Switch BSOD no obvious hardware issues

Got a model 2017 switch from ebay with bsod, already checked usual ICs for shorts, could not find hardware issue. Seller told that the switch suddenly blacked out and did not came alive again.

Switch can access RCM, already dumped the keys with “biskeydump.bin” payload. Currently trying to restore the NAND with this guide gist.githubDOTcom/Karunamon/d3938368c548a61a5a41281bb89f0e49#step-3-blowing-out-the-cobwebs-and-hacks, injecting memloaderv3 via RCM works but fails at executing umms_emmc.ini with “got 0 bytes, expected xxxxx bytes, error during lzma decompression”. Also if I use the TegraExplorer RCM payload, accessing the emmc info view causes a reset.

Are there any other ways to restore the switch? Thanks in advance!

Use this tutorial:


Do not forget:
You need a fully populated hactool keys.txt file. These cannot be shared as doing so is illegal. Google is your friend.

Thanks for the reply. I’ve only got a key dump with biskeydump. Every try to inject lockpick_rcm resulted in a non booting console -> display stayed black. Is the key dump with biskeydump (devices.key) sufficent?

Also the link you were providing requires to inject memloader via RCM which seems to fail in my case because of the “got 0 bytes … error during lzma decompression” error message. Can’t figure out why.

I sent you a private message with keys.txt file. Save it as Text Document (.txt)

Study these two tutorials.

1 Like

Thanks for the tutorials. Both of them show the step where they inject memloader payload into the bricked console to access the eMMC chip via a pc tool.

In my case, I’m not able to access the switch as a physical drive. After injecting memloader like this

The switch receives the payload but is then stuck at this screen with “Error during lzma decompression, got 0 instead of xxxxxxx bytes out!” nevertheless which eMMC section I’m trying to mount.

Right now im doing some more research on the hardware. The bare logic board is drawing around 0.13 - 0.15 amps at around 5 volts. Is this normal behaviour or shouldn’t it draw any current at all?

Is ok, but I think you have a problem and with emmc memory…

I managed to get a separate emmc memory module which I can access via tegraexplorer (just like the old one) but the problem with memloader is still present. What else can cause this issue?

Possible problems with the Tegra processor or any of the memory ram.

Short update on this: I reflowed the memory ram and the tegra processor and now memloader can load the nand successfully and I’m able to mount the drive on my pc. Unfortunately I’m not aware if the tegra processor or the ram was the problem. Thanks so far!

1 Like

I’m glad I was able to help you