Is the SNES a realible console? So worried about reliability

I've seen quite a few dead ones, but I don't think the failure rate is excessive for early 90's consumer grade electronics. At the time the SNES was made, semiconductors were sufficiently unreliable that parts that were going to be used in military/aerospace/hi-rel applications were typically high-temperature soak tested to weed out all the bad parts. Obviously, they didn't do this with consumer grade stuff so the bad parts normally ended up as early-life failures after they got to the customers.

And IMO swapping the chips only makes sense if you have multiple broken units - say you have a board with a bad CPU on it and another that has a good CPU but a bad PPU - you can make a good unit out of two bad ones. Of course, I have hot air equipment and a preheater, so swapping over a QFP isn't that much of a challenge.

 

As long as you never buy a second hand SNES you can be confident of that.

 

I'm just surprised because my SNES made it around 22 years before having that failure for no apparent reason. Technically the system still works fine. If you don't mind an annoying vertical line through your sprite cells then there's no problem. I do agree that it only makes sense to swap parts around if you're dealing with 2 or more boards with bad chips. Unless you have some secret source of NOS custom chips.

 

Shit, I've been testing with SF2 for years. Never had any problems but I guess I should try other games too from now on.

I'd say roughly 20 - 30% of the Super Famicoms I've ever bought had problems, mostly strange colours, boot glitches, or no boot at all, and I've had maybe 200 or so.

 

Well, I would certainly keep the board, since it sounds like it still has a good CPU and that's by far the most common failure on these systems. In fact, that sounds like it might be a RAM problem, which is really easy to change (OK, it could also be one of the PPUs, which is rather less easy to change).

It's worth getting or building a burn-in test cart if you are dealing with a lot of systems. You can also put the burn-in test ROM image on a flash cart, but that sometimes won't boot on systems where the bare ROM will.

This was a few years ago, but I also got a batch of 100 untested SFCs and after testing them 7 had electronic faults, which matched up well with the "about 10%" failure rate I was seeing.

 

While I haven't done any test software to check if my original unit has a RAM problem, when observing a variety of software it seemed unlikely since the only single fault was a vertical line in sprite cells, never background tiles, and I believe it was for each 16x16 sprite cell not every 8x8 cell. I had determined which bit of the 4bits of the pattern data was behaving as though it was stuck as set to 1. It's been awhile but I was fairly convinced the fault was inside one of the two PPU chips and not due to any bad traces, joints, or ram chips. I may have even run the usual test or burn-in cartridge roms to test the workram and video ram.

Maybe one day I'll dig it out and look into it again. I'm keeping it ofcourse since it is my original SNES console PCB. But it was easier to "fix it" by swapping PCBs from a working system in a yellowed case. Mine had no yellowed at all.

 

In general, the custom chips seem to be the highest failure rate items - the main reason I suggested the RAM is that (if you have a hot air machine) it's really easy to change - just heat the chip up, lift if off with a pair of tweezers add some flux and drop the new one in place. The burn in cart can also detect RAM failures (although I have seen it get it wrong and show a RAM failure when the problem was actually one of the PPUs).

 

Yay just bought 2 more dead units 1 was water damage other was a dead cpu

 

The early models, CPU-01, have problems in 50% of the units I buy from Japan.

I had one myself I ran for about a year, and then suddenly developed graphic issues.

On the good side, later units with other motherboards don't tend to have those problems.