New metal seed backup method

[jakob6102]

Hi there, I have been working on a new method for creating metal backups and would greatly appreciate any feedback on my method. I have written up a description of how to create the backup cards here: https://github.com/jakob6102/seedcard

This method is:
-Cheap and simple to create
-Requires minimal tools
-Small form factor
-Language agnostic
-Very durable

I believe this could be a good way to make metal backups simple and accessible to a wider audience as it only costs a couple of dollars to make each one and with the support of wallet software can be made very quick and simple to do.

[pooya87]

This is an interesting approach. Thanks for sharing.
Just want to point out one thing, mnemonics can have different lengths[1] such as 12, 15, 18, 21 and 24 with both 12 and 24 words being the most common. You should also include a grid for these word lengths with dimensions to make it easier to print a grid for 12 words for example.

[1] https://github.com/bitcoin/bips/blob/master/bip-0039.mediawiki

[dansus021]

https://github.com/jakob6102/seedcard <-- this is really fresh idea  its like some sort of morse code. would be great if you sell this 

[slaman29]

Very nice share, I have been thinking about something similar using cheap metal but that is also durable, and yet hard enough not to make accidental punches. This is for sort of inheritance needs for me. Morse and binary are good ideas! I just wonder if I am overestimating ability of closeby relatives to understand what this is so am researching other means of signal language that could even work as backup maybe metal plate copies in Morse, binary and something else:)

[aysg76]

Really good way to backup your seed phrase on metal sheets and what i like is storing and punching them in binary code first of all and then center punch it.Many times we saw people lost their backup phrase and don't have any access to their funds and regret of being careless.It is good to see that you have shared your idea on GitHub for whole community.You should also keep in mind different lengths for seed phrase as suggested by @pooya87 and make sheet accordingly for them.

You can have a look at test of various metals here : https://jlopp.github.io/metal-bitcoin-storage-reviews/

This can be more helpful to you in choosing the best metallic plate to know about heat or melting point and save them in any condition.

[ranochigo]

You can have a look at test of various metals here : https://jlopp.github.io/metal-bitcoin-storage-reviews/

This can be more helpful to you in choosing the best metallic plate to know about heat or melting point and save them in any condition.

I think OP isn't selling the plates? Just the idea of etching the seeds onto a metal plate using an obscure method.

It would be a good idea to test it out on a stainless steel and get someone to stress it if possible. I know certain methods tend to result in the backup being really hard to see after that. It can be dependent on the kind of metal used for the backup as well though.

[NeuroticFish]

I've seen another approach not long ago. I don't know if it's new, but I find it rather cheap if you think that you do keep the seed on steel:
https://portofele-hardware.ro/pastreaza-seed-ul-intr-un-mod-ieftin-si-sigur/

Unfortunately it's in Romanian, but the images may be just enough. Or, translated with Google:
https://translate.google.com/translate?sl=ro&tl=en&u=https://portofele-hardware.ro/pastreaza-seed-ul-intr-un-mod-ieftin-si-sigur/

[jakob6102]

This is an interesting approach. Thanks for sharing.
Just want to point out one thing, mnemonics can have different lengths[1] such as 12, 15, 18, 21 and 24 with both 12 and 24 words being the most common. You should also include a grid for these word lengths with dimensions to make it easier to print a grid for 12 words for example.

[1] https://github.com/bitcoin/bips/blob/master/bip-0039.mediawiki

Thanks for the feedback  , I can certainly provide files for shorter seed phrases. The other option is to just cut down the 24 word grid to the size you need.

[jakob6102]

https://github.com/jakob6102/seedcard <-- this is really fresh idea  its like some sort of morse code. would be great if you sell this 

Thanks  . I might look at selling these if I feel there's enough interest to justify the capital requirements. This came about from me not liking the cost of existing options so it was designed to be cheap, I reckon they could be sold with the guide grid already attached for $5-7 dollars so way less than anything else on the market.

[jakob6102]

Very nice share, I have been thinking about something similar using cheap metal but that is also durable, and yet hard enough not to make accidental punches. This is for sort of inheritance needs for me. Morse and binary are good ideas! I just wonder if I am overestimating ability of closeby relatives to understand what this is so am researching other means of signal language that could even work as backup maybe metal plate copies in Morse, binary and something else:)

I used stainless steel which wasn't expensive and should be very durable to heat , acid and impacts. Yeah it not being recognisable is both a blessing and a curse, a thief or anyone that sees it isn't going to have any idea it's bitcoin related but then neither will no-coiner family members. Hopefully they should be able to realise it's important enough to not throw out considering its a steel card with a weird code stamped into in  .

[jakob6102]

You can have a look at test of various metals here : https://jlopp.github.io/metal-bitcoin-storage-reviews/

This can be more helpful to you in choosing the best metallic plate to know about heat or melting point and save them in any condition.

I think OP isn't selling the plates? Just the idea of etching the seeds onto a metal plate using an obscure method.

It would be a good idea to test it out on a stainless steel and get someone to stress it if possible. I know certain methods tend to result in the backup being really hard to see after that. It can be dependent on the kind of metal used for the backup as well though.

The one I made as the example is stainless steel so it should be very durable but I still need to properly stress test it. If I was to sell it I would run tests with different stainless alloys to find which one seems to be the best for preserving the data on it. This design doesn't need the metal itself to be etched or anything like some other metal backup options so if in the worst case the plate is too hard to read the data from, you could always sand the top layer to clear it up and the punches would be deep enough to not be erased.

[jakob6102]

I've seen another approach not long ago. I don't know if it's new, but I find it rather cheap if you think that you do keep the seed on steel:
https://portofele-hardware.ro/pastreaza-seed-ul-intr-un-mod-ieftin-si-sigur/

Unfortunately it's in Romanian, but the images may be just enough. Or, translated with Google:
https://translate.google.com/translate?sl=ro&tl=en&u=https://portofele-hardware.ro/pastreaza-seed-ul-intr-un-mod-ieftin-si-sigur/

Thanks for sharing, this is another good method and is currently the cheapest way of making metal backups I think, you can even find 3D files online for a jig that makes stamping the washers very easy. However I think my design could be better than this as this method requires an English alphabet stamp kit and a hammer while mine just needs a cheap auto centre punch, you also need something hard to stamp the washers against to make the letters visible. It also takes quite a while to do and if you lose too many washers from the seed then you're in trouble.

[NeuroticFish]

OP, I advise you make longer merged posts and reply to everybody in a "grouped" manner, your way to make consecutive answers is not OK.

Thanks for sharing, this is another good method and is currently the cheapest way of making metal backups I think, you can even find 3D files online for a jig that makes stamping the washers very easy. However I think my design could be better than this as this method requires an English alphabet stamp kit and a hammer while mine just needs a cheap auto centre punch, you also need something hard to stamp the washers against to make the letters visible. It also takes quite a while to do and if you lose too many washers from the seed then you're in trouble.

I agree that yours should be cheaper, just I find the word list to binary index file quite a drawback.
But it all depends on what each and everybody prefers, hence the more methods the better.

Why diameters of the holes on the following plate are differenent, does it mean something?

It may be because with the center punch it depends on how hard you hit with the hammer and you'll probably won't have all the punches at the same size (diameter).

[jakob6102]

~

Why diameters of the holes on the following plate are differenent, does it mean something? Could you proceed with the optical reader for such encoded plate that would be able to convert the binary code it holds into SEED phrase and present it on display? I think you could commercialize you device you  if you were manage to do this. IMO, without a reader it is of low interest.

https://i.postimg.cc/Wz17SFTK/dd-121140.png

Sorry for the confusion, the different sizes don't mean anything the dots are just supposed to represent marking the guide with a pen.

Yes I fully agree with wanting to be able to scan the plate directly with a hardware device, my next step of this project is probably going to be to speak to the amazing people working on the Seedsigner and see if they think it's possible to add that feature to their project. If it is possible then you could even have a mode on the device that will tell you the checksum for an incomplete seed, so you would generate offline the first 256 bits of entropy using say coin flips, mark them onto the guide then scan it and the device would give you the final 8 bits for the checksum. If scanning directly isn't possible I still think this could be useful, especially if the device lets you input the binary so you don't need to convert to words. While its 11 bits of data per 'word' vs 4 for an actual word, it should be less button presses as there's only two options (1 or 0) per character vs 26 letters to scroll through when inputting a word into a HW device.

[jakob6102]

OP, I advise you make longer merged posts and reply to everybody in a "grouped" manner, your way to make consecutive answers is not OK.

Thanks for sharing, this is another good method and is currently the cheapest way of making metal backups I think, you can even find 3D files online for a jig that makes stamping the washers very easy. However I think my design could be better than this as this method requires an English alphabet stamp kit and a hammer while mine just needs a cheap auto centre punch, you also need something hard to stamp the washers against to make the letters visible. It also takes quite a while to do and if you lose too many washers from the seed then you're in trouble.

