E
|
|
June 24, 2016, 08:58:07 PM |
|
I wanted to add some DPS-2000BB efficiency numbers to this thread, as I've recently been measuring DC and AC draw on my S9 at various clock speeds.
Counting 12V cable losses, I'm seeing 88.3 to 88.7% efficiency everywhere from 800W up to 1400W DC. Ignoring cable losses the efficiency is 89.7 to 90%.
This is all at a 208V input voltage... efficiency should be higher with higher input voltage but I haven't been able to test that yet.
Higher input voltage, implies lower input current (Usually). If your losses are dominated by I^2 * R, then it will help. 208V vs 240V efficiency numbers aren't commonly published, but the difference is likely to be less than 0.5% or so. For example, an unspecified HP PSU measured in 2007: BTW I'm making my measurements with a calibrated Xitron 2503AH ( http://www.xitrontech.com/products/general-purpose-power-analyzers/#2503ah) with an external 500A precision shunt for the DC current; the Xitron is good to better than 0.05% on the internal I/V measurements but the shunt is only spec'd to 0.1% so that dominates the error. One cool thing you can observe with this is how current draw from a miner increases by 10s of watts as everything warms up to operating temperature. I'll try to get a video of that because it is pretty interesting to watch!
|
|
|
|
|
|
|
Once a transaction has 6 confirmations, it is extremely unlikely that an attacker without at least 50% of the network's computation power would be able to reverse it.
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
Optimizer (OP)
|
|
June 24, 2016, 09:03:40 PM |
|
208V vs 240V efficiency numbers aren't commonly published, but the difference is likely to be less than 0.5% or so. For example, an unspecified HP PSU measured in 2007: BTW I'm making my measurements with a calibrated Xitron 2503AH ( http://www.xitrontech.com/products/general-purpose-power-analyzers/#2503ah) with an external 500A precision shunt for the DC current. [/quote] Nice. I am jealous! We may rely on your for accuracy. Let me know what numbers you find. @Philipma, be careful when testing to 2500. I have not tested these boards beyond 2kW for an extended period of time. It should be safe, but try not to exceed 130C anywhere on the board. -Optim
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 24, 2016, 09:15:56 PM |
|
Nice. I am jealous! We may rely on your for accuracy. Let me know what numbers you find. @Philipma, be careful when testing to 2500. I have not tested these boards beyond 2kW for an extended period of time. It should be safe, but try not to exceed 130C anywhere on the board. -Optim [/quote] no worries my tests will use 1 miner for each psu which will pull about 1200-1500 watts at the wall I just have the dps1600 1 with your board dps2000 2 in a j4bberwock 4k board dps2500 1 not set up waiting for your second board. basically any of the above will only run 1 s-9 I will over clock and under clock. and take heat measurements plus give my pdu amp readings, but I do not think I will ever exceed 1500 watts at the wall on any of the above psu's
|
|
|
|
Finksy
Legendary
Offline
Activity: 1022
Merit: 1003
|
|
June 27, 2016, 05:18:38 AM |
|
no worries my tests will use 1 miner for each psu which will pull about 1200-1500 watts at the wall
I just have the
dps1600 1 with your board
dps2000 2 in a j4bberwock 4k board
dps2500 1 not set up waiting for your second board.
basically any of the above will only run 1 s-9
I will over clock and under clock. and take heat measurements plus give my pdu amp readings, but I do not think I will ever exceed 1500 watts at the wall on any of the above psu's
The DPS-4K will easily power 2x S9's Phil, and depending on batch it could be closer to 3. Even the IBM 2880W can power 2x S9's.
|
|
|
|
sidehack
Legendary
Offline
Activity: 3318
Merit: 1848
Curmudgeonly hardware guy
|
|
June 27, 2016, 01:21:01 PM |
|
If the boards are built about like mine they'll handle 2500W fine. Before we sold anything we pushed the PSU to about 196A and left it running for a week. I know the rail layering with 6-pins is quite a bit different than for straight screw terminals so the thickness isn't there, but you also get 2-3 times the area, which reduces current density and improves heat radiation.
I'd have no qualms at all powering three S9 off a 4K setup. I know 2xDPS2K works well for 3xS7.
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 27, 2016, 01:30:50 PM |
|
I got the second 2k board today. the dps-2500 comes on tues.
So far 4 s-9s' and 1 s-7 are in the array
2 s-9's on the 4k board = flawless freq 600 2 dps 2000 psu 1 s-9 on optimizer's 2k board = flawless freq 600 a dps 1600 psu 1 s-9 on EVGA 1600 p2 = flawless freq 600
1 s-7 on EVGA 1600 t2 = good freq 600
all have 240 volt ac. the draw for the 5 pieces is about 25 amps
I think I am going to order another s-9 today
we will have 6 units which will be the max for the current setup.
