ANTMINER S3+ Discussion and Support Thread

[luckypyrate]

Have you tested the psu and cables?  If you have a modular psu you should replace the cables and try and see if that clears it up.  Any errors in logs?

[Nite]

Have you tested the psu and cables?  If you have a modular psu you should replace the cables and try and see if that clears it up.  Any errors in logs?

its partial modular, the cables are hard wired so can't replace them.

Might just be a bug or something, just like any operating system if you restart it, it helps it run more efficently, I guess I don't mine restarting it once a week.

[Xian01]

If it is space heaters that you REALLY want, AND you do not mind maintaining / reducing your hash-rate, you could be better served by acquiring a set of S1's, underclock & undervolt them (to output ~125 GH/s @ - and don't ask me for a link, search the forum!) and replace the fan with a couple of quieter ones in a push pull layout at full pelt, aka blue wire hack; For a tad less power for both compared to your bitfury rig, you'd be laughing!

Thanks for the advice.

I'm looking to make at least a 450GH/s quiet space heater @ < 500W 

[jackbox]

Just had a strange issue on my S3+. It was only hashing about half what it should be. It is clocked at 225 (default) and in room with aircon on. All the Asics showed working. Temps were 38C on one blade and only 23C on the other blade. Resetting pool configurations did not change this. It appeared that one blade was working at full speed and the other blade showed good but not doing any work. After being puzzled I finally just powered down the S3+, waited and powered it back up. Now both blades are hashing again and temps coming up. Is this just an issue of the controller getting confused? For now it is hashing at full speed again.

[Nite]

Just had a strange issue on my S3+. It was only hashing about half what it should be. It is clocked at 225 (default) and in room with aircon on. All the Asics showed working. Temps were 38C on one blade and only 23C on the other blade. Resetting pool configurations did not change this. It appeared that one blade was working at full speed and the other blade showed good but not doing any work. After being puzzled I finally just powered down the S3+, waited and powered it back up. Now both blades are hashing again and temps coming up. Is this just an issue of the controller getting confused? For now it is hashing at full speed again.

!!!!!!!!!!

I had that same problem above check a few posts above yours, interesting coincidence that it happened to you too, is thier some kinda virus or bad work units foating around that disables the miners, or a faulty design in the s3 or something 

it has to be a network wide issue if we both got it at the same time.

I'm using Ghash pool btw, are you on the same one? I'm thinking about changing pools if it happens again....

[jackbox]

Just had a strange issue on my S3+. It was only hashing about half what it should be. It is clocked at 225 (default) and in room with aircon on. All the Asics showed working. Temps were 38C on one blade and only 23C on the other blade. Resetting pool configurations did not change this. It appeared that one blade was working at full speed and the other blade showed good but not doing any work. After being puzzled I finally just powered down the S3+, waited and powered it back up. Now both blades are hashing again and temps coming up. Is this just an issue of the controller getting confused? For now it is hashing at full speed again.

!!!!!!!!!!

I had that same problem above check a few posts above yours, interesting coincidence that it happened to you too, is thier some kinda virus or bad work units foating around that disables the miners, or a faulty design in the s3 or something 

it has to be a network wide issue if we both got it at the same time.

I'm using Ghash pool btw, are you on the same one? I'm thinking about changing pools if it happens again....

I am hashing on Slush's Pool. Weird happened to both of us around same time?

[Nite]

ahh I see what happened, I was looking at the stats for ghash and it was an ORPHAN block, not our faults, so anytime that the one of the hashing boards goes off it means an ORPHAN block happened, keep that in mind.

The Orphan block must have confused the miner and caused it to go into a shutdown state. Wierd.

[jackbox]

ahh I see what happened, I was looking at the stats for ghash and it was an ORPHAN block, not our faults, so anytime that the one of the hashing boards goes off it means an ORPHAN block happened, keep that in mind.

The Orphan block must have confused the miner and caused it to go into a shutdown state. Wierd.

But why would it do that if I was on Slush Pool? And why does the chain stay offline and not resume work without a power down and restart?

[Nite]

ahh I see what happened, I was looking at the stats for ghash and it was an ORPHAN block, not our faults, so anytime that the one of the hashing boards goes off it means an ORPHAN block happened, keep that in mind.

The Orphan block must have confused the miner and caused it to go into a shutdown state. Wierd.

But why would it do that if I was on Slush Pool? And why does the chain stay offline and not resume work without a power down and restart?

no clue :/ but why it cuts the mining of the S3 in half is unknown, its something Bitmain will have to look into.

https://blockchain.info/orphaned-blocks anyway here is a page that lists all orphan blocks as they happen. the latest one said it was relayed by Ghash

[BITMAIN]

BITMAIN Promotion plan for S3+, S4 and C1

$50 coupon for S3+, $400 coupon for S4 and $50 coupon for C1 are issued to customers who ever bought miners from BITMAIN or our official distributors

[jackbox]

BITMAIN Promotion plan for S3+, S4 and C1

$50 coupon for S3+, $400 coupon for S4 and $50 coupon for C1 are issued to customers who ever bought miners from BITMAIN or our official distributors

In my account are the S3 and S4 coupon, but not any C1 coupons.

[Rabinovitch]

So as in my...

[MissouriMiner]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

[mdude77]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

That's interesting.  That means I should be OC'ing my S3+ units as I'd still come out better than a C1 will! 

Not what I expected.

M

[pekatete]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

I wonder how far an OC'd C1 can be pushed though .... seeing it is stock rated for ~1Gh/s @ 250 (and the S3+'s were never rated that high, albeit for bitmain's assertion that most will hit 500 Gh/s), that liquid cooling may have some mileage still, what with ramping up that solo fan via the blue wire hack in tandem ........ just a thought.

[mdude77]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

I wonder how far an OC'd C1 can be pushed though .... seeing it is stock rated for ~1Gh/s @ 250 (and the S3+'s were never rated that high, albeit for bitmain's assertion that most will hit 500 Gh/s), that liquid cooling may have some mileage still, what with ramping up that solo fan via the blue wire hack in tandem ........ just a thought.

I was thinking that too.  However, at what cost?  Just because it can run at say, 1.2 TH/s, would you want to?  Would the additional power usage really not make it worth while?  If an S3+ overclocked to the highest uses less power than a C1 at base speed, then C1 power usage might be off the chart for OCing.

M

[pekatete]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

I wonder how far an OC'd C1 can be pushed though .... seeing it is stock rated for ~1Gh/s @ 250 (and the S3+'s were never rated that high, albeit for bitmain's assertion that most will hit 500 Gh/s), that liquid cooling may have some mileage still, what with ramping up that solo fan via the blue wire hack in tandem ........ just a thought.

I was thinking that too.  However, at what cost?  Just because it can run at say, 1.2 TH/s, would you want to?  Would the additional power usage really not make it worth while?  If an S3+ overclocked to the highest uses less power than a C1 at base speed, then C1 power usage might be off the chart for OCing.

M

I don't think I agree with you there. Essentially a C1 is four S3 boards and the only way it can theoretically (I know we have empirical evidence above!) consume more power than 2 S3+'s at the same freq has to be something that can be rectified by a user, possibly heat handling (aka loose paste or the liquid cooling being inferior to the S3+'s, which I doubt) or voltage setting (the new firmware has this, do both the S3+ and C1 compared above have the same setting, if at all?).

[mdude77]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

I wonder how far an OC'd C1 can be pushed though .... seeing it is stock rated for ~1Gh/s @ 250 (and the S3+'s were never rated that high, albeit for bitmain's assertion that most will hit 500 Gh/s), that liquid cooling may have some mileage still, what with ramping up that solo fan via the blue wire hack in tandem ........ just a thought.

