Bitcoin Forum

Hi, I have searched on the forum and internet for this topic and, although it seems to be a common issue, there is no proved solution. I'm using:

- Antminer S9 13.5 Th/s
- PSU AWP3++ Power Supply
- Hardware version: 16.8.1.3
- Kernel: Linux 3.14.0-xilinx-ge8a2f71-dirty #82 SMP PREEMPT Tue May 16 19:49:53 CST 2017
- File System: Fri Nov 17 17:57:49 CST 2017 (just updated)

My issue is that after 3 or 4 hours of working one of the boards is missing from the Status table and the hashrate is 2/3 (9.3 TH/s more or less). I can confirm this "change" with the mining pool monitor (currently using Slush pool). If I reset the antminer, all boards are listed and everything seems to be ok.

In some threads people say that by changing the AWP3++ to a non-Bitmain 1600w PSU it fixes the issue. Is this true? (if so, which one do you recommend?)

Others say that by sending to warranty the board (or the entire unit) it would fix it, but some people say that the issue persists.

So, I'd like to know if someone has a solution for this or can guide me for troubleshooting this

Thanks!!!!

Presuming your running using 220V power, you likely just have a flaky hash card.  I've lost at least one hash card in 11 of my ~150 S9s, and I'm told I'm doing better than most.

Yes, I'm using a 220v line and two questions for you: Which PSU are you using? and Which pool do you recommend? Thanks

You just updated to the worst firmware bitmain has ever put out for the S9, id recommend rolling it back if you are having any issues.

I use APW3++s for S9s and 841s, Sorcerers for 821s and 841s.

Highly recommended kano.is as a pool, or if you insist on a fancier UI/Higher fee/lower variance - slushpool.

okay, what is so bad about this firmware?
What is currently the best firmware to upgrade to? I have a S9 as well that has one board gone, and checked: its the same fw..

Just a lot of reported problems with the auto-freq November firmware.  There was a previous release on the Bitmain site, but that have pulled it.

You can see if your controller will accept the 650 fixed frequency firmware, then use the minerAdvanced.cgi page to tweak its performance.  Not all controllers will accept that firmware however.

The other thing to do is to pull your system apart, pull the boards, and blow them clean.  Sometimes a card will stop simply because the airflow is bad (I recently recovered a bunch of cards just by cleaning them - all had small moths and other bug parts blocking the leading edge of the heat sinks).

To be honest with you, changing the PSU wont do anything for the problem your describing.

Your Hash board is failing and will die soon. It will work for a bit then die, then all together not come back online.

Be prepared to RMA the S9.

This is from experience on systems I have sent back.

did u found a fix? i have the same problem, 1 hashboard on 2 antminer stopped showing up in the miner status page, i was traveling and leave them unplugged, i came back plugged them back and they have this problem..
does anyone have a previous firmware i can try ? mines are 13.0ths

I have one S9 where one of the boards stopped hashing overnight.  All the ASIC chips still look good.  Here's a pic:

http://www.cstone.net/~dk/s9_7_18-18.JPG (http://www.cstone.net/~dk/s9_7_18-18.JPG)

So the board is still there, it is just not hashing.  I rebooted that unit, but that one board still would not start hashing.  Board dying?

This is a 13.5T model, got it from Bitmain in December.  Never messed with the firmware.

Try changing the power supply, we've had boards drop like that and a swap of the power supply fixes it, sometimes leaving the unit unplugged for a few minutes and then plugging it back in helps but doing this usually drops the board again within a few hours.

Have also found that process of swapping the power supply allows the boards to cool enough that repowering will sometimes bring them up.  Multiple experiences where I thought it was the power supply, but it was actually just the few minutes of downtime.