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 29, 2016, 03:50:48 AM Last edit: June 29, 2016, 03:19:52 PM by philipma1957 |
|
I orderedr the fifth s-9 and I have the second board from optimizer. I tested it with my eBay purchase of the dps 2500 psu this psu can draw 2320 watts. but it will run 1 s-9 so should never over tax itself to the board. I will photo it in a few minutes. two fans really cool the psu off ready for the s-9 switch and space for 3 fans two fans this size can do two of these psu's
|
|
|
|
TheYankeesWin!
|
|
June 29, 2016, 03:30:15 PM |
|
I know you mentioned this is a dps 2500 psu. So does this mean this breakout board can run the dps 1600, dps 2000, and dps 2500 psus?
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 29, 2016, 03:36:33 PM |
|
Yes = dps 1600 I do not know = dps -2000 other people have had great results I have yet to test my dps-2000 psus maybe = dps-2500 I have partial results. I have one of these breakout boards working as I type it is using the dps-1600 psu which can provide up to 1800watts. It works great. I have not used it on the dps-2000 psus so I do not know only because I have not tested it. Ohters have and say it works great I have the second breakout board in the photos above. it is using the dps-2500 psu which can provide up to 2320 watts. It has not yet been fully tested it powers on the the dps-2500 the two attach fans turn on.
|
|
|
|
Optimizer (OP)
|
|
June 29, 2016, 04:39:50 PM |
|
I know you mentioned this is a dps 2500 psu. So does this mean this breakout board can run the dps 1600, dps 2000, and dps 2500 psus?
It should work with all of the PSUs matching the IO config. I have tested the 2000 PSUs to max load myself and MarkAZ put a write up on the 2000 test in the original thread posted at the beginning of this article. Phil is covering the rest. -Optim
|
|
|
|
NotFuzzyWarm
Legendary
Online
Activity: 3626
Merit: 2539
Evil beware: We have waffles!
|
|
June 29, 2016, 05:28:03 PM |
|
I know for a fact that the IBM 2kw psu's will comfortably run 3x s5's which is around 1,800w total. From before I updated to s7's https://i.imgur.com/XRlk5Tq.jpgdon't even need fans for them in this config
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 29, 2016, 06:03:29 PM Last edit: June 29, 2016, 06:22:46 PM by philipma1957 |
|
|
|
|
|
toptek
Legendary
Offline
Activity: 1274
Merit: 1000
|
|
June 29, 2016, 07:30:31 PM Last edit: June 29, 2016, 07:46:10 PM by toptek |
|
I was one of the first to buy this i bought three and sold one , i use two right now with the DPS 2000 watt PSU have been for months with out issue if im not mistaken it was designed for the Dps 2000 . i use both boards every day with out fail , planing to buy two more and try them out on the 2500 psu and 1600 psu sense one or two have . this PSU http://www.ebay.com/itm/IBM-BladeCenter-Delta-2000W-Power-Supply-Unit-DPS-2000BB-39Y7360-39Y7359-/381564359940
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 29, 2016, 07:38:01 PM Last edit: June 29, 2016, 08:02:48 PM by philipma1957 |
|
I was one of the first to buy this i bought three and sold one , i use two right now with the DPS 2000 watt PSU have been for months with out issue if im not mistaken it was designed for the Dps 2000 . i use both boards every day with out fail , planing to buy two more and try them out on the 2500 psu and 1600 psu sense one or two have .
So far I am very happy with these boards. a kit for 1 s-9 is about 85 usd plus shipping. that is 1 board and 10 pcie wires Now as a suggestion to the op I am in NJ you are west coast. you sent the second set to me in a small flat rate box. The post office would allow you to order small flat rate boxes for free and flat rate padded envelope for free. Stick the gear in the small flat rate box and stick the box in the flat rate padded envelope these are screen shots of usps order page for small flat rate and padded envelope https://i.imgur.com/QIk8MCg.png https://i.imgur.com/jRYeJP6.pngorder 6 packs of 15 envelopes = 90 they will come in a case like this your cost is about the same and the gear gets a little more protection. And a touch more room if the order is for a board and 12 cables vs 10 cables. Since the board can fit 12 cables some may want to use all 12 cables. I am clocking my s-9 and it is still in China not sure if I get it on fri the first or not.
|
|
|
|
Optimizer (OP)
|
|
June 30, 2016, 03:20:09 AM |
|
@Philipma
Sick. Thank you. I will try that.
Do you happen to have some efficiency numbers from your system??
-Optim
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 30, 2016, 03:41:34 AM |
|
@Philipma
Sick. Thank you. I will try that.
Do you happen to have some efficiency numbers from your system??
-Optim
I have 1 pdu running a dps-1600 with your board and with a evga 1600 p2. I get readings of 10 and 11 amps for two s-9s So 2400 watts to 2600 watts for 24 to 26 th. I have a second pdu so if I run 1 s-9 the pdu meter will be 5.0 to 5.5 amps as my pdu has only 2 digits. I will be able to do this on Friday or Monday
|
|
|
|
Optimizer (OP)
|
|
June 30, 2016, 08:03:40 PM |
|
@Philipma
Sick. Thank you. I will try that.
Do you happen to have some efficiency numbers from your system??
