S17+ (70TH) Hashchain missing (but not always)

[jakaba]

I have a brand new S17+ (70TH). From time to time Hashchain 0 is missing.

I found the following entry within my logs:

Code:

2020-09-10 11:15:27 driver-btm-api.c:1155:check_asic_number: Chain 0 only find 0 asic, will power off hash board 0

I have heard that this may be a common problem on the S17 series. What I don't understand though, is that this problem seems to come and go randomly.

When I turn off the miner and turn it on the next day it can occur. It can remain like this for a couple of days. Then without any clear indication why, all three Hashchains are found and working again. (Without any problems).

The problem either appears or disappears only when I turn off the miner and turn it on the next day. Once it's running (with 2 or 3 boards), it stays like that.

I turn the miner off every night (going to sleep mode first, too cool down). The past 6 days the miner was started every morning without any problems. Today the problem occured again. A number of reboots, power disconnects did not solve the problem (like the last times).

All temperatures are within a normal operating range.

Things I have tried:

• different power cables & outlets
• switching data cables (results in other hashboard number not being found, depending on how i switch them)
• Install Control Board Recovery Image: https://support.bitmain.com/hc/en-us/articles/360033757513-S17-S17Pro-S9-SE-S9k-Z11-control-board-program-recovery-SD-card-flashing-with-customized-PW-
• Reset Miner using web-intercace & hardware reset button

The thing is I don't really want to send it in for repair, as it seems all parts of the hardware (can) function properly. (and it will take forever).

What may be the cause of this, and is there a potential fix?

[1714693590]

1714693590

[1714693590]

1714693590

[1714693590]

1714693590

[1714693590]

1714693590

[favebook]

Firstly, yes, that is usual thing for 17th series of Antminers, and you have a sign of dying hashboard. Lose heatsink is most probable cause and there is not much if anything that you can do.

You can try putting it on one of the sides and gravity might help you. But since it is occasionally working fine, I'd recommend you to turn it on and when you see it having three boards up, do not turn it off EVER again. And pray to electricity gods that you won't have blackouts or brownouts.

Secondly, I am not sure why are you turning it off every night nor why are you "letting it cool down". There is not much benefit to second thing you are doing and first thing is obvious, you lose up to or more than 12h per day of your earnings. That way, you will never get ROI on it, let alone earn anything after it, especially if you do not have free electricity.

Thirdly, I'd like to give you few more advice.
1. When hashboard drops, do not change anything in config, just hit apply or use reboot button on web interface to restart it. It will be faster than full reboot and it will save you some time.

2. If you are up to it, you can try custom firmwares (they MIGHT help you if you underclock it and in some cases overvolt or undervolt just that hashboard that you are having problems with)

I recommend Taserz/Vnish or thierry firmware for 17th generation. You could try Braiins from slushpool, but I have never tried it so I cannot recommend or say anything about it.

[jakaba]

So if your diagnosis is right, it seems there is a serious hardware error. I was hoping there wasn't (like some loose cable on the PSU or something). In that case I suppose sending it in for repair and hoping for the best is the smartest move?

Putting it on the side actually did work for me before. I read about this somewhere in this forum. Tried it again today, this time without success.

The nature of my project does not allow for constant electricity supply. I do have free electricity from time to time, but at other times its rather expensive. The miner is supposed to be installed inside the tower of a windmill.

Thats why I am coding a control software that will tell my miner what to do, based on the current price of electricity (if selling the energy is more profitable I won't mine). By my calculations it will be able to run profitable about 70% of the time though, which should result in profit long term.

Regarding your tips:

1. time saving is not important right now. This is a research project and I need a functioning S17+ to do proper research. Profitability is not relevant (yet).
2. I am running BraiinsOS on all our S9s (12 in total). It has not released yet for S17+ (but will be this year). I am planning to use it for this S17+.

I have also tried Awesome Miner Firmware. It has the same issues (missing hashchain). Setting a low power profile did not resolve this.


The retailer that sold me the S17+ offered me to send in the defective Hashboard and get a (used but working) Hashboard as replacement.

I think I will try this. Even though I am unsure about potential issues when replacing a Hashboard with another one.

[philipma1957]

since it is not about profit you could send it in.  or check on thierry4wd to use his firmware for it.

see below
I have the same issue with my 63th t17+. middle board drops out.

I now run it with stock  freq but 17.65 volts vs 18.28 that stock firmware uses.

yeah. it gets around 62th not 63th but the 3 boards work.

[jakaba]

Thank you for recommendations.

Before I run out of my 180 Day warranty I will send in the faulty hashboard.
I read that a different board my not work with my S17+ as it has a different "board-code"

But sometimes it will work. I will give it a try and report back when I receive the "new" board.

[favebook]

It is a serious problem, and as I said, there is not much you can do to fix it unless you are well versed in repairing them.

Bitmain will never take only one hashboard for repair, you will have to send a whole unit (might be expensive due to shipping across half the world - assuming you are from USA). Also, if you decide to go through with only sending one hashboard to your reseller, be careful not to send that unit at a later time to Bitmain as I think they will count it as voided warranty if all three hashboards are not from same miner.

It matters on which side you put it. Big heatsinks should be on top as they are usually the ones that are loose.

