Looking for people knowing how to design and manufacture bitcoin miners

[lebitcoin75]

Hi, we have two mining farms in the US and since it has been very hard to purchase miners in quantity lately we are looking to design and assembly our own miners in the USA. if someone here as experience with that matter kindly reach out to me. thanks

[sidehack]

Hard part's gonna be getting enough chips. The big guys are sold out because the ASICs themselves exist in limited quantity. Designing the innards of the chip is about 95% of the effort required to make a miner, and it's nearly impossible to do for ordinary outfits, so your best bet is to try and use existing chips from one of the big manufacturers - but why would they sell you chips if chip stock is the reason their products are on backorder?

[lebitcoin75]

Hard part's gonna be getting enough chips. The big guys are sold out because the ASICs themselves exist in limited quantity. Designing the innards of the chip is about 95% of the effort required to make a miner, and it's nearly impossible to do for ordinary outfits, so your best bet is to try and use existing chips from one of the big manufacturers - but why would they sell you chips if chip stock is the reason their products are on backorder?

So do you think this is an impossible mission?

[sidehack]

To do on any kind of large enough scale to merit filling a farm, probably.

You might talk to someone like BitFury or Canaan about chip supplies. BF may hook you up, and their stuff is fairly straightforward to work with, though only a little better than S9 for efficiency last I checked.


If you can get someone to sell you chips in sufficient quantity though, making a miner around them won't be too terribly difficult.

[Vantis]

If you can get someone to sell you chips in sufficient quantity though, making a miner around them won't be too terribly difficult.

Even so, i doubt it's good option since manufacturing cost in US is very high compared with China. Apple tried it with their Mac Pro, but it didn't go well.
But it's different case if OP find someone with necessary skill and manufacturing cost isn't big problem.

Hello,

While I am fairly new to mining, it has become a slightly obsessive bug for me the last few months. I saw this post and oddly I am a M.E. for PCB manufacturing in the states. I have been in SMT(Surface mount tech) for over 25 years. I am the guy who picks the machines, gets the floors laid out, quotes the business, develops the flow chart, PFEMA, Control plan etc. to run the PCBs.

I have worked in both high and low volume places. Currently in automotive, but have been in EMS, OEM's, etc. I have worked in about 12 US plants doing, including China, Mexico this so I know the manufacturing side very well all around the world and quoting US vs Mexico vs China, I do it very commonly.  I also know each machines in the process to expert or near expert level.

So I am pretty much tip to tail you give me raw parts (parts, including programmed or the program to flash the "IC, I don't care if it is a QFP, QFN, SOIC, TSOP, etc, prefer no BGA because X-Ray machines get spendy to check its solder joints and I can flash it, also need the raw PCB, if I get schematics I could probably try and toss them into a gerber/cad file to get a quick turn board). I can give you a completed product ready to plug in mechanically and electrically.

Anything with PCB's I don't know I can find out, but the design of the chip is the only part I am missing in this puzzle. Manufacturing lives in a separate world from the Chip designers because their manufacturing is just a separate industry and a raw material to me. An example is like a person who puts in your floor doesn't know the people who engineered the laminate material, we just have nothing we need from each other.

Feels like a job app. lol

In all honesty if you had the board schematics I could mass produce it in the states at nearly the same cost. In the states we have just automated more so we have removed most the labor, and keep pressuring on us to lower more relentlessly. The higher the volume the better the states can compete by adding more automated processes. So for cost we automate more, they toss bodies at issues.

Am I hired. lol, no honestly if you have questions I can either answer them or get you an answer. Except the chip, can't really help there.

** Just googled a Screen printer, placement and reflow machines to build them would be about 75k to 100k-ish.

[sidehack]

If you find the chips I can probably design you a board and then let that guy build you a factory. You'll need someone to write you a driver though.

[Vantis]

In all honesty if you had the board schematics I could mass produce it in the states at nearly the same cost. In the states we have just automated more so we have removed most the labor, and keep pressuring on us to lower more relentlessly. The higher the volume the better the states can compete by adding more automated processes. So for cost we automate more, they toss bodies at issues.

I barely know about electronic hardware/engineering, but how about cost of the RAW material and hardware which used to manufacture mining ASIC automatically? AFAIK most of those stuff are imported and currently the import cost isn't cheap.

If you find the chips I can probably design you a board and then let that guy build you a factory. You'll need someone to write you a driver though.

Shouldn't be too difficult to look one since such task is directly needed by various US company.

I should ask do you want the short answer or the longer one?

Short is $50-60 per PCB @ 10k a year, that is parts and all. Only variable is the IC's though and I can't guess that. The world is in the midst of a global supply chain issue with IC's atm.

[sidehack]

Yeah chip cost (and quantity) will be by far the biggest determining factor in BOM cost.

[Vantis]

Yeah chip cost (and quantity) will be by far the biggest determining factor in BOM cost.

Options would be
1. Find a chip on the market you can get (good luck they all sent out letters to major customers stating due to "waffer constraints", they can't make contracted shipments).
2. Find an old chip that is still efficient and try and just get hash rate by just pure volume because of cheaper manufacturing costs.
    a. Ask for it on tape and reel, I could handle a tray as well but perfer tape and reel.
    b. Find a PCB with the chip you want that is soldered and scrap or at scrap value and we can remove it. Probably a 5-10% scrap rate to do that. Each one would take about 3-5 minutes to remove so add that labor. 20 dollars an hour and we can remove one every three minutes = so add a $1.66 to each chip to remove it. Just rough est.