Chain 1 only find 6 asic, will power off hash board 1

[dm800pvr]

Hey guys.

On my t17 42 TH, it'll run just fine for 1 year, sometimes I've got 42 TH sometimes after one reboot 28TH which means 1 board is gone! I Checked PSU and control board and also check voltages on 3 cable of each board and all was ok so problem is seems not in PSU.

check_asic_number_with_power_on: Chain[1]: find 6 asic, times 0=>Chain 1 only find 6 asic, will power off hash board 1

Here is usefull log section as you can see on Chain1 only 6 asic founded and then power it off?!!

Code:

check_asic_number_with_power_on: Chain[0]: find 30 asic, times 0
driver-btm-api.c:1042:check_asic_number_with_power_on: Chain[1]: find 6 asic, times 0
driver-btm-api.c:1042:check_asic_number_with_power_on: Chain[1]: find 6 asic, times 1
driver-btm-api.c:1042:check_asic_number_with_power_on: Chain[1]: find 6 asic, times 2
driver-btm-api.c:1069:check_asic_number: Chain 1 only find 6 asic, will power off hash board 1
driver-btm-api.c:1042:check_asic_number_with_power_on: Chain[2]: find 30 asic, times 0

What should I do now? no warranty at the time.

Thanks in advance.

[favebook]

Hashboard does not power on unless it reads all chips from board. You are experiencing same problem as almost every single person who bought antminer 17 series (including me). I've had units for for more than year and then that happens.

What you are experiencing is dead hashboard or faulty PSU. You could try different PSU but I doubt it is the problem.

In some cases putting miner horizontally on one or other side fixes the problem temporarily, but only permanent solution is to fix the faulty hashboard(s) or RMA the unit.

I decided to RMA my S17+ as it started losing one hashboard like your unit and then after a week or two, it lost all 3 hashboards. From time to time it pops up and works fine with 1 or 2 hashboards but that is not how it is designed to work.

I'd suggest you to open a ticket at bitmain support and then ask for RMA. Repairs are not that expensive from my experience, but shipping might be due to human virus spreading around and usual prices are nowhere near cheap either.

[mikeywith]

If mine ran for a year I would be happy enough to throw it away, this is a good record and above average life-span.

Code:

driver-btm-api.c:1042:check_asic_number_with_power_on: Chain[1]: find 6 asic, times 2

the heatsink on chip no.7 has probably lost contact, the solder paste must have gone bad, this is pretty normal with all of these gears, at least 30% of hash boards die this way.

Fixing this problem is expensive, there is nothing really you can do, just remove that hash board and mine with the 2 good hash boards, there are a few things that you can try, most of them are mentioned in this topic and this topic.

Try them out, if nothing works, go ahead and remove that board since it actually slows down the boot time and might make your miner reboot every once in a while, which is something you don't want.

[dm800pvr]

But I do another test :

Changed the control unit cable and it changed the error now to hashboard 2! So it means I have weak amps on one of output (the faulty middle one) right? I will try to find another PSU to test and will post the result here!

https://i.ibb.co/jv1Yqwp/t17.jpg

Normal Mode:

1=>CHIPA
2=>CHIPB
3=>CHIPC

Chain 1 only find 6 asic, will power off hash board 1.

What I did:

1=>CHIPB
2=>CHIPA
3=>CHIPC

Chain 2 only find 6 asic, will power off hash board 2.

I think it may be power issue if you see the logs it detects 6 unit and not enough power to boot other asics? 2 is faulty I think however it has volatages it may not produce enough amps to power all board asics.

[wndsnb]

If you have another PSU to try, then that's a good/easy thing to try, but I don't think the test you did indicates a PSU problem.

If I'm understanding what you did correctly, you originally had board 1 connected to chain 1, then you swapped the control cables so that board one is connected to chain 2. The problem stayed with the board, suggesting that the issue is in that board. Although it could just be that that board just happens to be more sensitive to a PSU problem than the others.

Maybe try running one hashboard at a time by disconnecting the control cables for the other boards? I'm guessing the same board will fail.

[mikeywith]

This confirms my theory, even more, I think you are confusing yourself, the miner only unders 1,2,3 which is the 18pins attached to the control board, clear? now that we got this out of the way, you should understand that since:

Chain 1 was in location 1 = chain 1 is bad
Chain 1 moved to location 2 = chain 2 is bad

this means chain 1 (what you call chip) is bad!

If the result was:

Chain 1 was in location 1 = chain 1 is bad
Chain 1 moved to location 2 = chain 1 is bad

then chain 1 is good and the problem is the PSU or the control board, but this isn't the case.

if you are confused still, do this.

Remove chain B and C , keep only chain A and do 3 tests

- 1 > ChainA
- 2 > ChainA
- 3 > ChainA

The results will be:

-Chain 0 has 6 Asics
-Chain 1 has 6 Asics
-Chain  2 has 6 Asics

[dm800pvr]

I confirm that you are absolutely right. How much is each chip price i see some videos that one faulty chip on a hash board could cause problem.

[mikeywith]

How much is each chip price

Not sure if T17 uses BM1396AB or another chip, but the price should be in the 5-10$ range per chip, but that isn't the issue, the issue is how are you going to replace the chip? you will need a professional to do it, another potential problem is it's unlikely to get 1 chip order, you might need to get a whole strip of 10 chips, plus shipping and all of that, it makes fixing the hash board expensive enough to just forget about it, but of course, do your own math.

T17 40th goes for about $370 in china (used), so my guess that a used hash board would be 20-25% of that? so $70-90 is my guess.

i see some videos that one faulty chip on a hash board could cause problem...

Yes the way these boards are made, 1 bad chip stops 29 other chips.