster for guiding me to Ray's guide. I ended up removing the CD4066 and 74LS04 and adding sockets to make future maintenance easier. The 74LS04 tested as good so I kept that, but the CD4066 was defective. After doing that and verifying by swapping the bad one and a good one and seeing the difference I now had another problem... the characters were scrambled albeit in a mapped manner meaning every space was now a @, and * was J and so forth with some also being correct such as B was still B. Before all the characters mapped correctly... just the colors were off. I ended up getting a Penultimate+ cartridge for its Diagnostic tools (as I had planned to get one anyways once I had a working VIC 20) and then the machine starting booting to just a black screen... and then it magically started working. The Penultimate+ cartridge Diagnostic software verified the memory and ROMs as all fully functional. I'm a bit puzzled about all of this and it was consistent for many power cycles so how it just cleaned itself up is a ?. I can only wonder if when I was testing and I had the board just sitting in the plastic case vs on the paper/metal shield if it was in contact with something, but that seems farfetched. Anyhow, it's all back together now and fully functional. Yay!
 
On to the next project...

Good job!

The character mapping problem that solved itself sounds like bad contact somewhere. Maybe a ROM socket or something. Moving the board around enough could mobe the chip just a little in the socket.
When some characters being replaced by different characters while others remain correct can sometimes be explained by bit errors on data pins or address pins somewhere. Like if address bit 0 were missing, all odd addresses, would be off by 1 while all even addresses would be good.
I'm not sure what the correlation is between the characters you mentioned but if I had that problem I would reseat socketed chips and spray the sockets with contact cleaner and I would have started with the ROMs.

The Penultimate cartridge is an excellent product. Good choice!