C64 socket inconsistencies why?

[7800fan]

On the 64C that I use, the VIA and both CIA were socketed but nothing else. On the SX I am trying to fix, almost everything is socketed, only RAM are soldered in.

 

I scored a breadbin C64 through FB from my cousin, he said it didn't work anymore. When I checked, the original hazardous model PSU wasn't outputting any 5v it was dead. When I used my good PSU from the 64C on the new one, it worked fine. I had planned to borrow PLA, CPU, and kernel ROM to check my SX as the troubleshooting guide on Carlsen site suggested. But when I opened, only SID is socketed, nothing else.

 

When I looked through images on Google, it seems socket can vary somewhat. Some C64 had one CIA socketed, some has both socketed, some had PLA socketed, etc. and the 3 ROM chips can vary (kernel, character, and BASIC). A lot seems to have socketed SID.

 

I can only guess what they were thinking. Did they use a dart board on socket-of-the-day to randomize socketing? Or were those factory repair after the production line rejected the board due to failure and socket was added during repair? Or were the pick n place machine smoking marijuana?

 

Considering PLA and CIA tended to blow easily, one would think most, if not all 64 (pre-Short Board era) would have sockets for easier replacement. The short board (white 64C) used different PLA that integrated other chips and were less likely to fail. CIA can blow if you looked at the controller port the wrong way. Devices that used the 9 pin connector but were not Atari compatible pinout as well as some Genesis controller can blow it.

 

(Genesis can blow because of very slightly dfferent pinout. C64 and Atari uses pin 8 for 5v and pin 5 for paddle control. Genesis uses 5v from pin 5 and pin 8 for select. Using some controller can short C64's 5v to ground and either blow fuse, blow PSU, or blow CIA)

[carlsson]

That is the one million dollar question in the quiz show "Commodore Trivia". Many people have asked themselves the same question, and nobody has really come up with a reliable answer. I think a lot depends on which factory it was assembled at, but also which batch number it was and perhaps who assembled it. I don't know how much the sockets cost extra, perhaps on days when they got a financial report that showed black numbers they celebrated by putting in sockets, but in case the numbers were in any way red, they skipped sockets again?

 

The same holds true for VIC-20's, although pretty much all PET/CBM models which themselves were much more expensive to customers, have socketed chips among those you normally would want to exchange. Yet even there you might every now and then see a soldered in 6545, 6522 and so on.

[SSG]

That is the one million dollar question in the quiz show "Commodore Trivia". Many people have asked themselves the same question, and nobody has really come up with a reliable answer. I think a lot depends on which factory it was assembled at, but also which batch number it was and perhaps who assembled it. I don't know how much the sockets cost extra, perhaps on days when they got a financial report that showed black numbers they celebrated by putting in sockets, but in case the numbers were in any way red, they skipped sockets again?

 

The same holds true for VIC-20's, although pretty much all PET/CBM models which themselves were much more expensive to customers, have socketed chips among those you normally would want to exchange. Yet even there you might every now and then see a soldered in 6545, 6522 and so on.

 

Maybe the ones with less sockets came out last because adding the sockets would take long? Maybe they started rushing to get crap out on the market to meet demands and skipped the socketting process?

[+OLD CS1]

I always figured it was done as units had parts failures in QA. After they came off the machined lines plopping the chips onto the boards and (I think) wave soldering, it was less potential damage to the replacement chips to solder sockets onto the boards.

 

My guess.

[7800fan]

Another theory that may work: supply issue. If they are out of chip xx, they substitute socket and then add the chip later when it shows up.

[Hannacek]

I don't know how much sockets cost (very little when a big company buys them in the millions of units), but the Bean Counters at Commodore probably didn't like them. Maybe some production managers liked them, and some didn't. Maybe they did a cost benefit analysis to see how much money they would save by not using them, compared to how much it would cost authorized repair centers to fix faulty chips. Commodore did make money when people bought the replacement chips.

 

Another issue is that socketed chips make a less reliable connection than soldiered chips. However, the risks of damaging the board during desoldiering chips far outweighs the benefits of not using sockets.

My C64 recently stopped working. All the chips except the ram are socketed. The problem was the PLA was not making good contact with the socket. I removed all the chips and cleaned the legs, and it works as good as it ever did. There wasn't much oxidation on the chips, but apparently it was enough to make the PLA stop working properly. However, I doubt this was a concern, because it took decades for the chip legs to stop making good contact with the sockets.

[7800fan]

They still allowed oh about 95% of C64 with socketed VIA and a lot of SID are also socketed. One would think they'd make all of the PLA socketed as those usually blows, and CIA when idiot messes with the controller port. But then again, they spent millions to build a warehouse to store a few million defective C64 instead of taking the time to fix it and sell it as refurbished. (I think this was mentioned in one of the later RUN magazine)

[motrucker]

Another theory that may work: supply issue. If they are out of chip xx, they substitute socket and then add the chip later when it shows up.

You likely have come very close with this idea. Commodore was well known for not wasting anything. This may well be the reason we see some earlier model chips used in later serial number computers. I've been using Commodore machine since 1983, non stop, so I have seen all sorts of odd "builds" from them.

[7800fan]

If they are known for not wasting anything, then why did they build a warehouse to hold in all those returned defective C64? If I remember correctly from one of the RUN issue, they had hundred thousands that they choose to sit on rather than getting them checked out and repaired then sold as refurbished.

[motrucker]

If they are known for not wasting anything, then why did they build a warehouse to hold in all those returned defective C64? If I remember correctly from one of the RUN issue, they had hundred thousands that they choose to sit on rather than getting them checked out and repaired then sold as refurbished.

I remember reading something to that effect, but, I never heard anything to substantiate that story. Remember the old axiom to "never believe everything that is written" - or as some philosopher supposedly said "better nothing be written" .

With some of the computers I have seen from Commodore, I would tend to believe they used every part they could.

[shoestring]

Considering PLA and CIA tended to blow easily, one would think most, if not all 64 (pre-Short Board era) would have sockets for easier replacement.

 

 

The PLA tends to run hotter than all the other chips so perhaps in some factories they elected to solder them into the boards.

 

It's due to chip creep, which is a process caused by thermal contraction & expansion which tends to push the chip out of its socket as the computer is turned on / off.

Edited March 1, 2016 by shoestring