Board 324003 cursor blinks too fast

[djc6]

I have a VIC-20 board revision 324003 that starts up, but the cursor blinks too fast. I have a second 324003 board missing VIC chip and CPU. I put the VIC chip and CPU from the board that blinks too fast into the one missing chips, and the cursor blinks at normal rate. So doesn't seem to be CPU or VIC.

I tried swapping UAB1 and UAB3 (both 6522's) over to the machine with the too fast cursor, didn't help. Tried character ROM too since it was socketed.

Any ideas what the problem could be? Here is a video:

https://i.imgur.com/p3TIMiL.mp4

[orion70]

Alt+W reverts from warp to normal speed (just joking )

[Mike]

Could be a short to GND on the IRQ board trace.

Does the VIC react to any keyboard input? If so, how fast is the repeat rate if you press and hold the space bar?

Do you own a multimeter and know how to use it?

[djc6]

Does the VIC react to any keyboard input? If so, how fast is the repeat rate if you press and hold the space bar?

Do you own a multimeter and know how to use it?

Yes, I can type on the keyboard fine. If I hold down the space bar, it takes about 3 seconds for the cursor to go from left side of screen to right side. I tried the other machine, the space bar repeat rate is the same.

I have a multimeter and can use it. Unfortunately, the bottom RF shield is soldered on, which is never fun to remove. Doesn't appear to have ever been removed. What is preferred method here? Heat up the joint with soldering iron, and bend each tab out of the way?

[Mike]

Unfortunately, the bottom RF shield is soldered on, which is never fun to remove. Doesn't appear to have ever been removed. What is preferred method here? Heat up the joint with soldering iron, and bend each tab out of the way?

For the moment, there shouldn't be any need to remove the entire RF box. Trying to do so will likely damage the mainboard and surrounding traces.

Edit: At first I thought you meant the RF box containing the VIC chip. Then I remembered some early mainboards have a shield covering the solder side of the entire mobo. That's okayish to remove but still shouldn't be necessary at this point.

Yes, I can type on the keyboard fine. If I hold down the space bar, it takes about 3 seconds for the cursor to go from left side of screen to right side. I tried the other machine, the space bar repeat rate is the same.

O.K. that rules out the IRQ issue for the moment. We might need the multimeter later.

Do you have an effective means to transfer files between VIC-20 and PC? Do SAVEs work (i.e. can you reload saved programs)?

If so, please run the following program (download) with LOAD"...",8 and RUN on your suspect VIC-20 and attach the 3 resulting files in a *.zip archive here.

Are you from PAL or NTSC country?

[djc6]

Do you have an effective means to transfer files between VIC-20 and PC? Do SAVEs work (i.e. can you reload saved programs)?

If so, please run the following program (download) with LOAD"...",8 and RUN on your suspect VIC-20 and attach the 3 resulting files in a *.zip archive here.

Are you from PAL or NTSC country?

Thanks for helping me out!

I have a Pi1541 that I ran your basic program from. When I run it, I hear some disk activity for a while (via the speaker on Pi1541) but then machine locks up after a minute or so - I can't even RUN/STOP & RESTORE. I'm still playing around with running your program - trying to figure out how saving works on the Pi1541.

I'm in NTSC country. I was curious if sound works on my 6560 VIC chip, so I ran this software:

https://www.uber-leet.com/Commodore_VIC ... Kruuv).zip

To see if it matches output from youtube video: https://www.youtube.com/watch?v=XNqZXHnqALg

The reason I mention this sound check tangent is that some interesting things happened on the screen when I try and load the program. Here is some odd behavior when I do a directory listing:

https://i.imgur.com/pzlJSDx.mp4

Here is more odd behavior when I run the program. The word "RETURN" is supposed to be in reverse characters (dark background; white lettering):

https://i.imgur.com/mbsbaxF.mp4

Thanks again!

[Mike]

The reason I mention this sound check tangent is that some interesting things happened on the screen when I try and load the program. Here is some odd behavior when I do a directory listing: [...] Here is more odd behavior when I run the program. The word "RETURN" is supposed to be in reverse characters (dark background; white lettering): [...]

All reverse chars seem to have a unstable bit 7, i.e. they change erratically between bit 7 on and off. That also explains the strange appearance of the cursor.

A defective RAM would likely stop the VIC-20 on start-up with cyan border and empty white display window.

With your multimeter and the VIC switched off, can you check continuity of VA10 between VIC Pin 31, UD7 (Char-ROM) Pin 19 and UE8 ('245) Pin 16 (all three ways)? VA10 is high when VIC accesses reverse chars from the character ROM.

It is possible, that UE8 is damaged and partially pulls down VA10. If it is a MOS part, replace it anyway (that also includes UD8 and UF8 as preventive measure). MOS logic chips are notorious for defects.

[djc6]

With your multimeter and the VIC switched off, can you check continuity of VA10 between VIC Pin 31, UD7 (Char-ROM) Pin 19 and UE8 ('245) Pin 16 (all three ways)? VA10 is high when VIC accesses reverse chars from the character ROM.