I was thinking that too.  However, at what cost?  Just because it can run at say, 1.2 TH/s, would you want to?  Would the additional power usage really not make it worth while?  If an S3+ overclocked to the highest uses less power than a C1 at base speed, then C1 power usage might be off the chart for OCing.

M

I don't think I agree with you there. Essentially a C1 is four S3 boards and the only way it can theoretically (I know we have empirical evidence above!) consume more power than 2 S3+'s at the same freq has to be something that can be rectified by a user, possibly heat handling (aka loose paste or the liquid cooling being inferior to the S3+'s, which I doubt) or voltage setting (the new firmware has this, do both the S3+ and C1 compared above have the same setting, if at all?).

Am I reading the above chart wrong? 

M

[MissouriMiner]

I have 2x S3+ that run well at 250Mhz, and I have some C1's.  So I did a power comparison and am sharing my results.

Configuration   MHz Setting   Avg Gh/s   Watts      Efficiency W/Gh      PSU Model
2x S3+            250            1000         903         0.90                     1000W Coolmax 80 Plus Bronze
C1                  250            1000         984         0.98                     1000W Coolmax 80 Plus Bronze
2x S3+            250            1000         821         0.82                     1300W Rosewill Lightening 80 Plus Gold
C1                  250            1000         881         0.88                     1300W Rosewill Lightening 80 Plus Gold

Notes:  Actual GH/s is close to 1000.  2x S3+ averaged 504Gh/s each.  The power usage for the C1 includes the "ANT C1 Mining Cooling Kit" from syscooling.com.

I wonder how far an OC'd C1 can be pushed though .... seeing it is stock rated for ~1Gh/s @ 250 (and the S3+'s were never rated that high, albeit for bitmain's assertion that most will hit 500 Gh/s), that liquid cooling may have some mileage still, what with ramping up that solo fan via the blue wire hack in tandem ........ just a thought.

I was thinking that too.  However, at what cost?  Just because it can run at say, 1.2 TH/s, would you want to?  Would the additional power usage really not make it worth while?  If an S3+ overclocked to the highest uses less power than a C1 at base speed, then C1 power usage might be off the chart for OCing.

M

I don't think I agree with you there. Essentially a C1 is four S3 boards and the only way it can theoretically (I know we have empirical evidence above!) consume more power than 2 S3+'s at the same freq has to be something that can be rectified by a user, possibly heat handling (aka loose paste or the liquid cooling being inferior to the S3+'s, which I doubt) or voltage setting (the new firmware has this, do both the S3+ and C1 compared above have the same setting, if at all?).

Am I reading the above chart wrong?  

M

Both the C1 and S3's are in the same room, right next to each other, and the ambient temp is the same.  The C1 clock speeds are the default of 250.  The S2's are overclocked to 250.  So, both comparison have 4 hashing boards running at 250Mhz, and all miner voltages are set to 0725.  My wall outlet is 120V.  I'm using the exact same power supply in the tests.  I just swapped the power supplies.  I had the 1300W connected to one, and the 1000W connected to the other, then swapped.  Don't forget the PSU for the C1 config is also powering the water cooling unit, which is the pump plus 3 fans.  I have 2x C1's and both are using the same watts.  I can't explain it.  I'm just sharing.  I wish my C1's were using less electricity, and would like to know if I have a couple lemons or what...

[mdude77]

Both the C1 and S2's are in the same room, right next to each other, so the ambient temp is the same.  The C1 clock speeds are the default of 250.  The S2's are overclocked to 250.  So, both comparison have 4 hashing boards running at 250Mhz.  I'm using the exact same power supply in the tests.  I just swapped the power supplies.  I had the 1300W connected to one, and the 1000W connected to the other, then swapped.  Don't forget the PSU for the C1 config is also powering the water cooling unit, which is the pump plus 3 fans. 

I thought of the pump + 3 fans idea ... but two S3+'s have four fans altogether.  So I'd think it's roughly the same, unless there's something about those C1 fans and pump that is using more power than 4 fans on the S3+'s.

M