Coldcard MK4 microSD failure – works at first, then later stops reading cards

[Unsaint32]

I’m trying to see if anyone else has experienced this, or if this is a known edge case.

I’ve now had two Coldcard MK4 units show the same microSD failure pattern.

What happens:
MK4 works normally at first.

PSBT signing via microSD works for several transactions.

Then, suddenly, the MK4 can’t find PSBT files on the card.

After that, the problem gets worse:

The MK4 fails the microSD self-test

It can’t format cards

Even microSD cards that previously worked stop working on that device

This happened on my original MK4 and again on a replacement MK4.

Cards tested:
Multiple brand-new and known-good microSD cards

8 GB and 32 GB

FAT32, MBR

Nothing exotic

Linux workflow (in case it matters):
Linux laptop with built-in SD card slot

microSD → SD adapter → laptop

Save PSBT file to the microSD drive

Right-click → Unmount

Wait for popup: “Filesystem has been disconnected”

Remove SD adapter, take out microSD

Insert microSD into MK4

No unsafe removal, no write caching left active.

Why this seems odd:

Cards work initially, then stop

Failure is progressive

Same behavior across two MK4 units

Once triggered, there’s no obvious recovery path

This doesn’t feel like a simple “picky card” issue, because the cards do work at first.

[tenant48]

Have you tried formatting the faulty SD card to FAT 32 on your computer and then writing something to it?
Did you use regular SD cards or Industrial class cards?

Modern conventional flash drives and SD cards (except Industrial class cards) can withstand about 10 rewrite cycles for the following reasons:
• QLC/TLC of the cheapest bin
• Overprovisioning occurs (sometimes 0%)
• Primitive controller
• No DRAM cache
• No temperature control
• Often no proper wear leveling

[Yamane_Keto]

Did you use Sparrow to create PSBT files? I've read many similar cases, and they all involved Sparrow.

Have you tried formatting the faulty SD card to FAT 32 on your computer and then writing something to it?

There's a similar post on Reddit where the user replied that they reformatted their SD card to FAT32, but nothing changed.

[Unsaint32]

Have you tried formatting the faulty SD card to FAT 32 on your computer and then writing something to it?
Did you use regular SD cards or Industrial class cards?

Yes, the cards are fully functional in my Windows and Linux PCs. When I use them with MK4, MK4 cannot read the psbt file in it, cannot format it, cannot pass the sd card part of the device self test. They are all usual micro sd cards. HC.  One of them is Sandisk "Industrial" written on it. Don't make a difference.

[satscraper]

Why this seems odd:

Cards work initially, then stop

Failure is progressive

Same behavior across two MK4 units

Once triggered, there’s no obvious recovery path

This doesn’t feel like a simple “picky card” issue, because the cards do work at first.

But what would happen if you tried the completely new card?

Would it behave the same way?

If not, and it fails immediately, I would try cleaning the contacts in the SD slot ^{maybe blow some compressed air into it or clean it with alcohol}.

Disclaimer. Do this at your own risk.

[tenant48]

Yes, the cards are fully functional in my Windows and Linux PCs. When I use them with MK4, MK4 cannot read the psbt file in it, cannot format it, cannot pass the sd card part of the device self test. They are all usual micro sd cards. HC.  One of them is Sandisk "Industrial" written on it. Don't make a difference.

Then there are most likely two options:
Problems with SD reader contacts or mechanical wear.
Firmware errors that appear after certain sequences of operations.

I'm more inclined to the second option, since new SD cards seem to work somewhat initially. It's worth waiting for the new firmware to be released and testing with it.

[Unsaint32]

According to ChatGPT, the procedure I follow (below) adds invisible data to the sd card, and that might be breaking the MK4.

Right-click → Unmount
Wait for popup: “Filesystem has been disconnected”


When I format the card again in a Windows PC, now it works with MK4.

My Sparrow wallet is in a Linux PC so I have to use a specific code to eject the card in the terminal (not the file manager). Using the file manager did work for a while before so we shall see. I will update.

[dkbit98]

According to ChatGPT, the procedure I follow (below) adds invisible data to the sd card, and that might be breaking the MK4.

Either ChatGPT is stupid or c0ldcard device is stupid  
This is not the first time I heard people complaining about their c0dlcard device getting bricked.
I know that I would never spend my money on this hardware wallet knowing the ego-lunatic liar in charge of that project.

[satscraper]

According to ChatGPT, the procedure I follow (below) adds invisible data to the sd card, and that might be breaking the MK4.

Either ChatGPT is stupid or c0ldcard device is stupid  
This is not the first time I heard people complaining about their c0dlcard device getting bricked.
I know that I would never spend my money on this hardware wallet knowing the ego-lunatic liar in charge of that project.

Modern microSD cards seem to be a jinx for hardware wallets. I remember Lucius complaining about the brand new microSD card that refused to work with Passport Core. They solved the problem by inserting theold, pretty aged microSD card instead. The OP should probably try the same approach and take an older microSD card and insert it into Coldcard MK’s slot. In the past, manufacturers didn’t seem to be greedy using gold to gild the card’s contacts and made the relevant layers thick enough.