the delta 2400 watts from parallel miner were really good but they seem to be gone
the ibm 2880 watts and the ibm 2980 watt are good but loud
be careful they have exposed 240 volt on the breakout board almost killed myself burnt my hand from the shock . I would cover all of the exposed leads on them
Topic: S9 Power Variances (Read 276 times)
Additionally if you want to save some more (or you can't find breakout boards somehow) money you can buy a server PSU, find the pinouts, and then wire up the connectors yourself by soldering them on.
There are 2kw+ server PSUs, I remember J4bberwock (buckrogers?) or some other member sold boards, the PSU itself and whole packages of the PSU (was some sort of IBM PSU, pretty loud IIRC). You'll probably find it if you do a bit of digging here.
Searched "J4bberwock IBM" on google, and results show it's a 2880W IBM PSU.
There are 2kw+ server PSUs, I remember J4bberwock (buckrogers?) or some other member sold boards, the PSU itself and whole packages of the PSU (was some sort of IBM PSU, pretty loud IIRC). You'll probably find it if you do a bit of digging here.
Searched "J4bberwock IBM" on google, and results show it's a 2880W IBM PSU.
Oh this is board only...Silly me... Yea i found the PSUs around 199$ but its only 750-1400W, there arent 2k+ W 
But you can run multiples (just read the warnings and set it up right when doing so), $199 is too high for a server PSU. I'd search around eBay for used 750-1100W.
Here's an 1100w for $40.00, probably can find better with some more throughout searching.
https://goo.gl/FuKT6z
Oh this is board only...Silly me... Yea i found the PSUs around 199$ but its only 750-1400W, there arent 2k+ W
Uhmmm do i see those numbers really good? 2400W for 55$? That is really cheap... for price of 1 Bitmains + shipment, i can buy 3 of this.... Thanks for this! TIL!
You'll still need a server PSU
But yeah those can be had for pretty cheap, they are rated to run 24/7, and are usually at least Gold rated.
Uhmmm do i see those numbers really good? 2400W for 55$? That is really cheap... for price of 1 Bitmains + shipment, i can buy 3 of this.... Thanks for this! TIL!
Damn those are some scary numbers! :O Thats almost 40% of them are dead. I know that you shouldn't be cheap in therms of PSU, but which PSU is in same/similar price range(but better quality(not sure am i being cheap here)) then Bitmains? Some 1800W PSUs are around 400$ in my country which is rather...x3-x4 times more expensive then Bitmains...
A lot of folks are using server PSUs with breakout boards from https://www.parallelminer.com/product/x11-breakout-board-adapter/ or eBay. My preference would be Parallelminer's version.
Damn those are some scary numbers! :O Thats almost 40% of them are dead. I know that you shouldn't be cheap in therms of PSU, but which PSU is in same/similar price range(but better quality(not sure am i being cheap here)) then Bitmains? Some 1800W PSUs are around 400$ in my country which is rather...x3-x4 times more expensive then Bitmains...
Noted.... In the future don't use Bitmain PSU's.... I knew they were suspect, but this is really bad performance. If it's one or two units then not really a big deal, but when spread across an 800 unit deployment it definitely has an impact on our engineering for this particular deployment.
My last hosting customer with bitmain PSUS purchased 3704 of them. Out of that order there were over 1200 that were either DOA or having other stability issues.
I warn anyone against buying their PSUs.
I personally have not seen the bitmain PSUs pull more power than others like you are stating but I dont doubt it one bit.
Just completed that test.... When i hook an S9 up to an IBM Blade center 2880 it runs just under 1400 W. We hook it back up to the Bitmain and it shoots right back up to 1500ish Watts...... Garbage.
Noted.... In the future don't use Bitmain PSU's.... I knew they were suspect, but this is really bad performance. If it's one or two units then not really a big deal, but when spread across an 800 unit deployment it definitely has an impact on our engineering for this particular deployment.
I wonder if it is the PSU or the miner? Maybe swap a PSU from a 1.4 over to a 1.5 and see what happens?
Noted.... In the future don't use Bitmain PSU's.... I knew they were suspect, but this is really bad performance. If it's one or two units then not really a big deal, but when spread across an 800 unit deployment it definitely has an impact on our engineering for this particular deployment.
13.5T - rated @ 1323w +10% which is 1455w. I see mine within 25 watts of that so I would still consider it in spec.
Is that not within the +/- 10%?
Actually on edit the OPs #'s ARE outside the 10% for a 13.5. But I would chalk it up to prod variances, yes it's disappointing for sure.
Actually on edit the OPs #'s ARE outside the 10% for a 13.5. But I would chalk it up to prod variances, yes it's disappointing for sure.
I have multiple batches of 13.5th S9s. Some pull around 1380w and some pull around 1480w. I dont use those crap ass bitmain power supplies so I know it is not a PSU issue.
Guys,
Anyone else noticing that the Bitmain S9 13.5Th/s from the latest batch is not really pulling 1.4 kW. It seems that there are HUGE variances in manufacturing quality of the Bitmain PSU's.... We have a few units pulling what they should (1.4-1.44 kW) but there are a ton of units we deployed pulling 1.5-1.55 kW. Anyone else seeing this??
-AH
Anyone else noticing that the Bitmain S9 13.5Th/s from the latest batch is not really pulling 1.4 kW. It seems that there are HUGE variances in manufacturing quality of the Bitmain PSU's.... We have a few units pulling what they should (1.4-1.44 kW) but there are a ton of units we deployed pulling 1.5-1.55 kW. Anyone else seeing this??
-AH
Jump to: