• Please review our updated Terms and Rules here

C64 Repair - Stuck

O

OutpostKodelia

New Member
Joined
Jan 10, 2024
Messages
3
Hi all,

First post, first C64 repair, but I have a good amount of experience repairing arcade and pinball PCBs. We bought a working C64 with a 250407 board. We did some upgrades to the video output. Everything was fine. We then bought a broken C64 to repair. The only thing that matters here is that we took our good PLA and put it into the broken C64 to see if that was the issue based on testing we had performed. When we returned the PLA to our working C64, the working C64 was no longer working (black screen) and the PLA was blister-you-if-you-touch-it hot. Okay, bad PLA, likely ruined by the other bad C64 somehow. Ordered some modern replacements since we had no known good PLA on hand.

First one to arrive was the MOS equivalent from 8-bit classics. Popped it in to the formerly working C64, black screen. Checked the dead test cartridge, sometimes one white flash, some boots we get two white flashes. Okay, so bad RAM?

Broke out the scope. I see activity on all the pins except the R/W line was stuck high. Okay, so I assumed the dead test is likely confused since it essentially can't write to anything.

Pulled out all socketed chips except VIC, MPU, and PLA. Same dead test result, one or two blinks. Chips pulled at this point are both CIAs, U4 kernel ROM, and SID.

Okay, so why is R/W still stuck high? Pulled VIC, MPU, PLA, and cartridge. As I understand it, the ONLY things that should still be touching the R/W line at this point are the RAM modules. To recap, both CIAs, the U4 Kernel ROM (yeah, doesn't touch R/W, but including it here for thoroughness), the SID, the VIC, the MPU, PLA, and the cartridge are all removed at this point.

R/W is STILL stuck high. And I tested with the multimeter too, it's exactly the same 5V as the 5V rail, about 5.06V. So it has to be a bad RAM module with some kind of funky internal short between 5V and R/W, right? Well...

The last thing I tried was to cut and lift pin 3 (R/W) on each RAM chip (U12, 24, 11, 23, 10, 22, 9, and 21) one at a time, testing again between each cut/lift thinking that cutting one of them would finally allow the R/W line to drop from 5V. Nope. Still have 5V on the R/W line.

At this point, I can't find anything else on the schematics that touches the R/W line, yet it remains at 5V. Remember, this C64 was working fine before we swapped its (now damaged) PLA back in from the other bad machine. So I don't think we're dealing with a board issue. But still, I've scoured the board and cart socket looking for any kind of solder bridge or stray metal that could explain it, but I can't find any. I checked resistance between the 5V and the R/W line, which I would normally expect to be completely open since everything that's touching the R/W line has either been cut or pulled, but no, it shows about 1.5K Ohms.

I am completely baffled. Please tell me I'm missing something obvious here, or that there's something missing on the schematic (FYI, I checked schematics from multiple sources, all looked the same regarding the R/W line). I've watched a TON of C64 repair videos on youtube over the last year or two (primarily Adrian's, but others as well) and I thought I had a decent handle on these C64s, but this one has me stumped.

Thanks in advance,
Kaydee
 
Doest it have any of those horrible single wipe sockets? If yes you could have contact problems. Either way you can inspect the socket pins and check for continuity between all relevant parts
 
My understanding is that you removed a PLA IC that was working, put it in a faulty computer. Obviously the faulty computer didn't work, and you returned the PLA to the original computer, but when you did it got blistering hot ?

It highly improbable that the defective computer could have damaged the PLA. Faults in computers, unless there is something globally wrong like excessive power supply voltage (which you would have detected with your basic initial tests) do not generally damage IC's put in as test replacements.

There are two things that could have though:

Both relate to reverse applied power supply polarity to the PLA. Either when the PLA was transferred it was plugged into the socket on the faulty computer the wrong way around, or that happened when it was returned to the good computer. Or in the transfer process the PLA IC got ESD damaged.