-Optim
I have 1 pdu running a dps-1600 with your board and with a evga 1600 p2. I get readings of 10 and 11 amps for two s-9s So 2400 watts to 2600 watts for 24 to 26 th. I have a second pdu so if I run 1 s-9 the pdu meter will be 5.0 to 5.5 amps as my pdu has only 2 digits. I will be able to do this on Friday or Monday That's not quite as accurate as I was hoping. Are you able to measure the voltage at the supply output and at the end of the PCI-e cables? That is the real number I am looking for tbh, though running a power analyzer as described above is the best system level information. I am only measuring with DC load at the moment and I find it is much less reliable. -Optim
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
June 30, 2016, 08:57:53 PM |
|
@Philipma
Sick. Thank you. I will try that.
Do you happen to have some efficiency numbers from your system??
-Optim
I have 1 pdu running a dps-1600 with your board and with a evga 1600 p2. I get readings of 10 and 11 amps for two s-9s So 2400 watts to 2600 watts for 24 to 26 th. I have a second pdu so if I run 1 s-9 the pdu meter will be 5.0 to 5.5 amps as my pdu has only 2 digits. I will be able to do this on Friday or Monday That's not quite as accurate as I was hoping. Are you able to measure the voltage at the supply output and at the end of the PCI-e cables? That is the real number I am looking for tbh, though running a power analyzer as described above is the best system level information. I am only measuring with DC load at the moment and I find it is much less reliable. -Optim i can give 5 numbers ac amps at the pdu and with only one s-9 it will be 4.8 to 5.3 amps ac volts at the pdu it will be about 239 to 241 dc volts at your board with full draw dc volts at your board on an unused jack dc volts at the s-9 end
|
|
|
|
Optimizer (OP)
|
|
July 01, 2016, 05:34:20 PM |
|
@Philipma
Sick. Thank you. I will try that.
Do you happen to have some efficiency numbers from your system??
-Optim
I have 1 pdu running a dps-1600 with your board and with a evga 1600 p2. I get readings of 10 and 11 amps for two s-9s So 2400 watts to 2600 watts for 24 to 26 th. I have a second pdu so if I run 1 s-9 the pdu meter will be 5.0 to 5.5 amps as my pdu has only 2 digits. I will be able to do this on Friday or Monday That's not quite as accurate as I was hoping. Are you able to measure the voltage at the supply output and at the end of the PCI-e cables? That is the real number I am looking for tbh, though running a power analyzer as described above is the best system level information. I am only measuring with DC load at the moment and I find it is much less reliable. -Optim i can give 5 numbers ac amps at the pdu and with only one s-9 it will be 4.8 to 5.3 amps ac volts at the pdu it will be about 239 to 241 dc volts at your board with full draw dc volts at your board on an unused jack dc volts at the s-9 end Awesome. Those will work. Would it also be possible to monitor the PSU output voltage under load? The easiest place to check is on the large connector on the breakout board. I just talked to my contract manufacturer. He was able to salvage a single 4k board right now. I should get it in my hands next week and run some tests. I don't suppose you want to try this one out in the coming weeks? -Optim
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4116
Merit: 7863
'The right to privacy matters'
|
|
July 01, 2016, 06:52:27 PM Last edit: July 01, 2016, 08:27:55 PM by philipma1957 |
|
@Philipma
Sick. Thank you. I will try that.
Do you happen to have some efficiency numbers from your system??
-Optim
I have 1 pdu running a dps-1600 with your board and with a evga 1600 p2. I get readings of 10 and 11 amps for two s-9s So 2400 watts to 2600 watts for 24 to 26 th. I have a second pdu so if I run 1 s-9 the pdu meter will be 5.0 to 5.5 amps as my pdu has only 2 digits. I will be able to do this on Friday or Monday That's not quite as accurate as I was hoping. Are you able to measure the voltage at the supply output and at the end of the PCI-e cables? That is the real number I am looking for tbh, though running a power analyzer as described above is the best system level information. I am only measuring with DC load at the moment and I find it is much less reliable. -Optim i can give 5 numbers ac amps at the pdu and with only one s-9 it will be 4.8 to 5.3 amps ac volts at the pdu it will be about 239 to 241 dc volts at your board with full draw dc volts at your board on an unused jack dc volts at the s-9 end Awesome. Those will work. Would it also be possible to monitor the PSU output voltage under load? The easiest place to check is on the large connector on the breakout board. I just talked to my contract manufacturer. He was able to salvage a single 4k board right now. I should get it in my hands next week and run some tests. I don't suppose you want to try this one out in the coming weeks? -Optim yes I would like to as i can test it against the other 4k board I have. I am still waiting on the ups truck but should get the batch 4 s-9 sometime today. gear is here and it is mining at ck.solo pool if you look the freq is at 525 the fans are also set low 50% why did I do this .. I am testing in my house which means my wife will kill me if I set fans to 90% and try to see if I can go to 625 over clock. I did run at freq 600 and was close to 12,800gh As you see the down clock makes 0 errors and fans set at 50% are okay not nasty bearable more to come. on volts etc. pdu reads 5.1 amps same pdu 2 car garage is filled with new gear for my kitchen remodel pdu is under the british shoe box spinning to my left i face the back door and the s-9 is vent out the back door and is hidden by a moving blanket. the sound is okay in my den next to the garage.
|
|
|
|
|