S9i one board doesnt work

[mortadeli]

Hi there,

one of my S9i units' boards doesnt work, temp constantly stays at 15. However sometimes after a couple of hours, that one board randomly starts working again. Same process repeats if unit is rebooted.

I am running latest S9i firmware, for some reason I cant downgrade firmware anymore to test. I have tried resetting as well.

I have also tried changing ethernet cables, or using another PSU.

any advice is appreciated

Thanks in advance

Code:

Miner Type	Antminer S9i
Hostname	antMiner
Model	GNU/Linux
Hardware Version	30.0.1.3
Kernel Version	Linux 3.14.0-xilinx-g16220c3 #83 SMP PREEMPT Thu Jul 12 11:42:53 CST 2018
File System Version	Mon Mar 4 11:45:08 CST 2019
Logic Version	V2.57
BMminer Version	2.0.0
Uptime	1:14
Load Average	0.18, 0.19, 0.19

https://imgur.com/a/ZN8MZmg
https://imgur.com/a/XIS5ymD

[1714615195]

1714615195

[1714615195]

1714615195

[euhctcyn]

Based on what you have already tried out, the only thing you have left is to try to replace the 18-pin data cable that connects the controller board to the affected hashboard.

https://imgur.com/a/97bD1Gs

If you don't have a spare one (I recommend always having, just in case) you can turn off the miner, disconnect the PSU and swap one of the working ones from the other two boards into the affected one to check if it's alive. If after doing so the board still doesn't work that probably means it's dead and you will have to replace the whole hashboard for a new one  

[Artemis3]

Based on what you have already tried out, the only thing you have left is to try to replace the 18-pin data cable that connects the controller board to the affected hashboard.

https://imgur.com/a/97bD1Gs

If you don't have a spare one (I recommend always having, just in case) you can turn off the miner, disconnect the PSU and swap one of the working ones from the other two boards into the affected one to check if it's alive. If after doing so the board still doesn't work that probably means it's dead and you will have to replace the whole hashboard for a new one  

Before assuming this, test. Simple: disconnect all the other boards and swap cables.

Test each board one by one, and swap data cables one by one, and power cables one by one on all three boards while single connected. Then you will know for sure if the data cable is the culprit. Takes time, but gives you the right answer instead of wild guessing.

[BitMaxz]

Upgrading or downgrading the firmware or factory reset and replace ribbon wire(18 pins)doesn't work you have a faulty temp sensor that is why it gives you 15 temp.

Can you try to increase the fan speed and lower the frequency if you don't know how to do this you can check this method from here https://bitcointalk.org/index.php?topic=2690140.0 you can find the frequency in the same tab.

If still doesn't work you have a hardware issue the only solution is to replace the temp sensor from the hashboard.

[mikeywith]

for some reason I cant downgrade firmware anymore to test. I have tried resetting as well.

then this should be the first problem you need to solve, usually changing the firmware alone can fix a lot of issues, so you should explain what does it say when you try to change the firmware?

[mortadeli]

then this should be the first problem you need to solve, usually changing the firmware alone can fix a lot of issues, so you should explain what does it say when you try to change the firmware?

It says "Cannot find Signature", I've tried loading the mac fix fw upgrade first as well and it gives me the same error message.

Before assuming this, test. Simple: disconnect all the other boards and swap cables.

Test each board one by one, and swap data cables one by one, and power cables one by one on all three boards while single connected. Then you will know for sure if the data cable is the culprit. Takes time, but gives you the right answer instead of wild guessing.

I tried this and the data/power cables work on the other two boards, so its not the cables.

Upgrading or downgrading the firmware or factory reset and replace ribbon wire(18 pins)doesn't work you have a faulty temp sensor that is why it gives you 15 temp.

Can you try to increase the fan speed and lower the frequency if you don't know how to do this you can check this method from here https://bitcointalk.org/index.php?topic=2690140.0 you can find the frequency in the same tab.

If still doesn't work you have a hardware issue the only solution is to replace the temp sensor from the hashboard.

Thanks. I tried this as well, isolated only trying to run with the board that gives issues, and the board no longer loads at all when speed and freq are modified, had to reset settings. Miner status pages was showing all values zero.

Will need to look into replacing the temp sensor then, any info/tips on how it should be done?

[Artemis3]

Will need to look into replacing the temp sensor then, any info/tips on how it should be done?

If you think only the sensor is the problem, while having only that single board connected, try booting from Braiins OS on an SD card. There is a cgminer option that lets you disable the sensors when you use this alternative free open source firmware replacement.

If you use this method, you only need an sd card and move a jumper. To go back just take out the sd card and put the jumper back where it was.

If you really like Braiins OS, you can install it in the controller's NAND as well.

You can also read the Braiins OS thread if you want more information.

Even if you don't like it you can go back later to Bitmain's firmware, after you confirm the problem really is just a damaged sensor.

[BitMaxz]

It says "Cannot find Signature", I've tried loading the mac fix fw upgrade first as well and it gives me the same error message.

This is a common issue right now if you flash the miner already with the latest version(2019 firmware) and then you are trying to downgrade you can see this error "cannot find Signature" there's a method to bypass this error that you can find here but the problem the 2019 firmware disabled SSH access.

So the only solution that I think might solve your issue is to flash the control board through SD card. Recently Bitmain decided to release a method flashing S9s through SD card.

If you don't know how to do it you can follow it from this link below.

- S9 series (S9, S9i, S9j, S9 Hydro) Control Board Program Recovery.

[Artemis3]

Ugh, this is awful. Can you downgrade from the 2019 firmware back to 2018? Disabling SSH is enough reason to stay away from Bitmain software.

[BitMaxz]

Ugh, this is awful. Can you downgrade from the 2019 firmware back to 2018? Disabling SSH is enough reason to stay away from Bitmain software...

If you successfully flash it with t9+ firmware as the procedure from this you might revert the firmware back to any s9 firmware through web GUI.

So you need to follow the method first from bitmain creating your own control board program recovery then flash the control board and after successful open your web gui upgrade the firmware to old version of s9 most stable firmware is "Antminer-S9-all-201711171757-autofreq-user-Update2UBI-NF.tar.gz".

Let's hope that you can upgrade the firmware from t9+ to antminer s9 firmware if it won't work the final solution that might work is to flash it with Braiin OS which you already posted above.