Denial

chysn wrote: ↑Wed Oct 27, 2021 4:54 pm One issue is that the remains of the previous stem is broken off, and flush against the top. So that'll be interesting to try to get out of there.

You have to replace stems by unscrewing the back of the keyboard, they're a single piece.

srowe wrote: ↑Thu Oct 28, 2021 12:33 am

chysn wrote: ↑Wed Oct 27, 2021 4:54 pm One issue is that the remains of the previous stem is broken off, and flush against the top. So that'll be interesting to try to get out of there.

You have to replace stems by unscrewing the back of the keyboard, they're a single piece.

Oh, cool, thanks!

Huh. This is damn peculiar...
Right now, I kind of suspect a 6522, but I'm going to look into this a bit further. This is the same VIC-20 that I picked up and (mostly) got working last spring. Results from VIA 1 timer 1 were likewise only at the extremes.

And okay, right now the same program is producing credible results. The plot thickens.

Update: RND(X) seems to work when X > 0. PRG says that "X=0 generates a random number from a free running clock." I haven't looked into this yet, but I'll be that "free running clock" is VIA 1 Timer 1, which doesn't count down; it just alternates between 0000 and FFFF on this machine.

RND(0) uses VIA1 timers, which are also used for RS-232 timing. Can you test that on this VIC?

srowe wrote: ↑Sat Feb 12, 2022 9:19 am RND(0) uses VIA1 timers, which are also used for RS-232 timing. Can you test that on this VIC?

Hey! Yeah, I was just making an update as you were writing this. The VIA timer issue is something I noticed last week. But my reasoning at that time was, "If there's something wrong with a VIA timer, the failure would be more catastrophic." Apparently VIA 1 Timer 1 isn't super-critical. I don't do anything with RS-232, though.

The VIA might have been failing a year ago. Or I could have killed it when I was doing my early user port experiments. Either way, I think I first try a new 6522?

chysn wrote: ↑Sat Feb 12, 2022 10:00 am The VIA might have been failing a year ago. Or I could have killed it when I was doing my early user port experiments. Either way, I think I first try a new 6522?

If they're socketed you could try a replacement or just swap them around.

srowe wrote: ↑ If they're socketed you could try a replacement or just swap them around.

Unfortunately, while VIA#2 is socketed, VIA#1 is not. I’ve practiced a lot of chip removal on a dead C64, and I’m pretty confident with desoldering. But I do not look forward to the time it will take.

Obviously, I’ll install a socket while I’m at it.