Quote from: not.you on July 24, 2014, 09:30:32 PM

Quote from: Trammel on July 24, 2014, 09:10:04 PM

Quote from: not.you on July 24, 2014, 09:05:04 PM

Quote from: Trammel on July 24, 2014, 08:30:07 PM

Quote from: philipma1957 on July 24, 2014, 07:55:29 PM

Quote from: Trammel on July 24, 2014, 07:48:20 PM

Received my S3's today and set them up this evening. The first one has no lines in the 'ASIC status' but the second one has a single line in it's 2nd 'ASIC status':

00000000 00000000
00000000 000000-0

Both are running off an EVGA 1300 Gold PSU using two PCI-e cables each.

I took the lid off the second S3 and took the heat-sinks off and cleaned up all chips re-applied a small dot of thermal CPU paste, reassembled and the 'ASIC status' was initially all clear and then after a couple of reboots the line re-appeared in the 'ASIC status'. Any suggestions?

The first S3 has ~421 and ~420 GHs running at stock frequency of 218.75.

The second S3 has ~475 and ~419 GHs running at same stock frequency.

They've only been running for half an hour.

you can reboot the second one with the gui reboot option.

if the _ sticks change the freq to 212.5

both of my miners won't mine well at 218.75 they have 1 or 2 constant _ but both are good at 212.5

I changed the second S3 to operate at freq of 212.5 and still have a '-' in the ASIC status!!

I have one of those too. Or had. Yesterday when I got the miners I had it come back several times. A reboot would sometimes make it go away or change to an x but it kept coming back. First thing I did with both miners was try them at various freq but neither of them seemed to like OC very much. Hashrate was marginally better at each step but HW errors were a lot higher on both at any OC. I did try downclock also but it did not make the - go away and made the hash rate even more pathetic. Neither of my miners pull 441 Ghs at stock freq.

Yesterday I took off the heat sinks for the one with the bad chip and the chips were not drenched in thermal paste as some people have seen, but actually seemed like they could have used a little more. I was pretty meticulous about putting it on the entire surface of each chip without a lot of extra (and cleaning it up on the sides where it did run off and pool a little). But even after that it was getting the -. But yesterday after having done that thermal paste cleanup I OC'd to 225 and left it for the night. What I noticed this morning was that the HW errors were much improved compared to when I first unboxed them - I thought it might be due to the cleanup of the thermal paste but it could also have been that bad chip being presumably left out of the loop (since it had a dash all night) so I thought maybe it had been the source of the HW errors before. So thinking to try and nudge it closer to the 441 advertised hash rate I bumped it up to 237.5 freq this morning. Now it is actually just getting 441 but the problematic chip has also been showing an o ever since also. I doubt it is really fixed but I am happy about the hashrate for now. But I won't be shocked if the chip shows an x after the next reboot either. As long as I get 441 out of it though I'll live with it.

I also took off the heatsinks on my other miner since it also wasn't pulling 441. That one actually looked like it had a good job of thermal paste on both sides. No huge voids and not a ton dripping off the sides. But it still had a little bit pooled on one side or another of a couple chips. Again I was real meticulous about getting it just right and cleaning up the excess and lo and behold that one is suddenly working great at 250 freq with minimal HW errors. It's pulling 501 Ghs ave at 3 hours now. That one was a little weird though too, the 237.5 freq actually lowered the hash rate to about 400 so I was actually pretty surprised to see what happened at 250.

I was pretty bummed with both of them yesterday but I am happier about it today with the one pulling such a high hash rate and the other at least making the advertised 441. I run them both with fans at full (blue wires disconnected) and covers off in a server room (ambient temps maintained below 70 Fahrenheit). The temps in the GUI are all mid 30's.

I switched the PCI-e cables to the two rear sockets and the - in the ASIC status went. Are you using 4 PCI-e cables per S3?

I am using 4 yes.

I increased the freq and then got an x in the ASIC status, so took freq back to stock of 218.75 and the - returned to the ASIC status. I've now added a further 2 PCI-e cables (so running with 4 on the second S3) and all ok again at stock freq. I may try upping the freq slightly tomorrow.