The PLA heating up blistering hot like this now is clearly damaged. Double check the orientation of the PLA in its socket.

I think it is unlikely that the original computer is damaged in any way (In that its unlikely that the now original damaged PLA could have damaged this computer). However, there can be exceptions.

In all probability, two things could account for the initial good computer now not working with a new PLA:

1) The replacement PLA's you got are defective, also got ESD damaged or are possibly fakes.

2) As suggested, the PLA socket on the previously good machine is defective. Possibly damaged after a PLA IC re-insertion event.

Suggestions to fix the original computer that was working:

1) assume initially by Occam's Razor that there is only one problem and it relates to the PLA you removed, and that nothing else is damaged.

2) Get a defunct IC, cut off one of its pins close to the IC body. Solder to that a small wire handle about an inch or two long. Don't let any solder run down the the part of the pin that enters an IC socket. Use this as a probe to feel the tension in each socket claw of the PLA IC's socket.

3) Buy at least 3 or 4 replacement PLA from different sources/suppliers and try those.

4) Very carefully inspect the work that you did on the pcb searching for the fault, cutting tracks etc. Make sure that you have not introduced additional faults which will confound the experiments testing the PLA socket and trying more PLA's.

5) Also be aware, there is all sorts of stuff on youtube videos that is sometimes questionable, misleading and in some cases downright incorrect. And, anything you see there, is not necessarily a good resource, as it relates to some unique electronic problem that exists in any particular machine that you may be trying to repair, a fault, which in many cases, another person may have never seen before, let alone posted a video on it.
 
Last edited:
Hugo Holden said:
3) Buy at least 3 or 4 replacement PLA from different sources/suppliers and try those.
Click to expand...
I would advise against this, unless you're referring to modern replacements. Personal recommendation as C64 PLAs are failing parts and unreliable even when handled properly: PLAnkton for around 20€ or 2x GAL20V8 and some prototype board if you want to be really cheap (some assembly and programming required).
 
My main question was about the R/W line sitting at 5V with 1.5K ohms resistance to the 5V line. I have since - I believe - discovered that this issue was a red herring. In my exhaustion, I didn't notice R51 (1.5K ohms) on the schematic pulling R/W up to 5V, so I suppose that's intended behavior for the R/W line, defaulting it to read when it's not explicitly pulled to ground for a write.

As to how the PLA died, I know what happens to IC's when they go in backwards and how ESD can damage a chip... I feel like we didn't have those problems, but we weren't recording video that we can go back and reference, so anything is possible. We have more PLAs and RAM on order, so we'll keep at it until we can pin it down. I just got sidelined by the R/W line's now apparent non-issue. Thanks all. I'll follow up when we know more...
 
powerlot said:
I would advise against this, unless you're referring to modern replacements. Personal recommendation as C64 PLAs are failing parts and unreliable even when handled properly: PLAnkton for around 20€ or 2x GAL20V8 and some prototype board if you want to be really cheap (some assembly and programming required).
Click to expand...
Yes modern replacements fine, but advice still applies.
 
OutpostKodelia said:
I feel like we didn't have those problems
Click to expand...
Click to expand...

Unfortunately there are no feelings involved, it is the Physics of Electric Charges and the way the body picks these up walking over surfaces and the high capacitance of the body, compared to the junctions of the handled parts.

ESD damage can happen, especially when chips are moved between machines.

Mostly, when you are working with some machine, your body quickly assume its ground potential or average potential if its above ground, because you are touching conductors in it at times, or handling it. So your body shares any electrostatic charge with it, until current stops flowing and your body assumes the same average potential as the circuitry you are working on.

If you remove a chip from it, that chip still has negligible electrical potential compared to you, so handling it won't likely harm it, while you remain at the desk that is.

But, if you take a few steps across the floor and go to return the chip to either to the same socket from the machine you took it from, or any other machine, your body has acquired charge.

The amount of charge even after a few steps and the right environmental conditions is easily enough to destroy a small semiconductor junction in an IC.

If you were carrying the part, when you reach to install the IC in the socket, a current flows via some pins you are toughing and via some other pins into the socket, or into the IC pins directly when you pick it up from the bench. The same happens when you walk away from any machine, if you have chips out of it on the bench, and come back, your body can have acquired a significant charge. The problem results from the relatively high capacitance of the body, vs the low capacitance of a junction. The formula is charge Q = CV, = capacitance x voltage. But it is the charge to capacitance ratio Q/C that is important, as it determines the terminal voltage.

As an example, any fixed amount of charge applied to a smaller capacitance will result in a higher terminal voltage than it does for a larger capacitance. The human body capacitance is about 100pF, but that of a semiconductor junction, like cmos, barely a few pF. Therefore, even if the body only got charged to 10v, that amount of charge would generate a hundred of volts, if it is transferred to a small capacitance of 1pF.

One way to prevent this trouble, is before putting an IC in a socket, especially if you have been walking with it, or walking around before you do it, is to touch one of your hands to the ground system of the machine you are putting it in, and wait a few moments for your body to assume the same potential as the machine. And do the same, if you leave a machine where IC's are out of it, when you return to it, before touching anything, connect one hand to the machine's ground system and wait. I do this as a matter of course and second nature, and it helps a lot to reduce the chance of ESD damage. Of course cmos are the more vulnerable to ESD damage, much more so than TTL.

The above techniques generally work in the absence of anti-static mats and wrist straps etc. The trouble is that sort of kit, it makes people think they are ESD safe, unfortunately the risks remain, they are only partially mitigated. A technician can still find ways to unwittingly transfer charges to devices without knowing or thinking how they managed it, and even if you had video's of what happened on the day, it might not reveal how it happened.

Clearly though you have described a situation where a chip was definitely damaged after a transfer to another machine. And if you are 100% confident it wasn't inserted backwards, then the likely cause, above all other possible causes, assuming the power supply in the second machine was normal, is ESD damage.
 
New PLA and we're only partially back in business. Dead test runs no problem, no errors. Diagnostic cart runs, no errors reported but when it gets to the last part of the SID test where it plays static sounds, it starts to write some garbage characters to the screen (mostly but not exclusively @ and ! symbols), then stops writing garbage when the SID test finishes. Then more garbage at the same point in the second pass. Third pass it often just crashes before it gets to the SID test again. RAM appears to be getting corrupted. Garbage happens with or without the SID chip installed. Boot up without the diagnostic cart installed goes directly to a colorful garbage character screen. I suspect one or more of the logic ICs failed, in particular I'm suspicious of U13/25 based on what I'm seeing as compared to other reported failures. Also possible it's not that, but bad RAM, though I'm leaning logic chips since the diagnostics can't seem to decide on which - if any - RAM is bad, depending on the run... though recent runs aren't reporting any RAM issues at all. I've ordered replacements for quite a few chips including the RAM (we have other C64s to repair after this one)... more to come once they arrive.

As to ESD, once again, I know how ESD works and in many many years of working on PCBs from the 70s-90s or with the countless computers I've built or maintained I've never had a problem with it, BUT in my frustration with the board I may have let my guard down and zapped it. We'll never know... All I can do is keep marching towards the fixes, albeit a bit more cautiously w/rt the ESD potential.
 
OutpostKodelia said:
New PLA and we're only partially back in business.
Click to expand...

If you are partially back in business with a new PLA, at least that confirms that the first lot of PLA's you tried as replacements were defective too, which is what I suspected from the scenario you painted.

If you have a working computer anywhere, before moving on, I would suggest you subs in these new PLA's into that and confirm 100% that these work for sure in a known good computer. If that is the case, fit them to your defective computer and then march on with the diagnostics after that.
 
You must log in or register to reply here.
Back
Top