takagari
Legendary
Offline
Activity: 1050
Merit: 1000
|
|
February 11, 2015, 06:41:09 PM |
|
The voltage and draw on the SP20, is 3280ish watts per plug or per pin? Trying to see what 14awg wire will handle
|
|
|
|
samsonn25
|
|
February 11, 2015, 07:24:28 PM |
|
The voltage and draw on the SP20, is 3280ish watts per plug or per pin? Trying to see what 14awg wire will handle 14 awg will handle more than the pins or connectors Housing them.
|
|
|
|
mavericklm
|
|
February 11, 2015, 07:24:49 PM |
|
max is 288watt/ 6pin connector
official recommendation 16awg, 40cm long
personally i would go with a bit thicker, like 15 or 14awg
|
|
|
|
klondike_bar
Legendary
Offline
Activity: 2128
Merit: 1005
ASIC Wannabe
|
|
February 11, 2015, 07:27:37 PM |
|
Nope. This will be a HUGE mistake if done. One can source 10-20-50 PSUs, but how about a few thousand of PSUs? We're out of stock on the SP35.
So next gen soon? Use a built in psu that's common like the dps2000, and ship without them? let us buy our own, Big savings there +1/ plus, having the users install the PSU makes troubleshooting more difficult since right now SP-T can verify that the SP3x units are shipping with a working PSU and act as inended. allowing users to dump used PSUs into the unit opens up a big margin for error also, the DPS2000 would be tough, since its airflow goes ACROSS the PSU, rather than down its legth like the current artesyn/emerson/murata PSU style. anyways, the PSUs SP-T is using are basically common-slot already. This is visible from the fact that they have multiple power ratings and brand names to choose from when assembling and shipping the SP3x series
|
|
|
|
takagari
Legendary
Offline
Activity: 1050
Merit: 1000
|
|
February 12, 2015, 03:56:37 AM |
|
The voltage and draw on the SP20, is 3280ish watts per plug or per pin? Trying to see what 14awg wire will handle 14 awg will handle more than the pins or connectors Housing them. Well, I was more interested in having the pins come out on 16awg for a few inches, than having the three splice together into a single line. of a heavier awg.
|
|
|
|
opentoe
Legendary
Offline
Activity: 1274
Merit: 1000
Personal text my ass....
|
|
February 12, 2015, 05:52:26 AM |
|
dps2000bb is very nice psu! the 2.88kw looks good too but is too noisy
You run your rigs at home? Only reason why a PSU would be too loud, right?
|
|
|
|
Guy Corem (OP)
Donator
Legendary
Offline
Activity: 1414
Merit: 1051
Spondoolies, Beam & DAGlabs
|
|
February 12, 2015, 06:47:16 AM |
|
another power cut - another non-responsive SP30
5 x SD recovery attempts
5 x 'xmas' alternating leds
0 x recovered
kinda sick of babysitting this thing - the SP20/SP20E didnt miss a beat, so wtf is wrong with the SP30.....again?!
It seems like a problem with the management board, maybe the emmc. If you're ok replacing it by your own, contact support@
|
|
|
|
RoadStress
Legendary
Offline
Activity: 1904
Merit: 1007
|
|
February 12, 2015, 07:29:39 AM |
|
another power cut - another non-responsive SP30
Did you try the boot method? Maybe the mng board got corrupted somehow and it just needs a uSD card as boot medium. Happened to one of my miners in Advania.
