S17+ Socket connection issue / 0 ASIC found Need help!

Hey everyone, I recently got my hands on two used S17+ units to learn the ropes. Complete beginner here. Initially, after setting up the NiceHash stratum, they both started hashing. But one of them only detected one out of three boards, while the other recognized two. After rebooting and flashing the firmware, one is working well on NiceHash, but the other seems like it might just be a fancy paperweight. I can't seem to shake off the 'Socket Connect Failed' error, and the machine won't recognize the boards. Honestly, I haven’t turned off the working unit yet because I’m worried it might end up bricked too. I plan to try powering it down once I can get the other one up and running, so at least one is hashing. I’ve tried pretty much everything I found on this forum and other places. Here’s what I’ve done so far. This is my last attempt to fix it myself before I consider sending it out for repairs. I'm really at my wits' end here. Here’s what I’ve tried: 1. Rebooting via the browser and also by turning the power off and on. 2. Resetting from the browser and the front panel. 3. Uploading the updated firmware from the browser and from an SD card, but the card might be too large (16G) since it didn’t work. 4. Opened up the control board cover to check the connections. 5. Made sure the machine is properly grounded. 6. I might have tried something else, but I can’t recall right now. 7. I let it run for about 10-20 minutes after each attempt before making another change. I’m attaching the Kernel Log from my last try. Any help from this community would be greatly appreciated! Thanks a ton in advance!

nicehash is not a great choice. that aside it looks like you have bad boards. try flashing braiins firmware to and sd card and boot from the sd card. if it boots and lets you do manual settings try freq 400 for each board and volts 16.3 good luck

I was afraid of bad boards. I'm trying your suggestion right now, thanks.

my miner seems to have some real issues. I formatted a 16g SD card to FAT32 and copied the BOS+ to work from SD and to Auto Install. On both the machine turns on but keeps both lights on and never seem to really boot. Seems like it gets stuck on booting from the card. Any suggestions? Maybe this is where I should ask for people that could check my miner and fix it? Thoughts? BTW completely agree with the NH statement, it was just the first one I tried / tested but am already looking into other options Thanks

Chain[1]: find 0 asic, times 2 Chain[1]: find 0 asic, times 2 Chain[2]: find 0 asic, times 2 All three boards are failing to boot and likely have bad asic chips. Way to common with S17 models. I have one board that if I reboot maybe 6 - 8 times in a row, it comes to life. And hashes without hiccup for weeks until I have to reboot. **You could try that just to see if you can get a board to go, but may not work and the board will still need repair eventually. I have had 6 of 9 boards fail in three machines. I was able to get 2 repaired, 1 unrepairable, and others pending repair hopefully. If you feel like it, you could carefully remove the boards, and may find loose or shifted heat sinks as well. That's a sign it got too hot. Loose heatsinks (completely detached), or even shifted, can sometimes be reset with a heat gun if you know what you are doing, and board MAY come back. This has worked for me a few times. But there is likely still an issue with that chip overheating in general. Shifted heatsinks, again because that chip got too hot, and heatsink did re-fuse, may not have re-fused completely or created solder balls under on chip shorting out asic. Finding a repair facility is hard. At least has been for me as they are backlogged. I used HMtech most recently but they have stopped accepting new customers with less than 25 units in need of repair due to backlog.

I'll give it a try on rebooting several times in a row. That's one thing I haven't tried yet I started looking for a repair facility and, like you said, no one is taking new orders, specially for only one machine. I'll also try to find time this weekend to open it up and check the boards. I'm also suspicious of the control board. Following another thread I disconnected the boards to check if it would connect to no success. But Besides the socket connect failed everything else seems to work, I can ping other urls and all seem to work. Then back to the hashboard option - maybe I'll try to learn myself a way to DIY and fix one to start.

In most cases, when the kernel log says hash board x found 0 Asics, it simply means the board is dead, it either has a dead chip somewhere or another burned component, in other rare cases a bad PSU will cause this error, the PSU needs to supply enough power to those hash boards in order for them to function and report the correct asic count, sometimes when the PSU is faulty but not completely dead (it only supplies 12v to the control board) you might get this 0 asic error. So long story short, try swapping the PSU and see if you get lucky.

Is there a way to check if it's the psu instead of a swap? Would I be able to measure the voltage with a multimeter? or would the kernel log report the actual voltage in the case of a faulty unit? Thanks

You could measure the voltage across all three board at the edge of the busbar, but honestly, I don't have enough knowledge about that PSU to confirm if that will be a sufficient and accurate test, I still think the best approach is swapping the PSU, it's not that hard, it takes like 5 mins or so.

hey, had some progress this weekend. Managed to install Braiins on the machine after installing the Antminer MP firmware. Got the Kernel log from the MP firmware, see below, and to my untrained eyes almost looks like it could recognize all the asics. However I could not configure it to mine so installed Braiins. With Brains the machine connects to NiceHash but shows zero activity on hashing. In my view proves control board is working but hashing are bad except for the MP firmware that seems to say something else. I also opened the top and measured the voltage whish was aligned with expectations, 21v solid. Unless anyone has any other suggestions I'm starting to reach out to places to send the machine for repair. I'm not sure I'd try to open it further and check the hashboards etc. Also not sure I would be able to fix it anyway, any thoughts? Thanks for the guidance so far.

In Braiins OS the logs may have a line that says x voltage requested, y given. Also for nicehash you have to use sha256asicboost urls, not sha256. I would also try Autotuning enabled and the lowest power limit. Remember, with Autotuning enabled, none of the manual settings (speed/voltage) apply.

hey, thanks for the note. I was pointing it to sha256 boost, my mistake not calling it out. I tried the different settings in Braiins and still got the same old 0 hash. Only difference this time is that NiceHash aknowledged the miner was connected but no hashing I'm packing the unit and sending for repair. Thank you all for your suggestions!! Was worth trying!!

you have to go to the repair house. but if you have the ability to handle the electronics, you can fix it by yourself. in my case, I can fix the older miner, for example L3+, S9, also having S17 series, just send to repair center due to it has still not to be the EOS.