I can't with my Antminer S17+... HELP

[g3rula]

From some time now… I started to repair that Chinese shit, called Antminer.
I was forced to do that, because I spent all my 5 years savings on 2 Antminer S17+ from China, but DHL brings me both devices destroyed.
Long story short…
I repaired my first hashboard, after months if work, and even when I replaced the PIC firmware with Pickit 3, one board have a different voltage than another two, and Asic power off.
I have to mention, if I use only 2 cards, in any combination, it’s working. But not all 3.
Any suggestions, please?  

Code:

2021-08-29 08:46:33 voltage[0] = 1940
2021-08-29 08:46:33 voltage[1] = 1960
2021-08-29 08:46:33 voltage[2] = 1960
2021-08-29 08:46:33 working_voltage_are different
2021-08-29 08:46:33 shut down

[BitMaxz]

It would be better if you could post the whole kernel logs here so that we can see more insight into this issue.
If you couldn't post the whole kernel logs here due to limitations you can paste them here https://pastebin.com/ and paste the link here.


What is your actual outlet voltage?
If it's 220v try to replace PSU because this is a well-known issue on this unit.

And try to directly plug your miner into the outlet and don't use any power surge.

[Artemis3]

when I replaced the PIC firmware with Pickit 3, one board have a different voltage than another two, and Asic power off.

This part is worrisome. We have seen people that mysteriously have the eeprom configured for a different model, and each model uses different voltages and speeds. Braiins OS+ catches this (and you can manually force the correct model), not sure about factory fw or mods.

Of course people wouldn't be messing with that if they hadn't their eeproms corrupted in the first place... Fortunately Braiins OS+ never ever writes those, but it reads for model identification and if there is a discrepancy the log shows an error and won't hash.

You absolutely cannot mix hashboards from a different model, so maybe the other fw is panicking and (correctly) refusing to work.

[mikeywith]

Code:

2021-08-29 08:46:33 voltage[0] = 1940
2021-08-29 08:46:33 voltage[1] = 1960
2021-08-29 08:46:33 voltage[2] = 1960
2021-08-29 08:46:33 working_voltage_are different
2021-08-29 08:46:33 shut down

Post the complete kernel log, each hash board will have a product id of some kind in the begging of the kernel log, but to jump to the end of the story, the stock firmware is very strict with the board voltage which is stored in the EEPROMs, using custom firmware will likely fix the issue, Atermis3 mentioned BO+, I also know that vanish ignores the voltage readings and will work as long as all hash boards are the same model.

If however using custom firmware does not fix it, it could be a PSU-related issue whereby the PSU simply can't handle 3 hash boards, you will need to test the miner with a different PSU shall the prime persist with using custom firmware.

[thierry4wd]

Code:

2021-08-29 08:46:33 voltage[0] = 1940
2021-08-29 08:46:33 voltage[1] = 1960
2021-08-29 08:46:33 voltage[2] = 1960
2021-08-29 08:46:33 working_voltage_are different
2021-08-29 08:46:33 shut down

Post the complete kernel log, each hash board will have a product id of some kind in the begging of the kernel log, but to jump to the end of the story, the stock firmware is very strict with the board voltage which is stored in the EEPROMs, using custom firmware will likely fix the issue, Atermis3 mentioned BO+, I also know that vanish ignores the voltage readings and will work as long as all hash boards are the same model.

If however using custom firmware does not fix it, it could be a PSU-related issue whereby the PSU simply can't handle 3 hash boards, you will need to test the miner with a different PSU shall the prime persist with using custom firmware.

exactly ! very well explained!

to solve the problem, you reprogram the hash cards or you install a custom firmware (I offer one)

but in any case, this does not come from the controlboard or the psu