PET 4032 motherboard problem

[billdeg]

First of all, I realize that there is more testing I can do, but for laziness reasons, does anyone recognize this symptom...pictured below of what chip is bad here? I have replaced the socketed chips, none helped. When I power on my PET 4032 I see this:

I dred replacing unsocketed chips without a plan. Any ideas which are the likely culpret? The power supply is good, the processor and all the socketed chips are good. I suspect RAM, but there have been times when RAM was not the issue, and there are a lot of chips to desolder to find out I was wrong.

Thanks

Bill

 

[cosam]

Definitely looks like a memory issue, whether it be RAM, video RAM or even ROM. If by "replace" you mean you actually replaced the socketed chips with known good ones, as opposed to just reseating them, you can of course eliminate ROM.

Any of the RAM (or video RAM, 2114s near the middle of the board, IIRC) running overly hot or cold would be the first suspects. If that doesn't turn anything up, try piggybacking some good RAM over the soldered in chips. If you're still none the wiser, it's time to break out the logic probe or scope and figure out exactly what's wrong.

 

[gsteemso]

Have you changed the capacitors? My 4032 needs it done and displays (a much smaller degree of) garbage characters until it warms up for 20 minutes or so, after which it works OK.

 

[tezza]

Gosh Bill, yes that could be RAM, or ROM, or a dodgy capacitor. Unfortuntely it's a pretty generic symptom unless that dark horizontal line in the centre means anything. Or is that just the camera catching the strobe?

Tez

 

[billdeg]

I agree with all...that's why I moved on to a different project. What happens in these situations is you get inspired one morning, get the solder sucker out, and spend the day replacing things.

I think first however I will check the voltages to all of the chips and components vs. the schematic. There may be an obvious short that I can detect.

It may not be RAM because I recall from experience that on an older PET with socketed RAM we could still boot a computer with some bad RAM in any position except the first few slots, the system simply reports less memory.

So, I will id which are the RAM that hold the ROMS and the bottom of memory (0000-1000) and replace them first. I don't think it's video RAM, what I believe we're seeing on the screen in the pic is what appears when you first power on a PET, but is then quickly replaced when the system loads BASIC into memory IIRC. Maybe someone else can confirm this.

I also remember that on some PET boards (the 2001-8's) the order of the RAM is different, some of them start from the upper left to lower right, some the exact opposite. My hunch is that the RAM in the first slot is the first RAM in the line, etc. So, I am not going to start in the middle of the board. I think I would only do that if the system was acting flaky, or dying after a few minutes. That's the kind of environmental evidence that would indicate a mid board problem, all easy stuff already checked.

Bill

 

[cosam]

What happens in these situations is you get inspired one morning, get the solder sucker out, and spend the day replacing things.

That's one approach, I suppose ;-) I prefer to spend my time figuring out which component is bad, then just replacing that one.

I don't think it's video RAM, what I believe we're seeing on the screen in the pic is what appears when you first power on a PET, but is then quickly replaced when the system loads BASIC into memory IIRC. Maybe someone else can confirm this.

Yes, there's always garbage on the screen when you start up, until the part of the ROM is run to zero out the video RAM. It'd be wise to check the CPU is running (has a clock signal, activity on the address/data buses) and that the reset or interrupt lines aren't permanently asserted.

 

[geoffm3]

Does it do the sequence of startup beeps? Do the contents of the video display change at all, when you press keys, etc?

 

[billdeg]

no beeps and no reaction to keypresses. The system is dying at some point while the RAM is counted and/or before the ROMs are loaded into memory. The more I think about this ..

Here is what I plan to do, probably soon because it's on my mind
1) Check the processor as cosam suggests
2) Spot check voltages based on the "if this is getting x volts then that means this section on the board is ok" kind of logic, etc.
3) Replace the RAM chips in order of boot use. Because the 4016 and 4032 ROMS are identical, and from my experience witnessing the removal of RAM while still being able to boot BASIC, I assume that BASIC is not loaded into a specific location in memory, but instead it's placed in a higher area where room permits. I can look this up.

Bill

 

[MikeS]

It's probably not relevant and maybe just semantics, but the ROMs (including BASIC) are not "loaded" anywhere; in a normal PET/CBM (i.e. without any expansion boards) the ROMs and the code they contain (Kernel, BASIC, DOS, Editor) are permanently located in the upper 1/4 or more of the memory map (Old ROMs $C000-$FFFF, new ones $B000-$FFFF with a gap in part or all of the $E800-$EFFF block for the I/O chips).

Unfortunately almost any fault anywhere can prevent the system from completing the startup sequence.

If you have bad RAM that only presents incorrect data then the computer might well boot with reduced RAM as you say (or odd screen data if it's the video RAM).

But if the fault is a data or address line shorted to ground or another line then it will interfere with the startup process (even if it's the video RAM) and you won't get very far.

mike

 

[billdeg]

no, that is relevant, you're right the C000 page would be way out of range for 32K. My strategy is probably correct though. I hope there no are little comonents that are burned out just RAM.
b d