I agree ... there are a number of hosts in eastern Canada, if the one you are using isn't helpful I'd suggest you look for another host to move your miners to.

What exactly is the current condition? From the logs it looks like one of the three has 0 working hashboards (or maybe a dead PSU), and the others have one dead hashboard each. Do the two with single dead hashboards work at 2/3 of the expected hashrate?

Did the host try swapping a PSU on the one with 0 working boards?

What I'd suggest is:

1 - Swap PSUs between the miner with 0 working boards and one of the other ones.

2 - Take a working hashboard from one partially working miner and use it to replace the bad one in the other miner. You might need to use 3rd party firmware to get this to work (like asic.to).

With a bit of luck, that might get you 2 fully working miners back.

After trying that, I might be interested in buying whatever pieces you have left. I'm located in the northeast USA, so shipping shouldn't be too crazy.

Yeah, only dead ones. Anyone that has working boards probably won't have trouble finding a buyer if they want to sell, but I'm only interested in trying to repair the dead ones.

I am also interested in working PSUs, working control boards, and cases.

So if you want to part out an S17 with only a single working hashboard, I'd take everything but the working hashboard.

I guess it isn't very surprising given the rates they are charging. On average they probably can't spend more than 2 or 3 hours on a hashboard to break even.

I think the Bitmain training is probably pretty specific, focusing on the most likely issues. If the problem doesn't fall into the scope of their training, there isn't really any information out there on it. I've been going off the Bitmain repair manuals and then filling in the missing information by reverse-engineering the boards. Pretty time consuming ...

You'd probably be fine with the current case if you just remove one of the fan plates. There are some companies that specialize in making custom cold plates, I wonder what quantity you'd need to make the $ work. In hot climates, you might save 50-75W per miner.

75W would only be 657Kwh/year, so not likely to ROI on energy usage. That's only $33 a year for 5c power.

5 or more ASICs I can understand, but not sure why a boost circuit failure would make a board "unrepairable".

Even though they are unrepairable I'd still be interested in your dead hashboards. Even if I can't repair them, I can still use them for spare parts.

If one of you is willing to sell the other your working hashboard, I'd also be interested in the PSU, case, and control board left over.

Bitmain already announced a liquid-cooled miner, the AntRack, last year.

It was discussed in this thread: https://bitcointalk.org/index.php?topic=5287628.msg55547933#msg55547933

Also, I haven't personally seen the guts of a S19 yet, but I've been told that they no longer have individual heat sinks, but large heat sinks that spread across multiple chips. If that's the case then it could be possible to swap out the heatsinks for liquid cold plates to do a liquid-cooled conversion.

Full log would help, but I think it's probably a flaky chip or connection to a chip. I've seen this type of thing multiple times. It's also a pain in the ass to fix since a lot of times the issue is hard to reproduce on a test fixture.

Could you post the full log?

Are you using an APW9+? I think the power_api.c deals with the communication to the PSU, so if you are using an APW9 maybe that is what is causing the "power type version error".

That is more like it. Unfortunately, you've probably picked the absolute worst time to start. There's a shortage of new gear, and with the rise in price and not much new gear coming online the profitability has gone up a lot and looks like it will stay that way until the miner manufacturers start delivering more or the price goes down. Eventually, the manufactures will manage to get some fab capacity and will dump as many miners as they possibly can, which will reset profitability again.

But right now, people see the price going up, and profitability really high, and are paying prices for gear based on thinking the price is going to go even higher.

With your power cost you could be profitable with an S9 even if the difficulty doubled. And even though the price of used S9s have nearly 10Xed in the last 6 months, (I think they are up to  $250 per miner with psu), you'd be able to make back your investment in 3 months if conditions remain the same.

So I'd look at S9s or Avalon miners in the 60 to 80 W/TH range. Maybe Canaan Avalon 1066 would work, those you can still buy new, and they have a reputation for being higher quality. With power that cheap I don't think it makes sense to pay a premium for super-efficient gear.

Sorry to hear that, did they say what was wrong with the 2 bad boards?

Don't think you are going to have much luck finding used hashboards, if someone was selling them they'd probably want more for 2 hashboards than what you could buy a full 2nd hand miner for.

Also, if using factory firmware they are not really swappable. They match up the hashboards at the factory and then write settings into the eeprom, if the settings don't match up between the hashboards factory firmware just refuses to start. Aftermarket firmware just ignores all the settings from the hashboard, so it will work in more situations.

Ok, that is not extremely cheap. These days you are competing against industrial miners that are paying closer to 0.02 GPB/kwh. They are the ones buying the most equipment, so they drive the miner prices based on their profitability.

First step is to just use an online profitability calculator (like this) to see what the #s look like for any given miner, and how long it will take to break even.

For info on altcoins, head over to the altcoin section... https://bitcointalk.org/index.php?board=160.0

First off, define "extremely cheap", in GBP/Kwh or USD/Kwh.

If your power is cheap by mining standards, then you probably should not be looking at S19 pros that are currently selling for ridiculously high prices. Right now the price is so high for S19 pros on hand that even if your power was free, the BTC price stays steady, and diff doesn't go up it would take over a year to make back the investment in the miner.

New generation miners all require 200-240V, and >3000W per miner. An S17pro specs say 3250W, so around 13.5Amps at 240V. That puts it above what you can safely put on a 15A circuit (typically 80% of peak rated current for a continuous load), so you could either have your electrician friend wire up individual 20A circuits for each miner, or find a >40A pdu to put two miners on one circuit. It would probably be cheaper to get your electrician friend just put in a sub-panel and individual outlets than bother with PDUs.

Could you post a kernel log and a screenshot of the status page for each miner?

Did both messages start at the same time?

The hashboard can use a few different types of temperature sensors, but in a single miner all hashboards need the same type. So that is what the check is for, but I don't know exactly what sensor 8 means. If it was working before, then it suggests that the I2C interface to that hashboard is not working right or the temp sensor is failing. Sometimes the I2C interface can get screwed up by failing asic chips.

If both started happening at the same time, maybe it is a PSU or control board problem.

Yeah, also can't think of any way for a device on the network would do that, especially only to some of the miners. Do the miners go down when you turn the laptop back on?

Also might be interested in T17 ot T17+.

So now I'm more confident that your issue is the control board. I had this exact problem on a S17 a few months ago, the same one that gave me the "Malformed power response" error. I eventually had to replace the control board to get it to work reliably.

My suggestion is to try doing a factory reset and reflashing the firmware, then try a known good control board, and if that doesn't work try a known good PSU.

Looking for S17 series miners with dead hashboards.


If you've got multiple miners with dead hashboards, I can help you consolidate good boards into fully working miners and I'll buy the dead ones.


Send me a PM with what you've got and the condition, and I'll make you an offer.

Will pay BTC using forum escrow.

I'm pretty sure I've seen this error from a flaky control board before, so it might be worth trying another control board.


The 1.8V for each voltage domain is supplied by a regulator shown in the photo below, and then it is transmitted through each chip. The schematic in my previous post shows the pinouts for the regulators, and the photo from my previous post shows where the voltage can be measured at each chip. I have not worked on a T17+, but on other miners there is 2nd 0.8V regulator near the other side of the board that generates the 0.8V for the chips on that side of the board for the same voltage domain.