Reviving An SD Card With Shorted Capacitors

A nice thing about SD cards is that even in their non-micro format they are conveniently small. This is however a bit of a problem when an SD card stops working, as they are not exactly designed to be easy to service, or to recover data from. There is however a very good chance that the Flash memory and controller are still fine, and it’s actually one of the passives on the tiny PCB that failed, as with the 32 GB SD card that [Yevgeniy Kapishon] recently diagnosed and recovered data from for a customer.

Tiny capacitors in an SD card package. (Credit: Aeson Labs)
Tiny capacitors in an SD card package. (Credit: Aeson Labs)

A big hint during initial diagnostics was a clear short between the supply rail and ground, but as became clear when taking the SD card apart, this one was built as a monolithic package, without exposed components on a PCB as in older SD cards. Correspondingly an X-ray machine and thermal camera were used to figure out what was inside the package, and where the short was located.

By combining the hot spot image with the X-ray it was determined that the problem was with some passives near the edge of the package. Some careful material removal later two miniscule capacitors were found to be the culprit and gently removed. With this the short on the power rail vanished, and the SD card started working again.

Having a shorted MLCC or similar passive component is a very common failure mode in general which can cripple even the most expensive device. Even if SD cards still aren’t really repairable, it’s at least reassuring to know that in many cases the data is fairly easy to recover once you have identified and removed the offending part.

19 thoughts on “Reviving An SD Card With Shorted Capacitors

  1. 32gb sdcard. Mmh $22.99 on the sandisk website. So I wonder how much the xray and thermal imaging cost? What was so important on that sd cart to spend that kinda money to say professional photographer don’t use 32gb cards usually anymore.

    1. The blog shows it’s photo’s in this case. It’s easy to imagine photo’s from a loved one that has passed away where you would be willing to spend a couple of hundred dollars to get these recovered.

  2. Would be interesting to see if a brute force of this would work. From the thermal image the capacitors appeared to heat up more than the traces leading to them. If the information on the card wasn’t worth all the intricate and $$$ effort, I’d try using a power supply set to the normal rail voltage and very slowly crank up the current limit until the capacitors cook themselves to an open circuit.

      1. I’d argue that a charred PCB under the capacitors turned to carbon PCB would’t become an issue.

        I don’t know the details of how ceramic capacitors typically fail to short. But if they are heating up, that tells me it’s a relatively high resistance short. From experience, burned to char through hole resistors typically measure very high. I’d use thermal camera and monitor the board while heating up the caps till they burn open enough to allow the rail voltage rise to an operable voltage.

        In my real world experience the only applications where a charred PCB causes more problems than the overheated/failed component itself is in high voltage, or whatever voltage is enough for it to follow the carbon and flash over. In this SD application, it’s a very low voltage.
        However, I now know what my next experiment will be if I ever find a shorted ceramic cap!

  3. I placed my dud SD card on a small anvil and reduced the thickness of it by about a millimetre or slightly more. Several bashes did the trick and the card gave no more trouble!

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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