Major comparison thread on 240v breakout boards for dps-2000 psus + pcie cables

[philipma1957]

Okay  I have decided to run a comparison thread.  On 240 volt psu setups for the s-9 and s-7.

I have pcie cables from   ITOP  a set of 10 https://bitcointalk.org/index.php?topic=1535584.msg15451043#msg15451043

I have pcie cables from  Optimizer a set of 10  https://bitcointalk.org/index.php?topic=1373092.0

I have pcie cables from  Finksy/J4bberwock   https://bitcointalk.org/index.php?topic=1308296


I have 3 different server  psu's


http://www.ebay.com/itm/IBM-BladeCenter-1800W-Power-Supply-P-N-74P4400-FRU-P-N-74P4401-DPS-1600BB-/262530302169? one unit

http://www.ebay.com/itm/Delta-Electronics-DPS-2000BB-IBM-74P4452-2000W-Power-Supply-/262375058818?   three units

http://www.ebay.com/itm/IBM-39Y7400-2320W-Watt-Power-Supply-for-BladeCenter-Delta-DPS-2500BB-39Y7405-/400977770416? five units

1 atx psu
https://www.amazon.com/EVGA-SuperNOVA-PLATINUM-Crossfire-220-P2-1600-X1/dp/B00NJG61JQ/ref=sr_1_1?


I have 2   2kwatt  boards from optimizer  one I purchased one was a demo

I have 1   2kwatt boards from finksy/j4bberwock note this was sold to me by Holyscott    http://holybitcoin.com/

I have 1  4kwatt board from    finksy/j4bberwock this was a demo from them

Note I will be deleting posts here  as I want to do a cleaner thread.

[judypug1956]

This is a spacer thread for info and neatness only.

[philipma1957]

First off I want to thank ITOP for pcie-cable sample.

They are 3 foot long look like 16 gauge and have a nice wrap.




they are on j4bberwock 2k board and a dps-2000 psu


test bed solar array and s-7 has which evga 1600 t2 hooked up in photo

[Finksy]

What kinds of tests are you planning on running the PSU's/boards through?  I would check voltage drop on those ITOP 36" cables.  If they're anything like the run-of-the-mill aluminum-wired 16awg PCIe cables that are available everywhere, the voltage drop over the length of the cable may be significant, especially under load/heat. 24" was the practical limit from testing the same 16awg wires that I carry.

[philipma1957]

What kinds of tests are you planning on running the PSU's/boards through?  I would check voltage drop on those ITOP 36" cables.  If they're anything like the run-of-the-mill aluminum-wired 16awg PCIe cables that are available everywhere, the voltage drop over the length of the cable may be significant, especially under load/heat. 24" was the practical limit from testing the same 16awg wires that I carry.

I am waiting for some splitters and meters to arrive .

I should be able to test with consistent methods for all cables .

I have about six sets to test.

Goal will be a test on four boards for every cable

Also an s-7 and an s-9

And three psus

 dps- 1600
Dps - 2000
Dps - 2500

Almost every long term miner has had a melting issue or worse.

Many reasons why.

I want to find some lower limits for the cautious .

Ie 1 s-9 for 1 2k board
2 s-9 for 1 4k board.

Summer is here it hot so my space is pushing 100 f with the heat.

I also have an infra red gun will try to catch heat issues.

Lastly Avalon may have a new model
They used a two jack board  that could pull close to 600 Dc .

Which is over pcie specs for two jacks. This was the Avalon 6.
I did not melt your cables or your jacks but they darkened .
I was over clocking in a very hot garage.
I was lucky to not have a meltdown.

I consider the Avalon 6 to be a bit unsafe with your boards.
The 4k and the 2880
I think your boards are safe with the s-7 and s-9

Most of the time but as you know you had a melt down.
Looks like it was the s-7 boards not,your gear at fault .

So far itop's cables are driving an s-7 using your 2k board at freq 600 in a spot that reaches 100.0 f
No issues  with the s-7 the cables or the 2k board.

[TheYankeesWin!]

Do you have any photos or results?

[philipma1957]

This is first test it is of itop 3 foot pcie cables
I give them a good score for running s-7 at a stock clock. about 145 watt power draw on the cable
I need more time to evaluate them.

I need a bit more time to figure the testing jig out better.


Just got in 4 psu's



these will do 91%  and 2300 watts but at 1 per s-9
they will not be stressed at all



so this test is with an s-7
a j4bberwock break out board

itop pcie cables



3m foam tape really helps with this psu and breakout board


test meter shows 12.1 volts at the s-7
the other end is 12.2 volts  so the 3 foot cables under load drop less the 1% that is good


too much glare in this photo have to do it again



this side shows an evga 1600 p2 with its own cables running an s-9 a lot of voltage sag.
but most is in the psu as it is 11.7 volts out of the psu and 11.5 volts into the s-9

[Finksy]

You don't have a more accurate voltmeter by chance do you? I am very surprised at how much voltage drop is on that EVGA 1600 PSU (on top of the low output voltage).  Maybe the S9 has too much draw for it, what does it put out with no load or less load?

[philipma1957]

You don't have a more accurate voltmeter by chance do you? I am very surprised at how much voltage drop is on that EVGA 1600 PSU (on top of the low output voltage).  Maybe the S9 has too much draw for it, what does it put out with no load or less load?

That plat does 11.95 with no load.

[TheYankeesWin!]

Hey Phil  what is the story with this thread?   Did you ever finish any tests?

[philipma1957]

Yes I was busy with ETH coin  and other stuff will post results tonight.

Itop 3 foot pcie cables worked great.  test was over a 21-22 day time period.

They ran 24/7 for the 22 days.   volts on the breakout board were 12.2  volts on the s7 were 12.1

these stayed that way the entire time.

the psu was a dps-2500 flawless
the breakout board was an optimizer board 2k style flawless
the cables were jacketed and worked just fine.

This unit pulled 1250 watts on 10 cables.

The temps in the room reached as high as 102f ambient .

Other gear that ran with no issues were:
j4bberwock cables
optimizer cables
three optimizer 2k break out boards
one j4bberwock 2k board
one j4bberwock 4k board

two dps 2000 psus fine
one dps 1600 psu  fine
one dps 2500 psu fine
one evga 1600 plat 2  ran but volts dropped between 11.5 and 11.3
one dps 2000 psu runs fine
so six antminers  five s9's  and one s7 all run fine


bad gear :

one dps 2000 smelled burnt when it arrived from amazon it overheats and powers off



my conclusions are this  if the gear pulls under 200 watts a pcie all the gear is good and in every case above the pcie cables were pulling under 180 watts

So optimizer cables
    j4bberwock/finksy cables
    itop cables

are all good to run an s7 or s9

optimizer 2k breakout board is better built then

j4bberwock 2k breakout board.

if you are running an s7 or s9 pulling 150 -180 watts you can kind of take your pick. 

But for me since the solar array does get hot in the summer I will pick the optimizer 2k breakout board over the j4bberwock 2k board.  It costs more but is beefier and in a 100f room is most likely safer..

Lastly the j4bberwock 4k board I have is old been working since last fall and was used to run avalon 6's before it was used in this test.  It has stood up so far  so I can't say it is bad gear.  I will continue to use it running two s9's using dps 2000 psus.


as for the psus.


the evga 1600 p2 is disappointing it is doing 11.4 volts under load not good enoguh in my book  so I do not think I would use it to run an s9 or s7.

The dps 1600 >>>>>> Good
The dps 2000 >>>>>> Better
The dps 2500 >>>>>> Best

of the three I tested

[ITOP]

Thank you for your testing ,
Knowing you are working at ETH ,we have a the latest version of PCIE 1x to 16x riser cable, that makes much more stable and make graphic card longer life.
here is the link for your reference .
http://www.ebay.com/itm/272456808280?ssPageName=STRK:MESELX:IT&_trksid=p3984.m1558.l2649

[philipma1957]

Thank you for your testing ,
Knowing you are working at ETH ,we have a the latest version of PCIE 1x to 16x riser cable, that makes much more stable and make graphic card longer life.
here is the link for your reference .
http://www.ebay.com/itm/272456808280?ssPageName=STRK:MESELX:IT&_trksid=p3984.m1558.l2649

I have yet to test your pcie adapter, but  I will keep link here. As your pcie cables are still working fine for me.

If you want to send one  of these to me I will send a pm with my address.

To all his price is a set of 6 for 52 with free shipping it is a competitive price.  If he sends me one I will do some photos and compare it to the ones I got on amazon.

With an eBay purchase you are protected if you pay with paypal.

[valkir]

Thank you for your testing ,
Knowing you are working at ETH ,we have a the latest version of PCIE 1x to 16x riser cable, that makes much more stable and make graphic card longer life.
here is the link for your reference .
http://www.ebay.com/itm/272456808280?ssPageName=STRK:MESELX:IT&_trksid=p3984.m1558.l2649

I just ordered these. Will let you know the status of it. 

[MarkAz]

Thank you for your testing ,
Knowing you are working at ETH ,we have a the latest version of PCIE 1x to 16x riser cable, that makes much more stable and make graphic card longer life.
here is the link for your reference .
http://www.ebay.com/itm/272456808280?ssPageName=STRK:MESELX:IT&_trksid=p3984.m1558.l2649

I just ordered these. Will let you know the status of it.  

I've had some for a month or two now, and they run really well - just don't using the included SATA power cable, which is just as bad as the original.  No issues with powering GPU's from these - but in all fairness, I never had any problems using Sidehack's board - so really either of those solutions is ideal.

I don't see how these would be any improvement over the traditional on in terms of prolonging GPU life - the only thing this buys you really is more convenient cabling, and more available power at the PCI slot, but really they shouldn't be drawing that much power from them anyway.

[Optimizer]

Phil,

I updated some of my efficiency numbers.  Check the thread!

https://bitcointalk.org/index.php?topic=1667131.20

I came across a 6-digit meter.  So far it puts me around 1/3% of power loss around 1900W.  Ambient temperature is always good!

Cheers,
-Optim