ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH

[thedreamer]

Both my 4.86TH Batch 1's when downclocked to 575Mh, got 0.0001 or 0.0000% error rates at just over 4.73TH average over a 12 hour period while testing.

I also found, that by loosening up the screws on the fans front and aft (all the sound became a much less 'droning' and a smooth 'hooosh' type of sound.
Originally the sound was like a humming type , pretty loud.

And yes, like I said I would, I got rid of my S7's as I'm done with Bitmain as a business after the crap they pulled with this miner.

[-droid-]

Both my 4.86TH Batch 1's when downclocked to 575Mh, got 0.0001 or 0.0000% error rates at just over 4.73TH average over a 12 hour period while testing.

I also found, that by loosening up the screws on the fans front and aft (all the sound became a much less 'droning' and a smooth 'hooosh' type of sound.
Originally the sound was like a humming type , pretty loud.

And yes, like I said I would, I got rid of my S7's as I'm done with Bitmain as a business after the crap they pulled with this miner.

Ahh that's why they were all loose lol

[Biodom]

Both my 4.86TH Batch 1's when downclocked to 575Mh, got 0.0001 or 0.0000% error rates at just over 4.73TH average over a 12 hour period while testing.

I also found, that by loosening up the screws on the fans front and aft (all the sound became a much less 'droning' and a smooth 'hooosh' type of sound.
Originally the sound was like a humming type , pretty loud.

And yes, like I said I would, I got rid of my S7's as I'm done with Bitmain as a business after the crap they pulled with this miner.

Ahh that's why they were all loose lol

mine are still tight, did not want to stop the hashing during first 24 hr.
I also did the trick with screws on B1, but forgot this time.

[thedreamer]

Both my 4.86TH Batch 1's when downclocked to 575Mh, got 0.0001 or 0.0000% error rates at just over 4.73TH average over a 12 hour period while testing.

I also found, that by loosening up the screws on the fans front and aft (all the sound became a much less 'droning' and a smooth 'hooosh' type of sound.
Originally the sound was like a humming type , pretty loud.

And yes, like I said I would, I got rid of my S7's as I'm done with Bitmain as a business after the crap they pulled with this miner.

Ahh that's why they were all loose lol

LOL. Forgot to tell ya. Didn't think you'd mind the less noise. :-)

IMO these are actually quieter at full speed but fans at 37 or below, than the S5's with stock fans running PWM.

[cl4ud1u]

One miner was giving me lots of HW errors and when i removed the fans to check the heatsinks, surprise...one missging







What can i do ? Write to Bitmain or  buy heatsinks? Where from ?

[notlist3d]

One miner was giving me lots of HW errors and when i removed the fans to check the heatsinks, surprise...one missging







What can i do ? Write to Bitmain or  buy heatsinks? Where from ?

If in warranty write bitmain.  If you say you opened it up without their permission they can void warranty just a FYI.  So you might ask them if you can open it in a email or support ticket...  if you say you did it on your own I  would hate to have warranty voided.

But yes get in contact with them.

[klondike_bar]

One miner was giving me lots of HW errors and when i removed the fans to check the heatsinks, surprise...one missging


What can i do ? Write to Bitmain or  buy heatsinks? Where from ?

not sure if bitmain would be fast or particularly helpful to respond. An easy DIY would be to buy a small adhesive heatsink (look on ebay/amazon for "tiny heatsink, bitcoin heatsink, rpi heatsink" or something along those lines, they stick on easily and while smaller than the original, should help somewhat.

alternatively, ignore it. accept the loss of ~80GH so long as the system appears to be getting ~98% of the nominal hashrate

[hawkfish007]

One miner was giving me lots of HW errors and when i removed the fans to check the heatsinks, surprise...one missging







What can i do ? Write to Bitmain or  buy heatsinks? Where from ?

Bitmain replaced 2 of my B1 hashboards, both had loose heatsinks with chip so I only had 1 working board. The exchange process was quite simple and it took less than 2 weeks.

[dmwardjr]

Not sure if this has already been mentioned in the thread:

BITMAIN did correct the specifications on Batch 6 to say the following:  "5. Chip quantity per unit: 135 x BM1385"

I still find it confusing it says:

2. Power Consumption: 1042 W + 10% (at the wall, with APW3, 93% efficiency, 25C ambient temp)

3. Power Efficiency: 0.25 J/GH + 10% (at the wall, with APW3, 93% efficiency, 25°C ambient temp)

The reason this is confusing to me is because it does not say, "+/-" but only "+"



The BM1385 chip is capable of 32.5 GHS per chip at just 0.216 Watts of power usage per GHS with 0.66V core voltage.

I'm just totally confused at the moment.  Can someone chime in to help me understand if it is 1042 watts at the wall @ 600MHz or 1,146.2 watts (+10%) ??

[RichBC]

Not sure if this has already been mentioned in the thread:

BITMAIN did correct the specifications on Batch 6 to say the following:  "5. Chip quantity per unit: 135 x BM1385"

I still find it confusing it says:

2. Power Consumption: 1042 W + 10% (at the wall, with APW3, 93% efficiency, 25C ambient temp)

3. Power Efficiency: 0.25 J/GH + 10% (at the wall, with APW3, 93% efficiency, 25°C ambient temp)

The reason this is confusing to me is because it does not say, "+/-" but only "+"



The BM1385 chip is capable of 32.5 GHS per chip at just 0.216 Watts of power usage per GHS with 0.66V core voltage.

I'm just totally confused at the moment.  Can someone chime in to help me understand if it is 1042 watts at the wall @ 600MHz or 1,146.2 watts (+10%) ??

I have put my thoughts in the S7 "Lite" Thread


https://bitcointalk.org/index.php?topic=1219211.msg12902245#msg12902245

https://bitcointalk.org/index.php?topic=1219211.msg12911092#msg12911092


Still very confusing and in summary far from certain that Bitmain knows what the spec is yet. 

Rich

[chalkboard17]

Not sure if this has already been mentioned in the thread:

BITMAIN did correct the specifications on Batch 6 to say the following:  "5. Chip quantity per unit: 135 x BM1385"

I still find it confusing it says:

2. Power Consumption: 1042 W + 10% (at the wall, with APW3, 93% efficiency, 25C ambient temp)

3. Power Efficiency: 0.25 J/GH + 10% (at the wall, with APW3, 93% efficiency, 25°C ambient temp)

The reason this is confusing to me is because it does not say, "+/-" but only "+"



The BM1385 chip is capable of 32.5 GHS per chip at just 0.216 Watts of power usage per GHS with 0.66V core voltage.

I'm just totally confused at the moment.  Can someone chime in to help me understand if it is 1042 watts at the wall @ 600MHz or 1,146.2 watts (+10%) ??

The same instant bitmain put up b6 up, and was ignored like I usually am.
https://bitcointalk.org/index.php?topic=1165628.msg12893262#msg12893262
and another post regarding efficiency 2 posts down

In a way or another this last batch isn't as efficient as all previous s7. Up to (not saying it IS) 15% less efficient I'd say. And not a single word has been mentioned by them regarding this and they couldn't even put actual efficiency. They just put this "+" stuff to misguide people.

[bbOOmm]

Not sure if this has already been mentioned in the thread:

BITMAIN did correct the specifications on Batch 6 to say the following:  "5. Chip quantity per unit: 135 x BM1385"

I still find it confusing it says:

2. Power Consumption: 1042 W + 10% (at the wall, with APW3, 93% efficiency, 25C ambient temp)

3. Power Efficiency: 0.25 J/GH + 10% (at the wall, with APW3, 93% efficiency, 25°C ambient temp)

The reason this is confusing to me is because it does not say, "+/-" but only "+"



The BM1385 chip is capable of 32.5 GHS per chip at just 0.216 Watts of power usage per GHS with 0.66V core voltage.

I'm just totally confused at the moment.  Can someone chime in to help me understand if it is 1042 watts at the wall @ 600MHz or 1,146.2 watts (+10%) ??

The same instant bitmain put up b6 up, and was ignored like I usually am.
https://bitcointalk.org/index.php?topic=1165628.msg12893262#msg12893262
and another post regarding efficiency 2 posts down

In a way or another this last batch isn't as efficient as all previous s7. Up to (not saying it IS) 15% less efficient I'd say. And not a single word has been mentioned by them regarding this and they couldn't even put actual efficiency. They just put this "+" stuff to misguide people.

What do you expect? Really, Bitmain is the only manufacturer of mining equipment for the "Home" miner. They have the monopoly and can do whatever they please. Their mentality is - If you don't like their policies and practices, well, go find another manufacturer.

Till there is real competition, expect this crap.

My opinion is .... there is a lot of instability within the chip. That is why there are so many different versions, hashrates and clock speeds. So instead of just tossing the QC failed boards, they have been individually testing each hashboard to grade the board for the different batches therefore mitigating their production losses.

This board with less chips, lower hashrate, but with greater power usage, My guess would be this was the original batch 1 that just absolutely failed expectations and resulted a re-design leading to the boards that are in the B1,B2,B3 etc .... They did say they had a very limited number of them.

It would make sense considering their past behavior of recycling the QC fails.

[dmwardjr]

This board with less chips, lower hashrate, but with greater power usage, My guess would be this was the original batch 1 that just absolutely failed expectations and resulted a re-design leading to the boards that are in the B1,B2,B3 etc .... They did say they had a very limited number of them.

I suppose that is possible with Batch 6.  I was looking for them saying, "limited number of them" but I could not find it.  Only time will tell when we see B7, B8, etc...

[Biodom]

This board with less chips, lower hashrate, but with greater power usage, My guess would be this was the original batch 1 that just absolutely failed expectations and resulted a re-design leading to the boards that are in the B1,B2,B3 etc .... They did say they had a very limited number of them.

I suppose that is possible with Batch 6.  I was looking for them saying, "limited number of them" but I could not find it.  Only time will tell when we see B7, B8, etc...

I disagree, it was B4/B5 that was limited. this one is not.
Also, boards are completely different with 135 vs 164 chips.
I believe that from now on it will be just B6 or alike.
Perhaps, this configuration is more amenable for the next gen of chips.

With btc at 388, I don't know, but if we correct to $325 and btc price/machine stays, I would be getting some.
Besides, you can safely use worse PSU's if you got them.

[dmwardjr]

I disagree, it was B4/B5 that was limited. this one is not.
Also, boards are completely different with 135 vs 164 chips.
I believe that from now on it will be just B6 or alike.
Perhaps, this configuration is more amenable for the next gen of chips.

With btc at 388, I don't know, but if we correct to $325 and btc price/machine stays, I would be getting some.
Besides, you can safely use worse PSU's if you got them.

Well, if your prediction is correct, I hope these are actually 1,042 watts instead of an additional 10% on top of that.  

However, f it is another 10% on top of that, it's not a deal breaker for me.

[bbOOmm]

This board with less chips, lower hashrate, but with greater power usage, My guess would be this was the original batch 1 that just absolutely failed expectations and resulted a re-design leading to the boards that are in the B1,B2,B3 etc .... They did say they had a very limited number of them.

I suppose that is possible with Batch 6.  I was looking for them saying, "limited number of them" but I could not find it.  Only time will tell when we see B7, B8, etc...

I disagree, it was B4/B5 that was limited. this one is not.
Also, boards are completely different with 135 vs 164 chips.
I believe that from now on it will be just B6 or alike.
Perhaps, this configuration is more amenable for the next gen of chips.

With btc at 388, I don't know, but if we correct to $325 and btc price/machine stays, I would be getting some.
Besides, you can safely use worse PSU's if you got them.

Ack, my bad. I forgot these were batch 6. I had B4 abd B5 on the mind with the limited.

I have no clue then. Different design, different specifications, should have a different modeling number.... Its already a nightmare to figure out what S7's are what just with all the differences between B1 thru B5, now add in B6 with a different board design and specifications - then looking the same from the exterior..... buyer beware when buying from a re-seller or buying used.

[RichBC]

I disagree, it was B4/B5 that was limited. this one is not.
Also, boards are completely different with 135 vs 164 chips.
I believe that from now on it will be just B6 or alike.
Perhaps, this configuration is more amenable for the next gen of chips.

With btc at 388, I don't know, but if we correct to $325 and btc price/machine stays, I would be getting some.
Besides, you can safely use worse PSU's if you got them.

Well, if your prediction is correct, I hope these are actually 1,042 watts instead of an additional 10% on top of that.  

However, f it is another 10% on top of that, it's not a deal breaker for me.

Assuming it's still a string design now with 15 Nodes the Core voltage has increased from 0.66V to 0.8V. If you believe the BM1385 data Sheet, which I am now uncertain of, then the J/GH will be worse than 0.3J/GH, that translates to at least a 20% increase in power to 1250W.

However too many if's and unknowns in the equation for my liking...


Rich

[dmwardjr]

Assuming it's still a string design now with 15 Nodes the Core voltage has increased from 0.66V to 0.8V. If you believe the BM1385 data Sheet, which I am now uncertain of, then the J/GH will be worse than 0.3J/GH, that translates to at least a 20% increase in power to 1250W.

However too many if's and unknowns in the equation for my liking...

Rich

I was seeing the same thing, Rich.  Which has me wondering if the "+10%" is correct.  Meaning, 1042 watts + 10% is also based off of the lowest voltage of 11.6 Volts and not the highest of 13.0 Volts.  At 11.6 Volts the core voltage is 0.77V for 15 strings.  At 12.0 Volts the core voltage is like you said, "0.8V for 15 strings."  

So, it appears they may have kept the original specifications for 162 chips with 18 strings and simply want us to add 10% to those original specifications now that they are using 135 chips in 15 strings.  At least that's the way I'm understanding it at the moment.

[dontetris]

What percentage of hardware errors did this give you?

One miner was giving me lots of HW errors and when i removed the fans to check the heatsinks, surprise...one missging

http://s27.postimg.org/d65o6xw1b/20151107_094758.jpg

http://s27.postimg.org/utivregkv/20151107_094816.jpg

http://s27.postimg.org/o4gxp4kmn/20151107_094840.jpg

What can i do ? Write to Bitmain or  buy heatsinks? Where from ?

[UfoRia]

Anyone come up with a power cycle schedule for s7's? I have noticed that if they run for more then five days they tend to slow, but if I hard cycle them, then soft cycle them they are good for about five days.

I have good power, so I know it's not power related. Is it firmware issues or something with the hashing boards over time.

Is my observations of metrics an anomaly or did I miss a thread?

Thanks in advance for third party insights!



Ufo