ANTMINER S5: 1155GH(+OverClock Potential), In Stock $0.319/GH & 0.51W/GH

[cvokey]

So what could it run if the room is about -30C or lower?

Just curious, how do you manage that temperature?

they will be in an uninsulated garage which is rather drafty in Canada. and here it is common to get temps below -30C or even lower without a windchill.  so the room will easily reach those temps soon.

I have 6 S4's running there now and they don't even pass 40C with the fans running slow

[1715507083]

1715507083

[dogie]

Overnight, one of the blades curiously went xxxxxx xxxxxxxx xxxxxxxx xxxxxxo running at stock 350 with the temps at 63C. Rebooted and it's working OK again.

Hmmm.....I got back from New Years Eve celebrations earlier and noticed the same thing with mine, but all X's. They all came back after a reboot also, but it is a little worrying.....   Anyone else had this?

EDIT: Also at stock freq with temps @ 65.

that's alarming...

There is a known firmware bug where one module drops out but both still display as mining and temperatures fixed. This may be the same bug and will be fixed in the next one. And yes, its very much firmware.

[dogie]

So what could it run if the room is about -30C or lower?

Condensation, so... 0GH

Am I right in thinking that your top SP31 is sucking in air from the PSU side, but the SP20s and S5 are doing the opposite and throwing air out?

Doesn't that get a little 'circular' ?

no they are all blowing the same direction

You can quickly remember which way a fan is blowing as they always blow towards their frame.

[dogie]

not sure i would agree. The temperatures on my unit are just fine, and Im not really concerned at all with the heat of the unit. I run the fan on 5V and aim to have the board temperatures averaging around 60C. With a nearby window cracked slightly an intake of 5C allows me to overclock to 381.25MHz/1255GH without tripping my 612.5W PSU rail (fan is on a 5V rail so that frees up 12W+ for mining). With room-temperature (~20C) intake I have to operate at or slightly under the stock 350MHz.

The plastic sides would likely not deform until a minimum of 100C, and actual melting temp has to be well over 150C, or the temperature that the antminer shuts itself off or would otherwise be failing massively (which nobody has experienced yet).

keep in mind that the SP20 temperature is more air temperature than the board. Asic stats show that the chips are 85C underclocked or up to 125C in normal operation. Antminer chips go into the 75-95C range I believe, and the sensors on it are likely in contact with the hot side of the heatsink. 55-60C is an ideal temperature for the S5, above 65C may spell trouble

that said, 80C is very high, is your garage warm? I found that if you put a cover on the unit air crosses the heatsink instead of blowing out the top around the beaglebone.

Polypropylene is 130C and Polycarbonate is 155C but you're going to need the AIR temperature to get up to that!

[raskul]

So what could it run if the room is about -30C or lower?

Condensation, so... 0GH

Am I right in thinking that your top SP31 is sucking in air from the PSU side, but the SP20s and S5 are doing the opposite and throwing air out?

Doesn't that get a little 'circular' ?

no they are all blowing the same direction

You can quickly remember which way a fan is blowing as they always blow towards their frame.

oot the open door at the front of the garage  

[philipma1957]

Well I have decided enough of the fan is too loud and I am now going overclock.

I placed the loud fan back in the original spot opened a window in the bedroom it runs in. The intake air is now about 50f or 10 c

I added a second fan a 3 pin delta the can blow 148 cfm  since it is 3 pin it runs 2800 rpm to 3200 rpm and I put the S-5  at freq 375

the noise is nuts and I won't be able to run this gear with this noise,but many people can run the gear in a shed or a barn and do not care about noise.

this is a 10 minute run at freq 375  very nice numbers

watts are 636 at the kwatt

636/1224 = that is .519 watts a gh damn good numbers  it is on an evga 1300 g2  psu.  and part of that is 2 fans vs 1 fan

[raskul]

Well I have decided enough of the fan is too loud and I am now going overclock.

I placed the loud fan back in the original spot opened a window in the bedroom it runs in. The intake air is now about 50f or 10 c

I added a second fan a 3 pin delta the can blow 148 cfm  since it is 3 pin it runs 2800 rpm to 3200 rpm and I put the S-5  at freq 375

the noise is nuts and I won't be able to run this gear with this noise,but many people can run the gear in a shed or a barn and do not care about noise.

this is a 10 minute run at freq 375  very nice numbers

http://imagizer.imageshack.us/v2/1024x768q90/661/gOZWdC.png

mine gave a board of all x's when i tried it at 375 

[philipma1957]

Well I have decided enough of the fan is too loud and I am now going overclock.

I placed the loud fan back in the original spot opened a window in the bedroom it runs in. The intake air is now about 50f or 10 c

I added a second fan a 3 pin delta the can blow 148 cfm  since it is 3 pin it runs 2800 rpm to 3200 rpm and I put the S-5  at freq 375

the noise is nuts and I won't be able to run this gear with this noise,but many people can run the gear in a shed or a barn and do not care about noise.

this is a 10 minute run at freq 375  very nice numbers

http://imagizer.imageshack.us/v2/1024x768q90/661/gOZWdC.png

mine gave a board of all x's when i tried it at 375  

do a reboot .  one issue is this seems to happen on higher clocks  and a system reboot via gui has seemed to work.

 this miner looks good to overclock in a cold room  with a second fan .  and thinking fuck the noise.


I will boost to 412 in a bit.

  I think in my setup I would max out with that number.

Stock fan in the push  and a delta 3 pin in the pull. 
I think the stock fan at 4000 rpm moves 200cfm the delta  at 3000 rpm in the pull  moves 148cfm
At these speed it is not a home miner. it is about 79db and nasty  sounding.

But  the numbers below are really nice

[raskul]

Well I have decided enough of the fan is too loud and I am now going overclock.

I placed the loud fan back in the original spot opened a window in the bedroom it runs in. The intake air is now about 50f or 10 c

I added a second fan a 3 pin delta the can blow 148 cfm  since it is 3 pin it runs 2800 rpm to 3200 rpm and I put the S-5  at freq 375

the noise is nuts and I won't be able to run this gear with this noise,but many people can run the gear in a shed or a barn and do not care about noise.

this is a 10 minute run at freq 375  very nice numbers

http://imagizer.imageshack.us/v2/1024x768q90/661/gOZWdC.png

mine gave a board of all x's when i tried it at 375  

do a reboot .  one issue is this seems to happen on higher clocks  and a system reboot via gui has seemed to work.

 this miner looks good to overclock in a cold room  with a second fan .  and thinking fuck the noise.


I will boost to 400 in a bit

to be honest i'm happy at 980+GH/s at 300. temps are comfortable - and i don't need to worry about the noise as it's outside in the cold

[dogie]

Well I have decided enough of the fan is too loud and I am now going overclock.

I placed the loud fan back in the original spot opened a window in the bedroom it runs in. The intake air is now about 50f or 10 c

I added a second fan a 3 pin delta the can blow 148 cfm  since it is 3 pin it runs 2800 rpm to 3200 rpm and I put the S-5  at freq 375

the noise is nuts and I won't be able to run this gear with this noise,but many people can run the gear in a shed or a barn and do not care about noise.

this is a 10 minute run at freq 375  very nice numbers

http://imagizer.imageshack.us/v2/1024x768q90/661/gOZWdC.png

mine gave a board of all x's when i tried it at 375 

Something which will happen on these miners that we haven't seen before is PSUs matter more than ever. There is going to a be a big difference in power consumption and overclockability depending if a perosn's PSU is 12.2V under load or 11.6V.

[philipma1957]

Well I have decided enough of the fan is too loud and I am now going overclock.

I placed the loud fan back in the original spot opened a window in the bedroom it runs in. The intake air is now about 50f or 10 c

I added a second fan a 3 pin delta the can blow 148 cfm  since it is 3 pin it runs 2800 rpm to 3200 rpm and I put the S-5  at freq 375

the noise is nuts and I won't be able to run this gear with this noise,but many people can run the gear in a shed or a barn and do not care about noise.

this is a 10 minute run at freq 375  very nice numbers

http://imagizer.imageshack.us/v2/1024x768q90/661/gOZWdC.png

mine gave a board of all x's when i tried it at 375  

Something which will happen on these miners that we haven't seen before is PSUs matter more than ever. There is going to a be a big difference in power consumption and overclockability depending if a perosn's PSU is 12.2V under load or 11.6V.

yeah using the single evga 1300 g2 for just the one miner I am getting over 1307gh the mod was add a decent delta  purchased on ebay for 10 bucks.

I can not  run this at the noise level it does  but many can.

At 1307 gh and watts at the kwatt meter at 692  692/1307 = .529 watts per gh at the kwatt meter





1 hour test at 412 is good.

[Guy Corem]

...
regardless, i am 100% certain the cable was at fault, not the sp20e as it isn't even OCd.

upon replacement of the cable its back running at 1430-1510GH/s @ 940W



Temp Front / Back T,B
    31 °C / 71,72 °C
Fan Speed
    90
Start Voltage
    0.67 / 0.67 / 0.67 / 0.67
Max Voltage
    0.7
Max Watts
    215 / 215 / 215 / 215

Thank you for clarifying.

[Xian01]

... speaking of overclocking, here's some figures from overnight testing. This is using the stock push fan, with an additional pull fan added (although I question whether the pull fan is making any difference at all...)

I've found, in my testing so far, that keeping the board temperatures <60C is required for stability. Room ambient temperature is 24C where the miners are working.

[philipma1957]

  okay fast opinion on the two gears (s-5 , sp-20)  safety .

both are far safer then AM's prisma was when it first came out.

both can fuck up by melting cables.

melting pcie cables happen for more then one reason.

1) bad management by the operator.  ie overloading a cable beyond its specs.
2) poor cables
3) defective cables
4) defective psu's

I can see multiple ways to set up the s-5 or the sp20 that result in overloaded cables. --- that is operator error.

And we all know some psu's are not good
and we all know some cables are not good.

frankly use common sense as both units suck down a lot of power on the pcie cables.

[philipma1957]

...
regardless, i am 100% certain the cable was at fault, not the sp20e as it isn't even OCd.

upon replacement of the cable its back running at 1430-1510GH/s @ 940W



Temp Front / Back T,B
    31 °C / 71,72 °C
Fan Speed
    90
Start Voltage
    0.67 / 0.67 / 0.67 / 0.67
Max Voltage
    0.7
Max Watts
    215 / 215 / 215 / 215

Thank you for clarifying.

How does the current limiting on the PCIe plugs work, BTW? I'm not at home to check, but I would assume that the two plugs are not simply paralleled on the hashing board and that each 12V plug powers two of the ASICs in a loop.
IE, if you disabled a full loop on a hashing board you'd pull full power from one plug, and ~0W from the other.

Is that correct?

the sp20 seems to be 1 plug per loop.  so if you knock a plug loose on 1 board(each board has 2 loops)  the board only allows 1 loop to work.  this is good and it is safe.

I have knocked loose a pcie cable at the psu end and the sp20 just  ran 1 loop on that board.

[dlowings]

bitmain, if your going to remove the listings of batch 1 through batch whatever, can you please find a way to leave the ship date for each batch . I ordered batch two and went back to the page to reference the batch two ship date and its gone … I'm fine with removing it if the ship date has passed but by removing the ship date prior to the actual ship date makes it look as if you are trying to make the ship date a moving target based on your needs as opposed to customer needs .

[Dr Charles]

bitmain, if your going to remove the listings of batch 1 through batch whatever, can you please find a way to leave the ship date for each batch . I ordered batch two and went back to the page to reference the batch two ship date and its gone … I'm fine with removing it if the ship date has passed but by removing the ship date prior to the actual ship date makes it look as if you are trying to make the ship date a moving target based on your needs as opposed to customer needs .

Is the ship date not on your order confirmation or invoice?

[klondike_bar]

How does the current limiting on the PCIe plugs work, BTW? I'm not at home to check, but I would assume that the two plugs are not simply paralleled on the hashing board and that each 12V plug powers two of the ASICs in a loop.
IE, if you disabled a full loop on a hashing board you'd pull full power from one plug, and ~0W from the other.

Is that correct?

the sp20 seems to be 1 plug per loop.  so if you knock a plug loose on 1 board(each board has 2 loops)  the board only allows 1 loop to work.  this is good and it is safe.
I have knocked loose a pcie cable at the psu end and the sp20 just  ran 1 loop on that board.

problem is that the S5 will put about 270-320W per blade, so if you use two seperate cables thats a managable 150W/cable. If one fails, that 300W will push a cable to its limit. 16awg can handle that but an 18awg will certainly fail.

in comparison, the SP20 can draw up to 288W PER CABLE at full speed. (~150W/cable underclocked). Not much difference in safety as far as I see, if anything the SP20 is more likely to cause PSU failures (in lower quality supplies).

dont use a Bronze or lower PSU for either device, thats my suggestion. GOLD or PLAT power supplies will have better safety features and use thick wiring in its cables. Alternatively, I sell 16AWG PCIe cables that have either a connector on each end, or a connector at one end and stripped wires for server mods on the other. Ive used them to carry up to 312W each at which point they get warm but not hot

[IITravel01]

Thought my pair of S5's would ship on 08 Jan. but got notice they shipped yesterday (via BitMain on Amazon).

[pekatete]

Quick, clean and easy fan muffing for the those that have an S3+.
Swap the push fan from an S3+ with the stock S5 fan. At 3500 rpm, the S3+ fan hums (or purrs) whereas the stock S5 fan whines ..
On the orher hand, the stock S5 fan will run at a much lower rpm when strapped onto an S3, thus won't be as loud as when it is on an S5.