ANTMINER S3+ Discussion and Support Thread

[oskuro]

Is this normal???



same on both S3. Both with default frequency: 218.75

why so many discarded??

strangely enough, yes, but SOMETIMES rebooting from the webgui (not by power cycle) helps somewhat. Also it seems that discarded shares ## are higher on ghash and btcguild than on eligius (personal observation on the same miner by rotating pools).

Maybe if i change frequency... i rebooted using webgui but still the same

boring that i get all the time a lot of discards.... i will try later to change frequency.

I will change to 212,50 later to see what happens.....

[DevonMiner]

Lots of people got their coupons, but has ANYONE received their refund? I got the auto email response for querying the Bitmain website a day ago but no refund yet.

Nope ... not yet ... but I guess they will leave it a couple of days to make sure you haven't replied (and therefore changed your mind) before they refund.

[iglasses]

Lots of people got their coupons, but has ANYONE received their refund? I got the auto email response for querying the Bitmain website a day ago but no refund yet.

Nope ... not yet ... but I guess they will leave it a couple of days to make sure you haven't replied (and therefore changed your mind) before they refund.

Same here.  Got a confirmation email yesterday with the amount and payment address so it looks like they are getting ready to process them.

[hanti]

Is this normal???



same on both S3. Both with default frequency: 218.75

why so many discarded??

strangely enough, yes, but SOMETIMES rebooting from the webgui (not by power cycle) helps somewhat. Also it seems that discarded shares ## are higher on ghash and btcguild than on eligius (personal observation on the same miner by rotating pools).

Maybe if i change frequency... i rebooted using webgui but still the same

boring that i get all the time a lot of discards.... i will try later to change frequency.

I will change to 212,50 later to see what happens.....

whats the speed on pool side ?
if its 430 or more then everything is ok

[mrpark]

anyone know what the top room temperature would be to operate these safely. I run spondoolies and sometimes the room temp gets 100F and there is no problems.  Not sure if these A3's will hold up to the same abuse, maybe I should not subject them to that abuse either

[Trammel]

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!!

[BennyTheBall]

I see a lot of you guys are reflashing your S3s. I downloaded the new image, but I can't figure out how to tell which version is currently running and if this is actually a new version. Any ideas?

[oskuro]

Is this normal???



same on both S3. Both with default frequency: 218.75

why so many discarded??

strangely enough, yes, but SOMETIMES rebooting from the webgui (not by power cycle) helps somewhat. Also it seems that discarded shares ## are higher on ghash and btcguild than on eligius (personal observation on the same miner by rotating pools).

Maybe if i change frequency... i rebooted using webgui but still the same

boring that i get all the time a lot of discards.... i will try later to change frequency.

I will change to 212,50 later to see what happens.....

whats the speed on pool side ?
if its 430 or more then everything is ok

It says 873.29 GH/s for 2 S3 but anyways I don't think all those discards are normal

[-droid-]

For those getting the - status on the chip try changing the PCI plug location, instead of using front two try the back two or try one in the front and one in the back

[Trammel]

For those getting the - status on the chip try changing the PCI plug location, instead of using front two try the back two or try one in the front and one in the back

Thanks, I switched the PCI-e plugs to the back two sockets and the ASIC status is all zeros, fantastic 

[teeGUMES]

Is this normal???



same on both S3. Both with default frequency: 218.75

why so many discarded??

someone suggested that its possibly backwards, the accepted are actually the discarded, and vice versa. probably fixed in a new firmware upgrade soon

[Askit2]

Before I get more upset then I already am about this I had a question.

Has anyone ever received a full refund from 112bit?

I have been having issues with an order and since its sitting at the depot with very slow replies I am fairly sure it will be shipped back. I do not want to pick it up until I am satisfied the issues are resolved. Any help is appreciated (even if I seem less then grateful right now).

[ravvles]

Here are my 2 hashing away!

both are sitting around mid 40's in 30c room! not as bad as I first thought!

One is better than the other!

