s9 frying connections with PSU

[Magaiveris]

Any ideas what might be wrong with one of my s9? I have 4 of them, one of them got fried the connections.

https://i.ibb.co/V2JK7V3/80394330-E629-470-E-AE49-A70328118225.jpg

https://i.ibb.co/44fqpdC/072-A0-AEC-FD81-409-F-BCB0-20-A9373139-A8.jpg

[1715532077]

1715532077

[1715532077]

1715532077

[1715532077]

1715532077

[1715532077]

1715532077

[HagssFIN]

Did you use all ten pieces of the cables?

How thick is your wire gauge?

[philipma1957]

Any ideas what might be wrong with one of my s9? I have 4 of them, one of them got fried the connections.

did this fix photo?


okay what is your voltage rating?

what did you set the hash rate to?

looks like you have a apw3++ or apw3+ psu is that correct?

Assuming  you have an apw3+ or apw3++

you can melt out for a few reasons.

High clock like 15th or 16th
voltage drop from 240 to 200 this means more amps and more heat.
connections were not tight due to plugging and unplugging


I have found that a fairly high clock combined with a voltage drop  or with a loose connection does exactly what you have.

Try clocking all the rest of your units 1 or 2 th lower than they are right now.

[NotFuzzyWarm]

voltage drop from 240 to 200 this means more amps and more heat.

That is very wrong. Line voltage only affects current on the AC line side of a PSU and has zero effect on the output (as long as the line is high enough for the PSU to work that is...). All the rest is 100% correct.

Given the age of the gear I lean to the PCIe connectors being loose/worn out. As I've said many times before, those connectors have a very short mating cycle lifetime. The highest quality ones are only good for around 25 plugging/unplugging cycles before the pins are damaged. El-cheapo ones, well....

[Magaiveris]

Did you use all ten pieces of the cables?

How thick is your wire gauge?

Yes all ten, you mean main power cable to PSU? Not sure. How thick i should be? Using ordinary pc cables.

[Magaiveris]

Any ideas what might be wrong with one of my s9? I have 4 of them, one of them got fried the connections.

https://i.ibb.co/V2JK7V3/80394330-E629-470-E-AE49-A70328118225.jpg

https://i.ibb.co/44fqpdC/072-A0-AEC-FD81-409-F-BCB0-20-A9373139-A8.jpg

did this fix photo?


okay what is your voltage rating?

what did you set the hash rate to?

looks like you have a apw3++ or apw3+ psu is that correct?

Assuming  you have an apw3+ or apw3++

you can melt out for a few reasons.

High clock like 15th or 16th
voltage drop from 240 to 200 this means more amps and more heat.
connections were not tight due to plugging and unplugging


I have found that a fairly high clock combined with a voltage drop  or with a loose connection does exactly what you have.

Try clocking all the rest of your units 1 or 2 th lower than they are right now.

This is s9 14th with APW3++ on 220v using it with latest official firmware, no overclocking. It is operating always on low power mode with average  13,79TH. Together with other machines (total of 4xs9 and 1xS19) connected to one source of 10kw.

It was bought second hand so not sure about sockets usage, anyways ordered new ones from china together with wires for replacement.

The rest of machine have several slighly  brown sockets, but this one went all the way.
It shouldnt be the psu also since i already have replaced one, because i thought its the problem.


Thanks!

[Artemis3]

The molex megafit 3x2 Chinese edition is garbage, they do that after a certain number of cycles of plugging and unplugging, with all miners that use it, IF you never pay attention to the metal parts inside becoming lose.

It is not firmware related, nor psu. Its a connector issue. If its too easy to insert/remove, you should be wary. Get some tweezers or something to tighten the metal female (inside the plastic socket).

I'm a little surprised Bitcointalk didn't catch up with this (should be) common knowledge.