Not baking, but I've recovered data from spinning hard drives after freezing them, on a couple of occasions. The last one required keeping it in a cooler full of ice, because the data transfer was so slow that it would stop once thawed.

freezing helps solidify and congeal the grease that degrades and becomes stickier over time and causes the drive mechanisms to stop spinning or blocks the seeking heads. As the drive heats up again, that degraded grease becomes a sticky goop again and gums up the drive.

Heating devices to fix them is a form of "reflowing" where you're getting the solder on the board close to its melting point, helping the solder 'reflow' into the cracks that have formed between its component contacts, via thermal or mechanical stress over time.

But why would grease work better when it gets more solid?

Usually the lighter fractions evaporate away, leaving solid grease that has more resistance.

If you get grease into your garbage disposal grinder, I’ve heard you should use cold water to get it to congeal so the grinder can break it up and it will continue flowing once it’s in the drain. If you use warm water, it might flow, but only long enough to then solidify on your drainpipes.

Maybe something similar is happening on the hard drive: sticky is too viscous to flow at full speed, but solid just breaks apart and dances around on the platters while the heads can make contact with the magnetic media.

Xbox 360's that had a red ring of death problem could sometimes be solved by baking them or at least applying some heat via a heat gun (in the right place).

I got an extra 2 years out of mine before it died doing this

Also the trick where you wrap one in a towel and let it cook itself.

That was just a low-tech non-targeted way to perform the same application of heat without breaking the warranty sticker on the console.

And it didn't actually work, because you can't get it to heat up to soldering temperatures that way (if you did get it that hot, the CPU cooler pressure would instantly crush all the BGA solder balls and render it a massive short circuit).

That only kind of worked temporarily because putting the system through thermal stress made things shift around due to thermal expansion, and possibly make contact again.

Yeah, it was a classic characteristic of the G84 series of GPUs from Nvidia (8400M GS in the XPS M1330 I owned was one). The Ball-Grid Array solder would crack in a heat/cold cycle and so you'd effectively get an open circuit. By baking it, the solder would reflow and close the circuit again.

According to Louis Rossmann, it is a myth. Baking GPUs don't re-flow the solder, the recommended temperature is too low for that anyways.

Instead it just "agitates" stuff at the molecular level, which can get the card running for some time. It is not a permanent fix though.

Worth noting that Louis Rossmann is a professional repair technician, so he may be a little bit biased, but he has good arguments.

I tried the heating trick on several occasions and on a variety of devices. Usually with just a hair drier. It tends to work. Great as a stopgap, but expect to replace the part later.

"Molecular agitation" is a weird way to spell "sintering".

When my mother's smartphone wouldn't boot anymore, I took it apart and put the main board in the oven for a while. It worked perfectly fine again for a week or so. I repeated the procedure a few times, until her new one arrived.

No idea why it works though, or why it only lasts a week.

What made you think of trying this? I think I need to bake more broken things.

Solder can become brittle/flake and break connections

Putting it the oven softens it and allows it to reflow a bit, which can reconnect broken connectors

It came up a few times when I googled it. Since the phone was completely dead, I figured it was worth a shot, nothing to lose.

Also, baking stale chips makes them no longer stale.

If the early day of home fax machines, my mother tried it with a fax she had spilt water on. This doesn’t work.

I wouldn’t say it’s surprising, ultimately it’s just a controlled thermal cycling to repair some physical defect. It’s counterintuitive because there is a large mythos around “heat is always bad” for electronics and delicate items.

With the gpu or motherboard something has warped out of shape or a solder joint has become defective. Akin to heating up the top of a jar with a stuck lid.

It's counterintuitive but not surprising?

To be honest I’ve not really linked these two concepts together before so I could be missing something. Surprises are emotional responses to sudden or unexpected events. Counter intuition deals with the truth value of facts that go against what you might otherwise expect. I guess you could have a surprise reaction to learning something counterintuitive but I don’t think I have.