I've OC'd this one, it ran at 420avg at 218 but 461avg at 237!

http://i61.tinypic.com/2802khd.jpg

The other is still pretty good! runs at 420 at 218 but doesn't OC well and only hits 430 when OC'd to 237!

http://i59.tinypic.com/2ivky0x.jpg

Both units are quite loud due to the fans trying to keep them warm in the ambient temps but it's not unbearable! units themselves don't feel hot and the cables aren't even slightly warm to touch!

[not.you]

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.

[Duce]

anyone know what the top room temperature would be to operate these safely. I run spondoolies and sometimes the room temp gets 100F and there is no problems.  Not sure if these A3's will hold up to the same abuse, maybe I should not subject them to that abuse either

Well remember 100F is ~37C so if that is the ambiant temperature depending on how well you can exhaust the heat the S3 units should be fine. I do not have the luxury of a cool environment so the room my units are in hit about 95F, at that time they are reporting 43C well within operating parameters. I try to stay below 45C but I have seen others here running at 50C. The S3 units will turn off at 80C if you have that function enabled. I would look at the Spondoolies specification because the S3 units will heat that place up beyond 100F that is for sure.

I should add that I run one at 218 and the other at 237, higher clock rates will increase S3 temperature as well.

[Trammel]

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?

[Sitarow]

Here is a comparison on projected returns with network difficulty increasing 10% vs 20% after power costs assuming BTC US value remains at $600 per BTC.

Network Difficulty Increasing 10% @ $600 usd per BTC



Network Difficulty Increasing 20% @ $600 usd per BTC



Shared link https://docs.google.com/spreadsheet/ccc?key=0AmeuPljmUNHCdEpqX2RmMDFwemJyLURVUWFtZ3J3aGc&usp=sharing

For those who wish to checkout your results.

Just copy and save the document to your google drive.

[mrpark]

anyone know what the top room temperature would be to operate these safely. I run spondoolies and sometimes the room temp gets 100F and there is no problems.  Not sure if these A3's will hold up to the same abuse, maybe I should not subject them to that abuse either

Well remember 100F is ~37C so if that is the ambiant temperature depending on how well you can exhaust the heat the S3 units should be fine. I do not have the luxury of a cool environment so the room my units are in hit about 95F, at that time they are reporting 43C well within operating parameters. I try to stay below 45C but I have seen others here running at 50C. The S3 units will turn off at 80C if you have that function enabled. I would look at the Spondoolies specification because the S3 units will heat that place up beyond 100F that is for sure.

Oh good point, I should just look at the suggested operation temp range. Sometimes I am not thinking straight.  I think I am going to need to put in better exhaust venting in the area in question.

[Duce]

3. Beeper rings randomly?
We set a beeper ring alert in the first version of firmware, to alert users when there is no immediate hashing work due to bad pool configuration.
However this alert will be removed in the later version of firmware.
Users can download new version firmware on July 25.

Answered here:
https://bitcointalk.org/index.php?topic=671189.msg7598229#msg7598229

Remember new FIRMWARE tomorrow to ease the beeping.

[Biodom]

Here are my 2 hashing away!

both are sitting around mid 40's in 30c room! not as bad as I first thought!

One is better than the other!

I've OC'd this one, it ran at 420avg at 218 but 461avg at 237!

http://i61.tinypic.com/2802khd.jpg

The other is still pretty good! runs at 420 at 218 but doesn't OC well and only hits 430 when OC'd to 237!

http://i59.tinypic.com/2ivky0x.jpg

Both units are quite loud due to the fans trying to keep them warm in the ambient temps but it's not unbearable! units themselves don't feel hot and the cables aren't even slightly warm to touch!

good, but they stabilize slow. the results might be different in 24hr. one of mine accelerated well, but developed a stuck asic with a "-" in 24-36hr
I start to gravitate more to a position that overworking them at higher speed might result in earlier demise, so to speak. i don't care that much about 5% extra
Time will tell.