New S19k Pro Stopped Working after 1 week

[tuaris]

It was working fine for about 4 days.  Put it on sleep mode about a day ago, and turned it back on.  Now, it's giving this:

Code:

2025-01-08 17:51:00 no board pluged in socket 432
...
2025-01-08 17:51:51 Error, fan lost, only find 3 (< 4)
2025-01-08 17:51:51 fan_id = 0, fan_speed = 7000
2025-01-08 17:51:51 fan_id = 1, fan_speed = 3440
2025-01-08 17:51:51 fan_id = 2, fan_speed = 5168
2025-01-08 17:51:51 fan_id = 3, fan_speed = 6020
2025-01-08 17:51:51 Sweep error string = F:1.
2025-01-08 17:51:51 ERROR_FAN_LOST: fan lost
2025-01-08 17:51:51 stop_mining: fan lost
2025-01-08 17:51:51 cancel thread
2025-01-08 17:51:51 ****power off hashboard****
2025-01-08 17:51:51 power off
2025-01-08 17:51:52 ERROR_SOC_INIT: soc init failed!
2025-01-08 17:51:52 stop_mining: soc init failed!
2025-01-08 17:51:52 ****power off hashboard****
2025-01-08 17:51:52 power off

Also noticed this in the historical log

Code:

2025-01-05 19:04:20 chain[2] asic[5] [215] [0]
2025-01-05 19:04:20 chain[2] bad asic num: 6 6 0
2025-01-05 19:04:20 chain[2] asic[5] [215] 28651771-28651556 [0] 62-62

What happened? The room temperature never exceeds 75F, and the current room temp is 61F.  Other S19k's still humming along nicely.
Did I receive a bad unit?

Firmware Version Mon Mar 4 16:03:40 CST 2024
Already tried the basic troubleshooting (reboot, factory reset, power cycle)

[tuaris]

Applied a firmware update (Antminer-S19k-Pro-merge-release-20240409060820.bmu), now I get:

Code:

2025-01-08 18:42:54 pulse_mode = 1, ccdly_sel = 0, pwth_sel = 4
2025-01-08 18:42:54 start do_core_reset
2025-01-08 18:43:06 end do_core_reset
2025-01-08 18:43:06 noraml inc freq voltage
2025-01-08 18:43:06 set_voltage_by_steps to 1500.
2025-01-08 18:43:10 Version num 65536
2025-01-08 18:43:10 Mask num 0x1fffe000
2025-01-08 18:43:10 opt_multi_version = 65536, interval timeout = 2620657
2025-01-08 18:43:10 freq = 590, percent = 90, hcn = 4442, timeout = 2620657
2025-01-08 18:43:10 (null)
2025-01-08 18:43:21 freq 200 waiting board temp to 35,curr temp 24
2025-01-08 18:43:23 Error, fan lost, only find 3 (< 4)
2025-01-08 18:43:23 fan_id = 0, fan_speed = 1990
2025-01-08 18:43:23 fan_id = 1, fan_speed = 0
2025-01-08 18:43:23 fan_id = 2, fan_speed = 2400
2025-01-08 18:43:23 fan_id = 3, fan_speed = 2400
2025-01-08 18:43:23 Sweep error string = F:1.
2025-01-08 18:43:23 ERROR_FAN_LOST: fan lost
2025-01-08 18:43:23 stop_mining: fan lost
2025-01-08 18:43:23 uninit_temp_info
2025-01-08 18:43:23 do not read temp anymore...
2025-01-08 18:43:23 cancel thread
2025-01-08 18:43:23 cancel thread
2025-01-08 18:43:23 ****power off hashboard****
2025-01-08 18:43:23 power off
2025-01-08 18:43:25 temp monitor thread exit
2025-01-08 18:43:31 freq 200 waiting board temp to 35,curr temp 24
2025-01-08 18:43:42 Sweep error string = P:2.
2025-01-08 18:43:42 ERROR_TEMP_TOO_LOW: protect machine!
2025-01-08 18:43:42 stop_mining: protect machine!
2025-01-08 18:43:42 ****power off hashboard****
2025-01-08 18:43:42 power off

So I got a unit with a bad fan? Is the ASIC fine?

[philipma1957]

you lost a fan it happens. buy a fan and replace it.


https://altairtech.io/product/antminer-120mm-replacement-fan-for-s19-s17-s11-s15-and-s9-series/

make sure  you get a square pin if the bad fan has a square pin

make sure you get a rectangle pin if the bad fan has a rectangular pin.

