Antminer S9 SE Troubleshooting

[philipma1957]

To be able to enable the fan control follow the guide by mikeywith posted from the link below once you enable it try to run it at 85% or higher and let see any changes on the hashboard temp and it should help to cool down a bit your unit.

If he gets temp errors then setting static fan won't help him, the miner would push the fan to it's limit or closet to it before sending any warning, and thus mean, his miner has already tried what you are suggesting but it still failed to operate at reasonable temp, what he should do instead if

1- provide better airflow.
or
2- underclock the miner.

it has been a while since I worked with an S9 SE, but if he follows my guide that you linked to unlock the hidden settings, it's possible to find clock and voltage settings which he can play with.

I would simply detach the center board from the 3 power jacks. Mine ran well for years like that.

[mikeywith]

I would simply detach the center board from the 3 power jacks. Mine ran well for years like that.

That would certainly help cool the miner by a lot, unlike removing the hashboard altogether because that would mess with the airflow since air would pass easily through the no-resistance without removing enough heat.

The middle board is the best one to unplug especially given that it is the one that is subject to fail before the other two because of the stupid design on these miners where the front chips of the middle board are right behind the fan motor getting too little airflow.

[SolHarvester]

With forced air I had one inline 6" duct fan rated at 345 CFM in the enclosure inlet side. Now I have two inline fans running at maximum. I have settled on a design so the enclosure has two air tight chambers to separate the cool air intake from the hot air exhaust.

I did a battery of tests today but when the APW7 voltage was dropping I could tell it was bad. If I'm not mistaken the PSU should maintain a voltage of 11.9 - 12.9 even with all three hash boards powered.

I listed the tests in order of which I performed them.

Note:
The voltage is measured at a power jack that is not being used.

APW7 PSU:

Ambient temp 15C

No hash boards: 12.04v
Test 1 - Chain# 1: Elapsed 45 min., 60 ASICs, 5630GH/S (avg), didn't measure voltage
Test 2 - Chain# 1,2: Elapsed 10 min., 120 ASICs, 0GH/S (avg), 8.07v
Test 3 - Chain# 1,3: Elapsed 10 min., 120 ASICs, 0 GH/S (avg), 8.20v
Test 4 - Chain# 2,3: Elapsed 20 min., 120 ASICs, 0 GH/S (avg), 7.60v
Test 5 - Chain# 2: Elapsed 32 min., 60 ASICs, 3,430GH/S (avg), 9.69v
Test 6 - Chain# 3: Elapsed 36 min., 60 ASICs, 3,232GH/S (avg), 9.54v
Test 7 - Chain# 1: Elapsed 1h 5 min., 60 ASICs, 0GH/S (avg), 8.65v

Chain Force Cannon PSU (2000W) - Model LL2000MINI:

Ambient temp 15.6C

No hash boards: 12.25v
Test 1 - Chain# 1: Elapsed 23 min., 60 ASICs, GH/S 5,707 (avg), didn't measure voltage
Test 2 - Chain# 1,2: Elapsed 23 min., 120 ASICs, GH/S 11,423 (avg), 12.23v
Test 3 - Chain# 1,2,3: Elapsed 4 min., 180 ASICs, GH/S 16,432 (avg), didn't measure voltage

Yup. It's the power supply. So the miner works fine I suppose, not perfect, but it works. it's just trying to keep it cool and have at least a PSU that works.
I think the APW7 I have is pretty much trash, so i'll have to return it.