I have good continuity between those three pins on the respective chips.

It is possible, that UE8 is damaged and partially pulls down VA10. If it is a MOS part, replace it anyway (that also includes UD8 and UF8 as preventive measure). MOS logic chips are notorious for defects.

Mitsubishi part.

A defective RAM would likely stop the VIC-20 on start-up with cyan border and empty white display window.

I don't know if this is a useful test, but my only cartridge is "Jupiter Lander" (but it comes up as Super Lander" on title screen). It takes a few seconds once game appears for the lander to move across screen, but there are these vertical strips present that don't appear on the other machine:

https://i.imgur.com/SYEwNmk.mp4

Really appreciate your help!

[Mike]

I don't know if this is a useful test, but my only cartridge is "Jupiter Lander" (but it comes up as Super Lander" on title screen). It takes a few seconds once game appears for the lander to move across screen, but there are these vertical strips present that don't appear on the other machine: [...]

O.K. ... that's a dead giveaway: BD7 is (mostly) read as 0 by the VIC chip in those cases where it should be 1. That also corresponds well to the reverse character issues.

What I first suspected with VA10 now seems to be with BD7, something is pulling it down.

Does one of the RAM chips containing D4..D7 (these are UE2, UE3, UE4, UE5 and UE6) become exceedingly hot during operation?

What happens if you run this small program:

Code: Select all

1 POKE36879,238:GOTO1

Do you get a light-blue background within a blue border, an all blue screen or flashy stripes?

[djc6]

None of the RAM chips get hot, and when I run that program I get a light-blue background within a blue border. No flashing or all blue screen (I can still read the text). I was curious what you were up to and found this page:

https://www.atarimagazines.com/compute/ ... r_Tips.php

Seems light-blue background within a blue border expected.

[Mike]

I was curious what you were up to and found this page: [...]

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. Finally, replace all involved components one by one - the bottom shield can be removed in the way you had already described: heat up the metal lids, remove solder, bend them straight one by one, clean up the solder joints. No need later to put back the shield.

More easier though would be moving VIC and CPU over to the working mainboard and using the defective mainboard for spares.

[djc6]

If available, I'd check the signal on BD7 with a scope at the '245, VIC, VIAs, RAMs.

What are your thoughts on using something like a Elenco LP-560 logic probe to check these signals? I was thinking of picking one up just to learn how to use it.

More easier though would be moving VIC and CPU over to the working mainboard and using the defective mainboard for spares.

Yeah . Only benefit the defective board has is the power switch is easier to operate. The working board's power switch is very stiff. I can just swap the switch.

I did something stupid yesterday swapping VIC 6560 around between the boards. I had chip shifted to the left by one pin. So pins 1 and 40 on the socket had nothing plugged into them. Pins 20 and 21 on the chip were hanging off left side of socket. I turned machine on this way. Since pin 40 on the socket is +5V and nothing was plugged into this pin, I thought maybe I got lucky. But that meant Pin 19 on chip for audio output was connected to pin 20 ground on the socket. Do you think I harmed the VIC in any way by doing this? I noticed audio doesn't work on the chip - only 1 of 3 voices and the noise generator work. Not sure if the sound ever worked since I just got the machines.

Thanks again for your guidance!

[chysn]

If available, I'd check the signal on BD7 with a scope at the '245, VIC, VIAs, RAMs.

What are your thoughts on using something like a Elenco LP-560 logic probe to check these signals? I was thinking of picking one up just to learn how to use it.

Yes, the LP-560 is the very one I have, and it works great for checking pin states. I have an oscilloscope, but I prefer using the logic probe for checking signals, because of the audio cues. I connect the power clips to pins 1 and 2 of the cassette port.

[djc6]

Yes, the LP-560 is the very one I have, and it works great for checking pin states. I have an oscilloscope, but I prefer using the logic probe for checking signals, because of the audio cues. I connect the power clips to pins 1 and 2 of the cassette port.

chysn, having a hard time knowing when I should use TTL mode on the probe, or CMOS mode. Haven't tried it yet, still researching how to use it. Any tips on TTL vs CMOS? Depends on the chip? Example what would the 6560 use. Thanks!

[chysn]

Yes, the LP-560 is the very one I have, and it works great for checking pin states. I have an oscilloscope, but I prefer using the logic probe for checking signals, because of the audio cues. I connect the power clips to pins 1 and 2 of the cassette port.

chysn, having a hard time knowing when I should use TTL mode on the probe, or CMOS mode. Haven't tried it yet, still researching how to use it. Any tips on TTL vs CMOS? Depends on the chip? Example what would the 6560 use. Thanks!

I think the only difference is the voltage range that the probe considers “high” and “low.” If I recall, TTL is high at 5 volts and CMOS is 3 volts. Mine was set to CMOS last time I checked VIC-20 pins but it probably doesn’t matter. The other switch should be set to Pulse.