Denial

Mike wrote: ↑Mon Jan 03, 2022 7:55 am I wanted to check that BD7 (bit 7 of the VIC side of the data bus) reliably appears at VIC when sent from the 6502 over the '245.

At this stage, I'd take the entire mainboard trace of BD7 under extreme scrutiny - continuity, shorts to other traces, any cold solder joints, checking all socket pins. If available, I'd check the signal on BD7 with a scope at the '245, VIC, VIAs, RAMs.

Mike, your remote prognosis was correct! I tried the piggyback method with a RAM chip from the working board, and sure enough UE4 was bad. When I piggybacked the chip Jupiter Lander started working much better - the black lines on the screen were mostly filled in with red. So I socketed the board and put the donor chip in, and now jupiter lander works 100%.

However, I still get some weird behavior when reverse characters appear on the screen - for example directory listing.

This directory listing still looks the same: https://imgur.com/pzlJSDx

I went back to original VA10 hypothesis and put a new chip in UE8 - no change. I've tried the VIC 6560 and character ROM in the other board.

Possibly another bad RAM chip? Can I replace a 300ns RAM chip with 200ns? The 200ns 2114s much much cheaper and easier to find.

djc6 wrote: ↑Thu Jan 13, 2022 8:21 am Can I replace a 300ns RAM chip with 200ns? The 200ns 2114s much much cheaper and easier to find.

Yes, you can.


It appear as VA10 remain float (not droved) during the VIC cycle....
check the continuity from the VIC's pin 31 to the char ROM UD7 pin12, but remember: the probes have to be placed on the chips pins where them exit from the Chip plastic body (also sockhets contat will be checked...)


This test can be done for each VA line ....

Looks like VA10 also goes to a 74LS138 at UC4 on Pin 1 - I also have continuity with rest of VA10 there. This is a chip I haven't tried swapping.

Swapping 74LS138 at UC4 didn't help. I think its a RAM issue. I found this post about changing the location of screen memory:

http://sleepingelephant.com/~sleeping/i ... 0f98#p5511

I tried different locations from $1200 to $1E00 and I found that if I start screen memory at $1800 or $1C00 I get no cursor when I type, and I can't generate reverse characters

POKE648,24:SYS58648
POKE648,28:SYS58648

Now to figure out what RAM chip(s) these correspond to! I already found UE4 was bad (fixed Jupiter Lander issue).

The VIC-20 is mostly TTL logic levels, so that's what I'd set the logic probe to. BTW, I wouldn't go indiscriminately poking around with it at random - I don't know how happy it'd be with 9VAC!

Everything is working now!! But I don't know why

I have two VIC-20's, board A and board B. When I bought board B and plugged it in, there was no picture. I get a picture if I swap the VIC 6560-101 from board A into board B. However, board B has the issue with reverse characters and cursor that blinks too fast that I've been trying to diagnose in this thread. Board A does not exhibit these problems. Both are revision 324003.

I bought a MOS 6560-101 off Aliexpress - dated 47th week of 1983. When I put it into board B, all my problems go away.

Any guesses why the other 6560-101 only work properly in board A? The iffy VIC is dated 43rd week of 1980 and the VIC chip from aliexpress - that functions 100% - is a lot newer, 47th week of 1983.

The iffy VIC chip from 1980 has a blue sticker on it - don't know significance of that.

tl;dr my cursor issue and reverse character issue is resolved by getting another VIC 6560-101 from aliexpress.