Baking computer chips: Fixing chips in an oven

Does baking computer chips fix them when they’re broken? Can you fix computer chips in an oven? The answer is sometimes. Unfortunately, it’s difficult to know when it will help, and when it does, we may not know why. But, when faced with a broken chip, we don’t exactly have anything to lose, either.

Baking computer chips does seem to fix them, at least sometimes. Whether it works depends entirely on why the chip failed in the first place, which isn’t always possible to know.

History of baking computer chips

bake computer chips
If you have scarce computer chips that no longer work, like an Atari POKEY or Commodore SID, you have little to lose and everything to gain by baking them to try to bring them back to life.

Baking computer parts to fix them isn’t exactly new. Computer enthusiasts have been baking failed motherboards, video cards, and SSDs to reflow their solder joints for years. And baking an Xbox 360 motherboard was one way to get rid of the dreaded red ring of death. Regardless of the type of board you’re baking, the concept is the same.

For years, some people have suspected that baking the board might actually help the chips as well, under some circumstances.

In early 2020, vintage computer and gaming enthusiasts started experimenting with baking chips and discussing them on forums. And in December 2020, Adrian Black conducted experiments and posted them to his Youtube channel, Adrian’s Digital Basement. When he fixes a computer, Adrian collects any failed chips he finds in a plastic container as trophies. So he pulled a selection of scarce Commodore chips from his dead parts bin, retested them to note their failures, and put them in the oven for 30 minutes at a temperature of 150 degrees Celsius. Then he retested them to see if their behavior changes.

At the risk of spoiling the ending, his results were less than inspiring. Of 29 chips he baked, only two of them showed improvement. And out of those two, neither fix was either complete or permanent. He also tried a second round, at 200 degrees Celsius for 2 hours, and that didn’t improve the results.

So, waste of time, right? Well, we can’t exactly say. Because there are people who’ve said it has worked, particularly in the case of Atari POKEY chips. And there’s an industry standard for, guess what? Baking chips.

The JEDEC standard

bake computer chips
If you want to try to fix computer chips by baking them in an oven, use an old toaster oven that you no longer use for food preparation. Bake for 30 minutes at 250 degrees F, then bake for another 8 hours if the chip shows any change in behavior.

Computer chips are sensitive to moisture. If a chip has been sitting in a warehouse for too long before being used, JEDEC recommends baking the chip at 125 degrees Celsius for 8 hours. And there is a second standard, for certain less common chips, of baking them at 125 degrees Celsius for 24 hours.

Who is JEDEC? JEDEC stands for Joint Electron Device Engineering Council, an independent organization that develops open standards for making and using computer chips. JEDEC has representatives from about 300 companies.

So, based on the JEDEC guidance I found, it’s possible that Adrian didn’t bake his chips long enough, even though some people have reported some success baking chips at 150 degrees for as little as 30 minutes.

Based on what the people who’ve tried it have said, it seems that 30 minutes is the minimum. But if we want a more permanent fix, following an established industry body standard seems prudent. There are still no guarantees of course. We should expect the success rate of reviving new, unused chips to be higher than the success rate of reviving old chips that died of old age.

Why does baking computer chips work?

Why is it possible to fix computer chips in the oven? Honestly, no one knows. We know one reason it might work–when it works–is due to moisture.

One other theory is that baking computer chips could reflow the connections inside the chip. But those connections tend to be welded, not soldered, and therefore wouldn’t be affected by a 125-degree temperature. So that’s probably not it.

Adrian also posted a link to a blog post about electromigration. The theory there was that subjecting a chip to heat without electricity can reverse the electromigration process that breaks down chips by slowly converting the conductors within the chip into insulators. He recommended baking chips for an hour and retesting, then baking any chip that showed improvements for a few more hours, and that this would work.

I understand being skeptical. I wondered about the results myself. But science is full of mysteries. We don’t know for certain why Lithium works to treat bipolar disorder either. If a treatment works, we go with it, and continue to observe, hoping that someday our understanding will improve. The parallels of treating a chip of unknown chemistry and treating brain chemistry where it’s impossible to know all the possible factors seems interesting. And with much lower stakes.

What about when baking computer chips doesn’t work?

Of course, there’s nothing magic about putting computer chips in the oven. Computer chips can fail for any number of reasons. If it failed due to overvoltage, the oven isn’t going to fix that. If it failed because the chipmaker messed up the chemistry (the PLA chips in Commodore 64s is a perfect example), the oven isn’t going to fix that either. That’s one reason I’m not surprised most of Adrian’s PLA chips showed no improvement.

Could there be other factors at play? Possibly.

The thing to remember about making computer chips is it’s an inexact process. A chipmaker who wants to remain in business is revising its chipmaking process all the time, to find ways to make ever smaller, and therefore, ever cheaper chips. No two companies make chips exactly the same way, because they’d be infringing on each other’s patents if they did. Two chips made by different manufacturers, or even in different fabs by the same manufacturer, can have different chemical properties. Therefore, when they fail, they can fail for different reasons and in different ways.

We think we might know two problems that baking computer chips can solve. There might be a third, or even more. But any given chip may be more or less susceptible to any of those problems.

Based on what I know, I’m not surprised only two of Adrian’s chips showed any improvement. If I were him and had those two PLA chips, I’d put them back in the oven for 8 more hours at 125 degrees Celsius, or 250 degrees Fahrenheit. But no guarantees.

Advice on fixing your own chips in an oven

So should you try it? Honestly, I think you have about a 10 percent chance of it succeeding, based on the experience I’ve heard from others. That’s not great. If you’re a baseball fan, that’s comparable to the chances of the pitcher getting a hit when it’s his turn to bat. It happens sometimes, but don’t count on it. That said, you have nothing to lose. The chips don’t work, and it doesn’t cost much to try.

I will echo Adrian’s advice and suggest you not use the oven in your kitchen. You don’t know what toxins the chip will release in your oven and you don’t want those in your food. Get a cheap used toaster oven the next time you go to a thrift store or garage sale in search of vintage computers. Please take appropriate safety precautions if you’re doing this during the Pandemic. Or just wait until next year. There’ll still be plenty of old stuff to buy next year.

You could also treat yourself to a new toaster oven if the one you have is looking a bit nasty, and repurpose your old one. You can get a decent one for $30-$40.

A reflow oven would be better, as it would offer better temperature control. But we expect a success rate of 10 percent. Frankly, unless you have a large stash of questionable chips, I don’t think it’s worth that expense. Forty bucks and getting a nicer small appliance in the kitchen seems reasonable. Five bucks from a thrift store also seems reasonable. Not hundreds.

But once you have a cheap toaster oven, there’s not a lot to baking computer chips. Test your dead chips to confirm their behavior. Mark each chip and note how it behaves. Place your dead chips in the oven and bake them for 30 minutes at 125 degrees Celsius or 250 degrees Fahrenheit. Remove the chips and retest them. If the behavior in any of them improves, bake those for another 8 hours at the same temperature, then test again.

The chip may or may not be cured at that point. But you haven’t lost much. And if you manage to revive a chip that’s been out of production for decades, you recouped the cost of that toaster oven.

Did I fix any chips in my oven?

Of course I wanted to try it myself. But, believe it or not, I only had two bad chips to test, a Commodore PLA and a Commodore SID. Most of the vintage computers I’ve fixed had problems other than bad chips. But I did have an old toaster oven I use for baking paint. So I decided to try my hand at baking computer chips.

Of course, with an observed success rate of 10 percent, with two chips of unknown history, what do we expect? Zero. What did I get? Zero.

That said, as I find other bad chips, I have nothing to lose by trying it again. I already have the oven.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: