ANTMINER C1 Discussion and Support Thread

[grn]

SYSCOOLING

What is status of order 2014123011433?

It was shipped on 30,December,by DHL.Do you have the tracking NO?

no i don't, could you please provide? (pm or email)

[1714470881]

1714470881

[1714470881]

1714470881

[1714470881]

1714470881

[1714470881]

1714470881

[1714470881]

1714470881

[1714470881]

1714470881

[soy]

Interesting.  36.9°F outside.  Inside is 63.9°.  Maybe it's my connection or maybe it's the pool but my hashrate goes though a long slow rise then fall and so on.  The temperature of the water out port of the water block is cool to the touch while the air pushed thru the C1 is quite warm on exit, warmer than the exiting water.  I just unplugged the single fan so the two closest radiator fans to the ports run.  I wonder if letting the C1 mounted fan run at 12v continually would be a good idea.  Since disconnecting the single radiator fan the C1 mounted fan has started changing speed fairly often.  The hashrate has come up but that might just be the observed cycle.

Odd.  GUI Fan2 speed indication never changes from 1200 although I hear it increasing and decreasing.

If its annoying you can try plugging the fan into the other header that it can reach. You'll have to take the case off unfortunately (and all the PCI-E's) to get to it though.

You sure you pump seems to be running okay? I can't think of a reason why the fluid exhaust would be cooler than the air exhaust.

I think problem might be there is gap between waterblock and hash board.You could have a try to open C1 and make sure whether every screw on board is tight.

Considering it.  Just for practice today, since I have two S3+'s that work well tho one a little better than the other, I ripped apart the slightly slower S3+ and reworked the thermal paste.  There was a wealth of thermal paste there unlike my S3's which I reworked.  So my new thermal paste didn't make a great deal of difference.  It is still the slower of the two S3+'s but it's been producing more shares than the faster most of the afternoon and evening.  It may have just been the removal of accumulated dust on the controller board as much as the new thermal paste.

I plan to rip the C1 down and replace the thermal paste but I used most of what I had and just ordered more.

[RadekG]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

C1 at 100-275MHz has following chart, it should be similar as on other devices based on BM1382 chips - S3, U3 and S4.

100MHz @ 0.625V
(ASICs do not work below this voltage)  :   192W =0.48W/GH
175MHz @ 0.650V                              :   356W =0.51W/GH
250MHz @ 0.720V                              :   650W =0.65W/GH
250MHz @ 0.760V                              :   662W =0.66W/GH
275MHz @ 0.760V                              :   809W =0.74W/GH
275MHz @ 0.800V                              :   900W =0.82W/GH
300MHz @ 0.815V                              :  1052W =0.88W/GH

Hope anyone is interested.

Please note that best W/GH ratio can be reached by lowering input 12V voltage, such as 9V, but highest eff. power supply can be done only at high voltages (12V and up) which can also dismiss conducting losses between PSU and miner.

This measurement was done at miner, not at the wall.
 

[dogie]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

Really wouldn't recommend it. Even if the chips can make it up there, you don't know about the other components on the board (which also aren't watercooled).

[soy]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

C1 at 100-275MHz has following chart, it should be similar as on other devices based on BM1382 chips - S3, U3 and S4.

100MHz @ 0.625V
(ASICs do not work below this voltage)  :   192W =0.48W/GH
175MHz @ 0.650V                              :   356W =0.51W/GH
250MHz @ 0.720V                              :   650W =0.65W/GH
250MHz @ 0.760V                              :   662W =0.66W/GH
275MHz @ 0.760V                              :   809W =0.74W/GH
275MHz @ 0.800V                              :   900W =0.82W/GH
300MHz @ 0.815V                              :   not yet measured

Hope anyone is interested.

Please note that best W/GH ratio can be reached by lowering input 12V voltage, such as 9V, but highest eff. power supply can be done only at high voltages (12V and up) which can also dismiss conducting losses between PSU and miner.

This measurement was done at miner, not at the wall.
 

Interesting.  I may give your suggestion a test eventually.  One thing I did notice with my unnamed cheap adjustable power supplies, since I flipped my rad fans so I could have the ports and wires inside a window, and I note it was raining when I did the flip, my power at the wall has gone from 860 watts to 880 watts.  I check my voltage at the miner and it's 12.08vdc same as it had been when I set it up (setting the voltage at 12.15vdc no load), then I checked the wall voltage.  The wall voltage is at 118.1.  When I shut down the miner it comes back to 120.1vac.  The miner is on one of two lines I ran to the living room for AC and my treadmill.  I'm kicking myself for not running 12-2 Romex instead of 14-2.  These two lines run to to a pair of half size circuit breakers.  The room temperature has come up from mid-60's to high 70's.  Outside temperature has come up 20° since the low power days of 860 watts at the wall (includes the pump).  So, when it's raining and the indoor humidity is 56.5%, cooling is reduced (low mass air drawing away less heat) and efficiency drops.  Also perhaps my cheap power supplies efficiencies drop in a warmer room and high humidity.

I'm running the C1 on 4 PCI-E connectors.  I had ordered another 4 of the connectors late and they arrived at my post office box the other day and I'm about to run into town for those.  So, later today the C1 will be running on 8 PCI-E lines.  Presently the leads to the 4 connectors are warm, dropping voltage.  So, before I install the new leads I'll carefully check the power at the wall.  If it's as you indicate, the reduced voltage drop between my supplies and the miner should cause an increase in power consumption - unless the efficiency of my power supplies suffer as the miner improves - the supplies are running at ½ full load each.

[soy]

Odd.  I ssh'd into the C1 some time back and changed the root password using the unix command passwd, logged out and logged back in using the new password.  Today I went to log in and the password had reverted to admin while the GUI was still connected and unchanged.

[soy]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

C1 at 100-275MHz has following chart, it should be similar as on other devices based on BM1382 chips - S3, U3 and S4.

100MHz @ 0.625V
(ASICs do not work below this voltage)  :   192W =0.48W/GH
175MHz @ 0.650V                              :   356W =0.51W/GH
250MHz @ 0.720V                              :   650W =0.65W/GH
250MHz @ 0.760V                              :   662W =0.66W/GH
275MHz @ 0.760V                              :   809W =0.74W/GH
275MHz @ 0.800V                              :   900W =0.82W/GH
300MHz @ 0.815V                              :   not yet measured

Hope anyone is interested.

Please note that best W/GH ratio can be reached by lowering input 12V voltage, such as 9V, but highest eff. power supply can be done only at high voltages (12V and up) which can also dismiss conducting losses between PSU and miner.

This measurement was done at miner, not at the wall.
 

Interesting.  I may give your suggestion a test eventually.  One thing I did notice with my unnamed cheap adjustable power supplies, since I flipped my rad fans so I could have the ports and wires inside a window, and I note it was raining when I did the flip, my power at the wall has gone from 860 watts to 880 watts.  I check my voltage at the miner and it's 12.08vdc same as it had been when I set it up (setting the voltage at 12.15vdc no load), then I checked the wall voltage.  The wall voltage is at 118.1.  When I shut down the miner it comes back to 120.1vac.  The miner is on one of two lines I ran to the living room for AC and my treadmill.  I'm kicking myself for not running 12-2 Romex instead of 14-2.  These two lines run to to a pair of half size circuit breakers.  The room temperature has come up from mid-60's to high 70's.  Outside temperature has come up 20° since the low power days of 860 watts at the wall (includes the pump).  So, when it's raining and the indoor humidity is 56.5%, cooling is reduced (low mass air drawing away less heat) and efficiency drops.  Also perhaps my cheap power supplies efficiencies drop in a warmer room and high humidity.

I'm running the C1 on 4 PCI-E connectors.  I had ordered another 4 of the connectors late and they arrived at my post office box the other day and I'm about to run into town for those.  So, later today the C1 will be running on 8 PCI-E lines.  Presently the leads to the 4 connectors are warm, dropping voltage.  So, before I install the new leads I'll carefully check the power at the wall.  If it's as you indicate, the reduced voltage drop between my supplies and the miner should cause an increase in power consumption - unless the efficiency of my power supplies suffer as the miner improves - the supplies are running at ½ full load each.

So, home with the PCI-E-6 connectors.  The power supplies were seeing 118VAC, wattage 880 at the wall and output of the supplies was 12.09vdc.  The wires from the supply to the C1 were warm.  Only 4 of the 8 connectors on the C1 were occupied.  Completed a pair, shut down the C1, added the 2 new to one side and fired up.  At fire up wattage was 870 and after 20 minutes it was up to 889 watts at the wall.  Finishing the second pair of wires, shut down, added the final pair of leads, now all 8 connectors occupied, and at fire up the wattage was 838 watts at the wall, 118VAC.  After 20 minutes, wall wattage measurement is 856 watts, wall voltage 117.9VAC, power supplies voltages 12.10vdc.

What this is telling me is that the supplies' voltages rose 10 millivolts due to decreased load as a result of reduced voltage drop in the lines from the supplies to the PCI-E connectors.  The higher voltage at the PCI-E connectors allowed more efficient mining.  I was somewhat remiss in failing to record the voltage drop between the supplies and the PCI-E connectors before doubling the wiring.  The rise in the supplies' voltages would have been greater if the AC voltage at the wall had remained at 118VAC instead of dropping to 117.9VAC as the supplies' outputs to some degree track changes in wall AC.  This was shown when wall AC voltage rose exceptionally high when the neighborhood shut down air conditioning on cool evenings.

After an hour the wattage at the wall is varying between 854-857 watts.

[dogie]

Odd.  I ssh'd into the C1 some time back and changed the root password using the unix command passwd, logged out and logged back in using the new password.  Today I went to log in and the password had reverted to admin while the GUI was still connected and unchanged.

Reported to engineers, thanks

[soy]

Odd.  I ssh'd into the C1 some time back and changed the root password using the unix command passwd, logged out and logged back in using the new password.  Today I went to log in and the password had reverted to admin while the GUI was still connected and unchanged.

Reported to engineers, thanks

I haven't opened the C1 yet.  I understand it has an SD card similar to a BBB?  I may have rebooted since first changing the password from the linux command line after ssh'ing in.  Perhaps that password change isn't saved to the compressed files on the card.  I would think this is a security fault if the root password change remains for the GUI login but reverts back to admin for the ssh login.

[notlist3d]

Odd.  I ssh'd into the C1 some time back and changed the root password using the unix command passwd, logged out and logged back in using the new password.  Today I went to log in and the password had reverted to admin while the GUI was still connected and unchanged.

Reported to engineers, thanks

I haven't opened the C1 yet.  I understand it has an SD card similar to a BBB?  I may have rebooted since first changing the password from the linux command line after ssh'ing in.  Perhaps that password change isn't saved to the compressed files on the card.  I would think this is a security fault if the root password change remains for the GUI login but reverts back to admin for the ssh login.

Yes it is very nice on the SD micro card on the controller.   Worst comes to worse you are able to flash a SD card and stick it in.  I believe second post links to it.

[etile]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

C1 at 100-275MHz has following chart, it should be similar as on other devices based on BM1382 chips - S3, U3 and S4.

100MHz @ 0.625V
(ASICs do not work below this voltage)  :   192W =0.48W/GH
175MHz @ 0.650V                              :   356W =0.51W/GH
250MHz @ 0.720V                              :   650W =0.65W/GH
250MHz @ 0.760V                              :   662W =0.66W/GH
275MHz @ 0.760V                              :   809W =0.74W/GH
275MHz @ 0.800V                              :   900W =0.82W/GH
300MHz @ 0.815V                              :   not yet measured

Hope anyone is interested.

Please note that best W/GH ratio can be reached by lowering input 12V voltage, such as 9V, but highest eff. power supply can be done only at high voltages (12V and up) which can also dismiss conducting losses between PSU and miner.

This measurement was done at miner, not at the wall.
 

I have other table for my Antminer C1

250Mhz @0760 0.855 W/GH
225Mhz @0725 0.758 W/GH
200Mhz @0700 0.808 W/GH
175Mhz @0680 0.831 W/GH
150Mhz @0650 0.820 W/GH

This measurement was done at the wall with two Platinum PSU 650 Watt

What firmware you are using and how do you set voltage (web inteface or ssh + vi) ?

[charles2k]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

1,2TH/300MHz - is it stable? How many watts from the wall?

[RadekG]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

1,2TH/300MHz - is it stable? How many watts from the wall?

updated, now it draws 1052W, very stable, HW errors: 0.0000% last 15hours 6HW errors.

[IITravel01]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

1,2TH/300MHz - is it stable? How many watts from the wall?

updated, now it draws 1052W, very stable, HW errors: 0.0000% last 15hours 6HW errors.

At which voltage?  0775 or 0815?

[RadekG]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

Really wouldn't recommend it. Even if the chips can make it up there, you don't know about the other components on the board (which also aren't watercooled).

Disagree and also agree.

Power components are covered by water cooling block except inductors. Of course, you are right that DC-DC convertors are not designed for such big power. But, they should be designed to deliver full load + safety margin at highest operating temperature (80C), so it gives us some safe margin when cooled properly. Unfortunatelly Bitmain didn't replied me what is the maximum and I still didn't disassemble the unit to reverse engineering DC-DC supply.

Of course, I do not advise anyone to overclock his C1 as he is loosing warranty by this.

By the way, as Syscooling pumps are failing, I was pressed to tie two C1 miners together and do other tests.

Conclusion:

2x C1 at one water cooler: temps around 55C at 30C ambient
C1 can be completely cooled without liquid cooler when set at 0.8TH and lower at ambient temp 20C and less.

[charles2k]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

1,2TH/300MHz - is it stable? How many watts from the wall?

updated, now it draws 1052W, very stable, HW errors: 0.0000% last 15hours 6HW errors.

Room and inside miner temperature?

[RadekG]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

1,2TH/300MHz - is it stable? How many watts from the wall?

updated, now it draws 1052W, very stable, HW errors: 0.0000% last 15hours 6HW errors.

Room and inside miner temperature?

I have three OCed C1 at 300MHz:

1) 0.775V, 47HW at 10hrs, 54-56C with very low airflow through cooler.
2) 0.775V, 631HW, 4.5d, 36-38C, ambient at 20C
3) 0.815V, 15hrs, 6HW, 39-41C, ambient 20C

[IITravel01]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

1,2TH/300MHz - is it stable? How many watts from the wall?

updated, now it draws 1052W, very stable, HW errors: 0.0000% last 15hours 6HW errors.

Room and inside miner temperature?

I have three OCed C1 at 300MHz:

1) 0.775V, 47HW at 10hrs, 54-56C with very low airflow through cooler.
2) 0.775V, 631HW, 4.5d, 36-38C, ambient at 20C
3) 0.815V, 15hrs, 6HW, 39-41C, ambient 20C

Wonder if you upped the volt. a tiny bit to say 0780 if you can drop the hw errors down to single or double digits? Or is that over 4.5 days?

[soy]

I am not sure if I didn't missed something, but did anyone do a real OC? I mean OVER 275MHz?

I am currently doing 1.2TH at 300MHz @ 0.775-0.815V, did anybody tried more? how is the GH/W?

C1 at 100-275MHz has following chart, it should be similar as on other devices based on BM1382 chips - S3, U3 and S4.

100MHz @ 0.625V
(ASICs do not work below this voltage)  :   192W =0.48W/GH
175MHz @ 0.650V                              :   356W =0.51W/GH
250MHz @ 0.720V                              :   650W =0.65W/GH
250MHz @ 0.760V                              :   662W =0.66W/GH
275MHz @ 0.760V                              :   809W =0.74W/GH
275MHz @ 0.800V                              :   900W =0.82W/GH
300MHz @ 0.815V                              :   not yet measured

Hope anyone is interested.

Please note that best W/GH ratio can be reached by lowering input 12V voltage, such as 9V, but highest eff. power supply can be done only at high voltages (12V and up) which can also dismiss conducting losses between PSU and miner.

This measurement was done at miner, not at the wall.
 

I have other table for my Antminer C1

250Mhz @0760 0.855 W/GH
225Mhz @0725 0.758 W/GH
200Mhz @0700 0.808 W/GH
175Mhz @0680 0.831 W/GH
150Mhz @0650 0.820 W/GH

This measurement was done at the wall with two Platinum PSU 650 Watt

What firmware you are using and how do you set voltage (web inteface or ssh + vi) ?

Stock batch 2 C1, standard settings.  I don't want overclocking.  I want my miners to run dependably at rated speed with the option to lower settings for reduced power when the btc value gets so low I almost need to stop.

And thank you for the "225Mhz @0725 0.758 W/GH".

[ilostmyface]

With the upcoming diff spike and and the currant BTC price, prices for miners are becoming less attractive. Things are going to get interesting. I was about to pickup another C1 because of the quietness over the new S5s, but with the current conditions buying these miners is starting to become hard to rationalize. I hope BITMAIN continues to make water cooled models when & if things turn around.