S17 : WARN: chain[1] - 0 of 48 chips detected

[nordmann666]

hello,

bought used and got this on kernel log:

Code:

[2020/02/20 23:20:42] INFO: Starting FPGA queue
[2020/02/20 23:20:42] INFO: Initializing hashboards
[2020/02/20 23:20:43] INFO: chain[0] - Initializing
[2020/02/20 23:20:55] INFO: chain[0] - 48 chips detected
[2020/02/20 23:20:58] INFO: chain[1] - Initializing
[2020/02/20 23:21:10] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:11] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:12] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:12] WARN: chain[1] - 0 of 48 chips detected
[2020/02/20 23:21:22] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:23] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:24] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:24] WARN: chain[1] - 0 of 48 chips detected
[2020/02/20 23:21:33] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:34] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:35] ERROR: src/chip139x.c:957 chain[1] - Failed to receive chip replies
[2020/02/20 23:21:35] WARN: chain[1] - 0 of 48 chips detected
[2020/02/20 23:21:35] ERROR: driver-btm-s17.c:1189 chain[1] - Failed to detect ASIC chips
[2020/02/20 23:21:35] INFO: chain[1] - Disabling the chain
[2020/02/20 23:21:36] ERROR: driver-btm-s17.c:1343 chain[1] - Failed to init the chain

the original bitmain firmware dont show the chain anymore and awesome miner firmware shows:

Code:

2	0	0	18.5	15	0.00	0.00	0	0-0-0-0	0-0-0-0	Failed to detect ASIC chips

any hints?

[1715229864]

1715229864

[1715229864]

1715229864

[mikeywith]

First of all, always code the kernel log by using the hash/number sign in the toolbar.

Code:

[2020/02/20 23:21:35] ERROR: driver-btm-s17.c:1189 chain[1] - Failed to detect ASIC chips

The error is quite clear, the board is bad, sadly!

To break it down, there is at least a 90% chance that this board is dead and will no longer function, however very seldom there might be other reasons for the kernel to through this error.

A faulty PSU or a bad connection to the control board, you need to first check the ribbon cable that goes from the control board to hash board number 1, it might be loose, however, the best practice would be swapping the cable with a working one, ribbon to board 2 goes to board 1, and vice versa, and the same thing with the PSU, you need to switch the hash board positions.

Expected results:

1-The miner now finds 48 chips on chain 1 and 0 on chips on chain 2 (The hash board is dead)
2-The miner now finds 0 chips on  chain 2 and 48 chips on chain 1   (Ribbon cable/Psu connection is faulty)

very seldom this error might come from a bad firmware, a hard reset might (chances are slim) fix your issue, you should also post the complete kernel log so we know if the other boards work fine. If all boards are bad then this could be a control board related issue.

Meanwhile, you should start contacting whoever you bought it from, let them know what's wrong and that you are going to troubleshoot it using the methods I explained, if the return of product is an option, you should go for it as soon as possible.

[philipma1957]

remember the s17 has been out for about 10 months.

and it sounds like yours was overclocked since you mention aftermarket firmware.

so there is a very good chance you got two out of three.

chain 1 is a middle board

0
1
2

middle burns tend to die faster since the center board always gets less wind from the fans du to the fan having a center hub.

try for a return.
if not try for a 40% refund and be willing to settle for a 33% refund.

[BitMaxz]

Did you test to run it at low power mode? Maybe the chain 1 can't run at the higher power that is why it's not showing and getting zero chips from awesomeminer.

How about revert it back to the original firmware and test it with low power mode instead? A quick test will tell us if the middle hashboard is totally dead or if it can still run at low power mode.

In case you don't know where you can find the low power mode you can find it under Miner's Configuration tab on the WebGUI.

[mikeywith]

if not try for a 40% refund and be willing to settle for a 33% refund.

At 33% it means he bought only 3 hashbaords without a PSU, Control Board, Frame, 4 fans, etc. I think the best he could hope for is 25% assuming every hash board makes 25% (75% in total) and the rest of the components make the remaining 25%.

If the seller is honest and nice enough, they would give them more than that considering the downtime and the efforts in troubleshooting, but I don't see that happening because if the seller was honest they would have tested the machine before sending it out.

[nordmann666]

i checked all cables now and reseat them...the miner was at bitmain allready and look for yourself how they repair

cable ties?! are the original cables from the fans not long enough?
https://imgur.com/a/Vp1A4DQ

missing screws
https://imgur.com/a/LoSFhXZ


now all boards are showing up with all asics, but next problem is, that in normal mode the miner dont find temp sensor and if he did the temps are 100 degree and more in normal mode
in power save mode
https://imgur.com/a/HAQ15oF

and hw errors are also...i dont know...way to high


i dont wanna blame the seller...it´s a nice guy and i believe he was also suprised with this problems i have. but yes...if it need long time to get a repair i will have to talk seriously about refund :/

[philipma1957]

what makes you think this miner was repaired by bitmain?
i have seen bitmain use fan wire extensions
and cable tie tight is a good idea since gear vibrates

not sure the missing screw photo is correct.
but all my gear is 70 miles from here so i cant check that.

 
also i cant see your third image.  what are too many hardware errors.

i want to see the gui for the low speed setting.

fan speeds and hard ware and hash rate.

[nordmann666]

6000HW errors on all chains after running one hour.

thats to much for original fw without any modification on frequence or voltage.

[philipma1957]

6000HW errors on all chains after running one hour.

thats to much for original fw without any modification on frequence or voltage.

that is a meaningless number.

what is the error percent?

0.0050 is acceptable.

that would mean you get 9999.50 of every 10000 earned.

[mikeywith]

that is a meaningless number...

They may be not very accurate, by in most cases you can simply tell by looking at the number of HW without even looking at the percentage, 6000hw after running for just one hour is pretty bad, I am sure the % is pretty high, with that being said, I still agree, he must check the HW%

OP, the % is located in the bottom row under Diffa#, check the first column to the left where it has the number of the pool, the last row will "hw" follow it to the right side right under DIFFA# and it will be a number and a % sign.

Edit:

Here is an image to save the day

[philipma1957]

I have a t17   it has 66,000 hw's in 6 hours

it is 0.0285%

this would be   9997.22 out of 10,000 usd here is a screen shot

I can live with losing under 3 dollars of every 10,000

so this unit's hourly error rate is 11,000 for 40th

it is almost 2x that of the op

and it loses hardly any money his unit is losing maybe $1.50 out of every 10,000