[tuaris]

you lost a fan it happens. buy a fan and replace it.


https://altairtech.io/product/antminer-120mm-replacement-fan-for-s19-s17-s11-s15-and-s9-series/

make sure  you get a square pin if the bad fan has a square pin

make sure you get a rectangle pin if the bad fan has a rectangular pin.

Hmm, maybe just getting the fan is less trouble than trying to work with the vendor I purchased from.

I wonder... I have a ton of spare fans from L3's and S9's.   Will they work here?

[philipma1957]

you lost a fan it happens. buy a fan and replace it.


https://altairtech.io/product/antminer-120mm-replacement-fan-for-s19-s17-s11-s15-and-s9-series/

make sure  you get a square pin if the bad fan has a square pin

make sure you get a rectangle pin if the bad fan has a rectangular pin.

Hmm, maybe just getting the fan is less trouble than trying to work with the vendor I purchased from.

I wonder... I have a ton of spare fans from L3's and S9's.   Will they work here?

maybe the s9 fan is fast enough. 

if it tops out at 4500 rpm it will likely not work but if it goes to 500 or more it should be good

[tuaris]

It's surprisingly difficult to figure out what fan is what and there is a serious lack of information about it.

According to Bitmain

J12 = fan_id = 3
J13 = fan_id = 2
J15 = fan_id = 0
J14 = fan_id = 1

But I have J11, J12, J13, J14, and according to the labels on the control board

J11 = Fan 1
J12 = Fan 3
J13 = Fan 2
J14 = Fan 4



So do I assume Fan 2 = fan_id = 1 = J13 = Rear0 = bottom back?

UPDATE: trail & error results:

J11 = fan_id = 1 = Fan 1 = Front0 = bottom front
J12 = fan_id = 0 = Fan 3 = Front1 = top front
J13 = fan_id = 3 = Fan 2 = Rear0 = bottom back
J14 = fan_id = 2 = Fan 4 = Rear1 = top back

[BitMaxz]

It seems according to your logs it is a front0 fan that doesn't have RPM
So there's a slight difference from the result you edited above.

fan_id = 0 = Front0 = top front
fan_id = 1 = Front1 = bottom front

If I'm wrong, you can also try to check the right location of the fan under the miners status, or a better way is to remove the fans one by one and then look for the logs to see if there's a change on the RPM to all fans so that you know which fan exactly is the broken one.

[danieleither]

With even a handful of miners it is essential to carry spares. It is not worth returning a miner for faulty fan or PSU, because they are so cheap and you'll lose more in downtime, plus the risk of damage in the post.

I have approx. 200 miners total across 5 small farms. I reckon I get through at least 50 fans and 5 PSU's each year, and 1 or 2 control boards.

I'd recommend to keep as spares:

1 PSU for each 20 miners (with min 1 in stock at all times)
1 control board per 50 miners (with min 1 in stock at all times)
1 fan per 2 miners

I buy all my spares from ZeusBTC (import from China).

[philipma1957]

With even a handful of miners it is essential to carry spares. It is not worth returning a miner for faulty fan or PSU, because they are so cheap and you'll lose more in downtime, plus the risk of damage in the post.

I have approx. 200 miners total across 5 small farms. I reckon I get through at least 50 fans and 5 PSU's each year, and 1 or 2 control boards.

I'd recommend to keep as spares:

1 PSU for each 20 miners (with min 1 in stock at all times)
1 control board per 50 miners (with min 1 in stock at all times)
1 fan per 2 miners

I buy all my spares from ZeusBTC (import from China).

Yeah they toss fans like mad

so its good to have them.

once you pull a dead fan and you are certain it is dead.

cut the power cable from it as close to the fan as possible and save the four wires attach to the connectors.

you wont mistakenly try the dead fan again. and once in a blue moon you may need the four wires and the white connector

[mikeywith]

With even a handful of miners it is essential to carry spares. It is not worth returning a miner for faulty fan or PSU, because they are so cheap and you'll lose more in downtime, plus the risk of damage in the post.

I have approx. 200 miners total across 5 small farms. I reckon I get through at least 50 fans and 5 PSU's each year, and 1 or 2 control boards.

I'd recommend to keep as spares:

1 PSU for each 20 miners (with min 1 in stock at all times)
1 control board per 50 miners (with min 1 in stock at all times)
1 fan per 2 miners

I buy all my spares from ZeusBTC (import from China).

i second this, and then eventually, when your miners have run for a while you won't even have to buy new spares, just interchange/swap parts, I have dozen of dead miners that i can use, some have a good PSU, some have good fans, no need to buy anything.