Apple IIGS graphic glitches

[lowlytech]

I apologize in advance as I have been having more than a few issues with a newly acquired iigs.

I have graphical glitches say in every 1 in 15 boots.  I have ran the IIGS diagnostics and it shows no errors.  I have a GGLABS 4MB card and an apple 1MB card and this anomaly will happen with either card.  There are no extra boards installed, and there is a fresh PRAM battery.   

When the issue occurs, the text prodos screens look normal, it is only when switching into the graphical mode.  When it does occur, turning the machine off and turning it back on seems to correct it.  It won't self correct until you power cycle the machine.  That being said if it didn't boot with corruption, so far it has never appeared until another power cycle.

Does this issue look like one of the RAM chips on board are flaky or something else?  

 

[cruff]

Looks like an addressing error where it is drawing a column too early in the scan, pulling the data from a later column.

 

[lowlytech]

Is there a somewhat finer approach to pinpointing what component I need to be looking at on the mainboard to fix this?  I will pull the CPU from the socket and reseat it, although I have pushed firmly on all socketed IC's but haven't pulled any.  Assuming that doesn't help, is there an IC I need to be looking at?  If it isn't SMT then I can replace it, however if whatever is responsible for the addressing lines is SMT, then I am gonna have to get a replacement motherboard as SMT and I don't get along.  The board in question is a 607-0173-A

 

[cruff]

I can't tell exactly, but it looks like the offset of the dislocated pixels is maybe 8?

 

[AwkwardPotato]

Yeah, it looks like the errors are 8 pixels apart from each other. The IIGS stores 4 pixels to a byte in this mode though and it doesn't look like entire groups of 4 pixels are wrong, so it might not be an addressing problem. You could try replacing the Standard RAM, and also reseating the VGC. If you do reseat the VGC, don't try prying it out of the socket with a screwdriver as the sockets are very fragile and will crack. You can buy PLCC extractors on Amazon/eBay for this purpose.

 

[Gorgonops]

I know at least some early units have known faulty VGCs, but obviously that's not a good diagnosis given the only easy source for one is another motherboard.

 

[cruff]

You could try running the RAM diagnostics from the diagnostics disk to see if that turns up anything about banks $E0 and $E1.

 

[lowlytech]

Cruff,  I will run the GS diagnostics again, but I left them running overnight with no issues found.  Maybe when I notice the glitches I will try to run them then without the power cycle.  

Gorgonops, Now you have me worried about the VGC.  Would the fact that it only happens in graphic modes indicate a bad VGC?  When it is booting up in a CLI/Prodos the graphic glitches are not present.   

If it is memory, where is the best place to get replacements.  Just wanted to make sure I got a compatible set of chips due to the fact that installing a socket looks to not be an option due to the tight fit in the case. 

 

[cruff]

What happens if you switch it into text mode by selecting the text control panel (command-control-escape) from the glitched graphics mode, then exit from it back to graphics mode?  Do you have a way to cool off the VGC when it is glitching to see if it is temperature related?

 

[AwkwardPotato]

The IIGS's standard RAM is comprised of 4x 150ns TMS4464 ICs. Compatible parts can be found on eBay, just make sure the number after the hyphen (which indicates the speed) is less than or equal to 15.

 

[Gorgonops]

Would the fact that it only happens in graphic modes indicate a bad VGC?  When it is booting up in a CLI/Prodos the graphic glitches are not present. 

When the IIgs runs in text mode (or the Apple IIe compatible graphics modes) the video is generated by the Mega II chip, not the VGC, so if you're only seeing corruption in the IIgs-specific modes that's probably not a good sign. (Although I did look it up, and symptoms for that known issue doesn't seem to map to what you're seeing.) That corruption is... weird, though, in how it looks like chunks of the screen are offset, almost like the address bus had a stuck/inverted bit that caused horizontal rows to be either written to (by the CPU) or read from (by the VGC) the wrong place. I don't have much confidence in suggesting it but replacing the RAM might be worth a "shotgun and pray" attempt.

The symptoms vaguely remind me of an intermittent address bus glitch in one of my Commodore PETs that I have yet to figure out.

 

[lowlytech]

Well I am knocking on yellowed plastic here, but since reseating the CPU I have power cycled this poor GS about 50 times and haven't once had the glitches.  Now with my luck the minute this post goes up it will glitch, but right now it is behaving. 

Thanks for all the input and info on this issue i was experiencing.   I will get the diagnostics to run in a loop overnight and see if anything interesting happens.   Are the best diagnostics the actual Apple GS ones I assume?  Didn't know if there was a better more detailed option out there I wasn't aware of.

Thanks again..

 

[lowlytech]

Still glitch free, maybe it was just a pin on the cpu not making good contact.  Diagnostics ran for two days in a loop, no errors in that time.