Good luck with your project, report here in new thread when you finish it if you are up for sharing it.

Try Thierry's or Taserz's firmware, I highly recommend both. First one has lesser fees but less features, second one has higher fees but a lot of features (you might find useful).

[mikeywith]

The retailer that sold me the S17+ offered me to send in the defective Hashboard and get a (used but working) Hashboard as replacement.

I think I will try this. Even though I am unsure about potential issues when replacing a Hashboard with another one.

You should accept the offer, the retailer can't send you a worse hashboard than the one you have unless if it doesn't work at all, the problem with your hashboard is very hard to fix, it isn't a matter of replacing a wire or a fan, you will need to locate the bad chip/s, remove them and then re-solder them on the board.

That might seem like an easy thing to do but it isn't, if you are lucky, putting some pressure on each of the heatsinks will be enough to take the bad chip off the hashboard (the bad chips on all the 17 series aren't exactly burned, they just have terrible connectivity due to cheap/bad solder paste) but that only solves half of the problem, the other half would be re-soldering that chip/heatsink back, and that is nowhere near easy and require a good level of skill and a set of tools.

In many cases, the heatsink/chip is indeed loose but not loose enough to fall under pressure, so you will need a fixture tool like this one just to know what chip/chips which need removing, and that tool isn't cheap by any means, knowing the bad chip/heat sink isn't the end, it's only the beginning, you will need a heat gun to remove it, a bit too much of heating and you could damage the board itself or make another good chip/heatsink lose contact to the board and thus create even more problems.

As long as the chip/heatsink hasn't completely fallen, then it all depends on your luck, it could run for a week then drop for a day then come back, it could do just fine for a few months, it could fall next week and never work again, so getting a replacement is the BEST option you have.

[jakaba]

UPDATE:

A few weeks ago the retailer send back my defective hashboard.
They put my hashboard into a different S17+. This caused the miner to not even start hashing on any of the boards.
From what I understand this means that my hashboard cannot work together with hashboards from a different S17+.

Therefore I am left with 2 options:

• 1. Do nothing and have my miner run with 2 "new" hashboards
• 2. Send my miner in, and receive a used S17+ as replacement

Option 1 will leave me with 2/3 of a new S17+
Option 2 will leave me with 3/3 of a used S17+

I am not really sure whats better, given that 3/3 of a used S17+ may have a shorter lifetime compared to 2/3 of a new S17+.


Any thoughts or recommendations? I am eligible for a used replacement miner  until December 2020.

[philipma1957]

mine with 2 boards until Nov 30th. then make the decision of getting replacement.

[NotFuzzyWarm]

To me it would depend on if the used miner has a full 90-day or whatever is standard BM warranty. If it has full warranty then I'd replace the miner.

[favebook]

I wouldn't say that used one may have shorter lifespan. It's actually luck based and your new one might break in a month while used one keeps working for years to come. You cannot really predict it.

I'd personally go with 3/3 of used one if it has been used in normal conditions and if it is 100% working stable for 48h+. Otherwise I'd go with Phil's suggestion. Also, if new one has warranty is an additional game changer.

Since beginning of cold weather, my S17+ started acting up as well. I kept loosing the same two boards over and over again after 12h+ of working perfectly every day. When I looked into him, I checked data cables, tried different combinations, checked heatsinks and found that only 2 specific hashboards are making problems even thought they never went past 80C nor have they been making problems for previous 1 year. Therefore, I went to check PSU and found that one PSU fan is not even turning. My assumption is that PSU is making my boards fail. But I am not sure why would it target only 2 specific ones (even though I tried putting them in different data cable ports as well as different PSU clips. I will revert back once I change PSU if anyone is interested.

[mikeywith]

A few weeks ago the retailer send back my defective hashboard.
They put my hashboard into a different S17+. This caused the miner to not even start hashing on any of the boards.
From what I understand this means that my hashboard cannot work together with hashboards from a different S17+.

I didn't know these S17+ had the compatibility issues, I encountered the same issue with the S9k, there are about 4 types hashboards, labeled as 1,2,3,4, so you only have a 25% chance of compatibility between different miners, which SUCKS.

Anyway, I am with favebook on this one, go with the 3/3 used, there is NO guarantee whatsoever that the NEW miner will mine any longer than the used one, i'l take the extra 20terahash with no questions asked.

Also, I wonder why doesn't your retailer get a tool like this https://www.zeusbtc.com/RepairDetails.asp?ID=236 to fix the compatibility issue, it's kind of expensive but it would save them a lot of money.

I will revert back once I change PSU if anyone is interested.

I am interested, please update.

[favebook]

Changing PSU did not help. A single board that sometimes had 0 chips instantly started hashing, while other two were not hashing like on old PSU but after a single restart it disappeared as well. 0 out of 65 chips read on all boards.

PSU that I put in now was new (ordered from Bitmain) and was tested on working unit, which means that it is not a problem. Control board cannot be the problem as I have tested it on working unit as well. Data cables I am yet to test but I doubt they are the problem.

I will quote this message in THIS THREAD as that is where it might help people. I will provide pictures of a single hashboard which I think started the chain reaction of all boards going down.