I run a small brand of stainless steel crypto backups and recently I put my most popular one inside a ceramic furnace for 12 hours straight, with a peak temperature of 1150 °C (2102 °F) which lasted 1 hour.

My backup works by converting each word of your seedphrase into its equivalent number using the BIP39 Standard, a static pool of 2048 words. Then you just grab a hammer or mallet and stamp the sequence of numbers into the plate.

Product: https://cryptonumeris.com/products/plate-s

Original wordlist: https://github.com/bitcoin/bips/blob/master/bip-0039/english.txt

So instead of shipping 27 alphabetical stamps I just ship 9 numerical stamps (6=9).

I use AISI 304 Stainless Steel which has a melting point of 1450 °C (2642 °F) so I knew it would not melt, this was rather intended to be an endurance test against high temperature for a long period of time.

Fully stamped plate

Macro shot of the stamped numbers

***

Anyway, so one of my dad's friends runs a ceramics factory which manufactures Raku (his name is Juan), and he was kind enough to let me put one of my plates in one of the furnaces.

On April 12 at 6:30 am the test had begun, the furnace was on. Juan told me the peak temperature was gonna be reached at around 4:30 pm so I went to the factory at that time to photograph the live furnace and and the thermostat.

Please note that the furnace was turned off at 6:00 pm but remained closed until the next morning at 9:00 am, an extra 15 hours which I do not have the graph for, but it's safe to say it remained very hot inside the furnace for a good few hours after it was turned off.

The live furnace.

Peak temperature

Temperature graph

The Aftermath

On April 13 at 9:00 am the test had concluded, the furnace was practically at ambient temperature. The first pic was taken by Juan when they first opened the furnace. I got really anxious because you can barely see any marks on the plates so for a moment I thought they were completely done. I arrived at 11:30 am to the factory only to be surprised by the results.

First pic after opening the furnace

Front and back

Close up

Macro shot of the stamped numbers

Result & Conclusions

The backup data was fully recovered, you can see all the numbers because the steel has almost all of its integrity even after being burned down for half a day.

After carefully reviewing both burned plates and the nature of this benchmark, I came to the conclusion that this test was a total success. I like that this ended up being an extreme endurance test that really shows the strength of the steel.

There's no secret, if the furnace would have reached a temperature of 1450 °C (2642 °F) both plates would have melted and redeemed irrecoverable. But realistically speaking, in what scenario would these crazy temperatures be reached? The average house fire temperature oscillates between 600 °C and 800 °C (1100 °F - 1500 °F). Source 1 Source 2

I do not claim to be an expert in house fires but I firmly believe that 12 hours of high temperature is a long amount of time and proves that this grade of steel is suitable for this kind of application.

I would love to hear everyone's opinion on my test, was it good enough? I do believe it simulates a severe house fire. If you have any cool ideas for other benchmarks against steel!

Full article: https://cryptonumeris.com/blogs/cryptonumeris/stress-test-1

Edit: Thank you all for your feedback! :]