|
|
|
|
aclass
|
|
February 12, 2015, 07:38:01 AM |
|
why is there such a big difference between frequencies of the loops: LOOP[0] ON TO:0 (w:149) 0: DC2DC/1/:[vlt1:636 vlt2:642(DCl:794 Tl:794 Ul:642) 54W 84A 38c] ASIC:[ 85c (125c) 710hz(BL: 710) 72 (E:193) F:0 L:0] 1: DC2DC/1/:[vlt1:638 vlt2:642(DCl:794 Tl:794 Ul:642) 53W 83A 51c] ASIC:[ 85c (125c) 690hz(BL: 690) 77 (E:193) F:0 L:0] LOOP[1] ON TO:0 (w:140) 2: DC2DC/1/:[vlt1:638 vlt2:642(DCl:794 Tl:794 Ul:642) 57W 89A 63c] ASIC:[ 85c (125c) 770hz(BL: 770) 79 (E:193) F:0 L:0] 3: DC2DC/1/:[vlt1:638 vlt2:642(DCl:794 Tl:794 Ul:642) 65W 102A 71c] ASIC:[ 85c (125c) 820hz(BL: 820) 61 (E:193) F:0 L:0] LOOP[2] ON TO:0 (w:153) 4: DC2DC/1/:[vlt1:638 vlt2:642(DCl:794 Tl:794 Ul:642) 57W 89A 38c] ASIC:[ 85c (125c) 740hz(BL: 740) 81 (E:193) F:0 L:0] 5: DC2DC/1/:[vlt1:636 vlt2:642(DCl:794 Tl:794 Ul:642) 56W 88A 52c] ASIC:[ 85c (125c) 740hz(BL: 740) 72 (E:191) F:0 L:0] LOOP[3] ON TO:0 (w:163) 6: DC2DC/1/:[vlt1:638 vlt2:642(DCl:794 Tl:794 Ul:642) 61W 95A 59c] ASIC:[ 85c (125c) 790hz(BL: 790) 85 (E:193) F:0 L:0] 7: DC2DC/1/:[vlt1:638 vlt2:642(DCl:794 Tl:794 Ul:642) 62W 96A 66c] ASIC:[ 85c (125c) 800hz(BL: 800) 78 (E:193) F:0 L:0]
the loops that get hotter (the back ones) get clocked higher and use more power
|
|
|
|
RoadStress
Legendary
Offline
Activity: 1904
Merit: 1007
|
|
February 12, 2015, 07:46:05 AM |
|
i havent tried the boot method yet this time. although it didnt seem to help the last time this one failed after a brief power cut.
as to replacing management board/emmc? i am confident if its relatively simple.
Give it a try
|
|
|
|
PatMan
|
|
February 12, 2015, 10:38:54 AM |
|
5 x SD recovery attempts
5 x 'xmas' alternating leds
0 x recovered
kinda sick of babysitting this thing - the SP20/SP20E didnt miss a beat, so wtf is wrong with the SP30.....again?!
Same problem here, but with an SP20E. It just disappeared from my network, stopped hashing & became inaccessible via ssh, gui or myminer.io. Tried the same as you above as well as the boot SD - nothing. Awaiting response from Barbara & tech support......
|
|
|
|
jtoomim
|
|
February 12, 2015, 11:24:03 AM |
|
Well, I was more interested in having the pins come out on 16awg for a few inches, than having the three splice together into a single line. of a heavier awg.
I've done this before on some of our early prototype PSUs. It works okay. It's not worth the time it takes to put it together, though, in my opinion, especially if you find a decent source for 16awg PCIE cables that are long enough etc.
|
Hosting bitcoin miners for $65 to $80/kW/month on clean, cheap hydro power. http://Toom.im
|
|
|
jtoomim
|
|
February 12, 2015, 11:40:51 AM |
|
I thought the original spec from Molex for the pins in the PCIe socket was the same as the original ATX socket, and I thought that was 5 amps. Does anybody have a more accurate recollection or specification on that? At 240W, that's 20 amps split across 3 pins, and also 3 wires. That's just shy of 7 amps per pin/wire, assuming all is well balanced.
It was originally 8 amps per pin, limited by the Molex connector, for 288 W. More recent PSUs (since 2005) are required to use 11 amp connectors, for 396 W. This assumes 3 pins for each of ground and 12V. The PCIE spec only requires 2 of each to be connected; although many PSUs include a 3rd of each, it is not universal. http://www.tomshardware.com/forum/274631-28-power-spec-power-plug#2674391The original PCIE power spec as a whole only specified 2 amps per pin, or 75 W total. The 8-pin revision of the spec specified 4 amps, or 150 W total. This is much less than the specifications of the individual components of the connector, which is why we disregard those limits entirely. http://www.playtool.com/pages/psuconnectors/connectors.html#pciexpress
|
Hosting bitcoin miners for $65 to $80/kW/month on clean, cheap hydro power. http://Toom.im
|
|
|
zvisha
|
|
February 12, 2015, 01:38:34 PM |
|
5 x SD recovery attempts
5 x 'xmas' alternating leds
0 x recovered
kinda sick of babysitting this thing - the SP20/SP20E didnt miss a beat, so wtf is wrong with the SP30.....again?!