I agree that yours should be cheaper, just I find the word list to binary index file quite a drawback.
But it all depends on what each and everybody prefers, hence the more methods the better.

Why diameters of the holes on the following plate are differenent, does it mean something?

Apologies for the individual replies, how do I make a longer merged post? do you mean reply to multiple quotes in the same comment? I'm sorry I'm new to bitcointalk.

Yeah the word to binary is currently the largest issue with this method; luckily it's something that can easily be remedied by wallet software giving the option to display the binary and trying to get some to do this will be my next task.

[NeuroticFish]

do you mean reply to multiple quotes in the same comment? I'm sorry I'm new to bitcointalk.

Yes, that's what I meant, sorry if I was not clear.
And you don't have to apologize to me  , just learn and do better for the future. I guess that the older posts will be (sooner or later) merged by a mod if you won't do that yourself.

[aoluain]

~

Why diameters of the holes on the following plate are differenent, does it mean something? Could you proceed with the optical reader for such encoded plate that would be able to convert the binary code it holds into SEED phrase and present it on display? I think you could commercialize you device you  if you were manage to do this. IMO, without a reader it is of low interest.

Sorry for the confusion, the different sizes don't mean anything the dots are just supposed to represent marking the guide with a pen.

Yes I fully agree with wanting to be able to scan the plate directly with a hardware device, my next step of this project is probably going to be to speak to the amazing people working on the Seedsigner and see if they think it's possible to add that feature to their project. If it is possible then you could even have a mode on the device that will tell you the checksum for an incomplete seed, so you would generate offline the first 256 bits of entropy using say coin flips, mark them onto the guide then scan it and the device would give you the final 8 bits for the checksum. If scanning directly isn't possible I still think this could be useful, especially if the device lets you input the binary so you don't need to convert to words. While its 11 bits of data per 'word' vs 4 for an actual word, it should be less button presses as there's only two options (1 or 0) per character vs 26 letters to scroll through when inputting a word into a HW device.

The beauty behind this system is that no tools other than a centre punch and hammer are
required. My own way of producing similar plates is to use metal letter punches, it takes
more time though.

I like the idea of possibly being able to scan the binary directly into a device but
it seems like it could be a bit more involved and could make it look more complicated.

[MusaMohamed]

Backup with metal seed is costly.

If you can save your backup safely, paper or USB is good enough to use. I don't have money to buy metal seed just for back up. I don't see this method is really safely and necessarily from the review in Metal Bitcoin Seed Storage Reviews and Stress test

[BlackHatCoiner]

Hi, jakob. Welcome to bitcointalk!

I read the seedcard from github, thank you for sharing this with us. But, I'll be judgmental; generally deriving addresses from a seed was implemented to be able to recover your funds using a fixed number of bits no matter the addresses you've funded. Then, we combined it with the mnemonics which was a way to avoid any human mistakes.

While I do agree on having hierarchical deterministic wallets, I'd never use a non-encoded seed to store my money. Your solution seems fine for someone who's okay with writing the bits of the seed (essentially the entropy), but I think that the odds for a human mistake are more than they seem.

You said it yourself:

There is some room for human error with the word>binary conversion and this conversion is hard to visually error correct however these issues are alleviated by wallets adding the suggested impovements to display and accept seed binary natively.

[jakob6102]

Hi, jakob. Welcome to bitcointalk!

I read the seedcard from github, thank you for sharing this with us. But, I'll be judgmental; generally deriving addresses from a seed was implemented to be able to recover your funds using a fixed number of bits no matter the addresses you've funded. Then, we combined it with the mnemonics which was a way to avoid any human mistakes.

While I do agree on having hierarchical deterministic wallets, I'd never use a non-encoded seed to store my money. Your solution seems fine for someone who's okay with writing the bits of the seed (essentially the entropy), but I think that the odds for a human mistake are more than they seem.

You said it yourself:

There is some room for human error with the word>binary conversion and this conversion is hard to visually error correct however these issues are alleviated by wallets adding the suggested impovements to display and accept seed binary natively.

Thank you for the feedback and I certainly agree, not only is there more room for error but you also can't visually check for errors like you can with the spelling of words from a set list. How much do you think the wallet having the option to display the entropy in a visual format and creating this seed backup method from that alleviates this concern? I'm thinking it could be displayed something like this (again using "satoshi:10111111011"): •^o••••••^o••