Mike wrote:
A valid point. I must confess that - for myself - I can't really come up with a solid statistics. There's only one VIC-20 in my possession, the first one I got in 1983 and it still is working without any issues. Granted, it has got some hardware upgrades over the last years, but beyond that I never found myself in the situation to actually repair it. It's a VIC-20CR and all its chips were originally soldered in, so - as you say - it ensures proper cooling over the pins cf. the chips being socketed.
You're right, it's very difficult to come up with solid statistics! The only method would be to run two new-old-stock chips from the same tube to death, one heat sinked and one not. And that would be barbaric
I'm very glad your VIC is still working in its original condition! It sounds like it's been well looked after
It's very rare to find one that hasn't been fettled with, as most of them seem to require repairs of some kind at some point. Especially if you've bought a used one more recently, as you can never tell its past. This one was dead-as-a-dodo, but only thanks to an inverter chip, a multiplexer chip, and the BASIC ROM. The BASIC ROM doesn't run as hot as the others chips and glue logic doesn't even run warm, so these must have been manufacturing flaws or something else.
Mike wrote:
I can agree with the first and third point, however can the maximum reachable temperature really be decreased unless you actually care for forced air cooling? Also, my concerns in the other thread mainly centered around that the added heat capacity near to the chip die then keeps the die warm over a longer period when the VIC already has been switched off again. Those effects presumably depend on the average power-up time nowadays, though.
You're right, it's difficult without forced-air cooling from a fan or something similar, as static air cooling of the heat sink limits its effectiveness. Also, the effectiveness of the heat sinks depends on the environment the computer is used in (ambient temperature), and they are basically pointless if the cardboard RF shield is still present inside the case or the cooling vents are blocked.
However, I am under the impression that - even with only convection cooling - having the greater surface area of the heat sink to exchange heat with the cooling air does reduce the maximum temperature (it varies between sources, but I've seen hotter chips run up to 15C cooler with a statically cooled heat sink). The sheer volume of the heat sink seems to aid this, as it shares the thermal power from the chip and dissipates it across a greater mass, again helping to reduce the maximum temperature.
Mike wrote:
No worries. As long as people keep in mind, the heat sinks are not the 'cure-all' of things, but rather those other failure modes also need to be taken into account (and one can't do much about moisture creep sadly), then everything's okay.
Absolutely!
It's important that we are aware of all possible issues, and that we make sure to fix the most impending culpits such as old PSUs first. It's also very important to remember that there aren't any magic cures when it comes to old electronics, so if a chip's dead or dying, there's nothing you can do. We can only try to slow their deterioration, and this even applies to healthy chips.
whereas as you mentioned paste is easy to get off but can be a bit messy. Paste never truly 'dries' unless it's very old or has been running very hot, so sinks can be knocked off with relative ease.