Same problem here, but with an SP20E. It just disappeared from my network, stopped hashing & became inaccessible via ssh, gui or myminer.io. Tried the same as you above as well as the boot SD - nothing. Awaiting response from Barbara & tech support...... If you get xmas lights, this means that the SD card operational. this means that you should be able to use the boot-sd (even if the mmc got corrupted) and run from SD-card Even in the x-mass state the platform should be available via SSH.
|
|
|
|
PatMan
|
|
February 12, 2015, 01:42:09 PM |
|
5 x SD recovery attempts
5 x 'xmas' alternating leds
0 x recovered
kinda sick of babysitting this thing - the SP20/SP20E didnt miss a beat, so wtf is wrong with the SP30.....again?!
Same problem here, but with an SP20E. It just disappeared from my network, stopped hashing & became inaccessible via ssh, gui or myminer.io. Tried the same as you above as well as the boot SD - nothing. Awaiting response from Barbara & tech support...... If you get xmas lights, this means that the SD card operational. this means that you should be able to use the boot-sd (even if the mmc got corrupted) and run from SD-card Even in the x-mass state the platform should be available via SSH.
That's the problem - I no longer get the xmas lights. I'm on skype now waiting for a reply from spondoolies-tech support 1 EDIT: They just went offline...........
|
|
|
|
zvisha
|
|
February 12, 2015, 03:01:59 PM |
|
5 x SD recovery attempts
5 x 'xmas' alternating leds
0 x recovered
kinda sick of babysitting this thing - the SP20/SP20E didnt miss a beat, so wtf is wrong with the SP30.....again?!
Same problem here, but with an SP20E. It just disappeared from my network, stopped hashing & became inaccessible via ssh, gui or myminer.io. Tried the same as you above as well as the boot SD - nothing. Awaiting response from Barbara & tech support...... If you get xmas lights, this means that the SD card operational. this means that you should be able to use the boot-sd (even if the mmc got corrupted) and run from SD-card Even in the x-mass state the platform should be available via SSH.
That's the problem - I no longer get the xmas lights. I'm on skype now waiting for a reply from spondoolies-tech support 1 EDIT: They just went offline........... contact me (zvi shteingart)
|
|
|
|
samsonn25
|
|
February 12, 2015, 05:03:01 PM |
|
I thought the original spec from Molex for the pins in the PCIe socket was the same as the original ATX socket, and I thought that was 5 amps. Does anybody have a more accurate recollection or specification on that? At 240W, that's 20 amps split across 3 pins, and also 3 wires. That's just shy of 7 amps per pin/wire, assuming all is well balanced.
It was originally 8 amps per pin, limited by the Molex connector, for 288 W. More recent PSUs (since 2005) are required to use 11 amp connectors, for 396 W. This assumes 3 pins for each of ground and 12V. The PCIE spec only requires 2 of each to be connected; although many PSUs include a 3rd of each, it is not universal. http://www.tomshardware.com/forum/274631-28-power-spec-power-plug#2674391The original PCIE power spec as a whole only specified 2 amps per pin, or 75 W total. The 8-pin revision of the spec specified 4 amps, or 150 W total. This is much less than the specifications of the individual components of the connector, which is why we disregard those limits entirely. http://www.playtool.com/pages/psuconnectors/connectors.html#pciexpressWell said.
|
|
|
|
klondike_bar
Legendary
Offline
Activity: 2128
Merit: 1005
ASIC Wannabe
|
|
February 12, 2015, 05:42:03 PM |
|
Well, I was more interested in having the pins come out on 16awg for a few inches, than having the three splice together into a single line. of a heavier awg.
I've done this before on some of our early prototype PSUs. It works okay. It's not worth the time it takes to put it together, though, in my opinion, especially if you find a decent source for 16awg PCIE cables that are long enough etc. *cough* shameless self-promotion *cough*
|
|
|
|
RoadStress
Legendary
Offline
Activity: 1904
Merit: 1007
|
|
February 12, 2015, 06:04:45 PM |
|
*cough* shameless self-promotion *cough*
I'm fine with it. At least he can deliver.
|
|
|
|
bigblind
|
|
February 12, 2015, 06:22:58 PM |
|
I've ordered today! That's like getting into mining for the first time again . Hopefully ASIC mining will be fun as GPU mining was.
|
|
|
|
|