Update Large review of Finksy/J4bbrwock server psus compared to ATXpsu-photos up

[nhando]

I couldn't believe how strong these Scotch Tape is!!  Once the PSU were taped together, I couldn't pry them open with my hands................but maybe I'm weak.  =)   You can see that just a tiny piece of the tape held the fans together so well.  Mr Phillip definitely have a lot of creative solutions.  I love this tape and the 2K PSU.  Also the Ultra KAZE fans Phillip recommend is also very good.  Super Quiet.  Just ordered the 4K board with voltage meter.  Can't wait to make my Avalon earn it's worth as it's gimping along at 3.2-3.3TH right now.  

Phillip - Waiting on your review for the 2980W PSU.  Want to see if it's worth the price premium for Platinum efficiency.   Rough math shows it's much more expensive than the 4K bundle but maybe over long term it will pay off with the efficiency?  The review will shed more light on it.  For now I will stick with the 4K bundle until I see better ROI on the 2980.

[dmwardjr]

My pdu has 1/10 or nearest tenth. It certainly will show close enough to see a large difference between the 2880w and the 2980w.

What we all really need to know is 100 watts a s-7 or 100 watts an Avalon really the difference between the two models.

I appreciate you doing this.  By the way, what PDU do you have?  I'm sure you've mentioned it elsewhere in another thread.  I need to get another PDU very soon before my power upgrade.  I wouldn't mind having a 240V PDU that gives decimal point reading as well.

Relax  everyone  I ordered this

 I had a coupon and paid about 140 not 150

http://www.amazon.com/gp/product/B00TMPKE96?psc=1&redirect=true&ref_=od_aui_detailpages00

should be a direct fit and easy to compare to this

http://www.amazon.com/gp/product/B00VMEB944?psc=1&redirect=true&ref_=ox_sc_act_title_1&smid=A25R0MJJKVE76I


this is 76 with shipping.


so say 77 vs 150  or 75 dollars difference .

    I will run with identical testing.   and buyers can tell the watt difference   both should work fine with the breakout board.

The only thing I am sure about today is my titanium 1600 watt evga gets better watts rating then the 2880watt psu.  at freq 600

I purchased a 2nd one for the hell of it.  One from NSR on eBay and one from NSR on Amazon at $150 each.  Even at 4 or 5 percent difference between the two models I believe it's worth it in my case.  My plans are to have 120 to 220 x S7's [or equivalent] by this time next year.  Which means I'm going to need much more than the 17 x 2880's I have at present.  I'm kind of laughing at myself now for buying the 8 x Corsair AX860's when I first started back in September, 2014.  It definitely pays to do your homework before taking the plunge.

[nhando]

Don't laugh as we all make the same rookie mistakes.  I have Dell 750W, HP 1000W, 2 EVGA G2 1300W and IBM 2K PSUs.  haha.  Now I'm debating if I should get the 2980....................................................

[dmwardjr]

APC7911 has decimal points in the web interface, but not on the PDU its self.

If you want a good/accurate meter, look into the Watts up? meter, the UO version of the .net version is ideal for this kind of testing. Elsewise, I think you'll be spending thousands for something better, like a fluke power meter/analyzer.

Wow, thanks for that info.  It's nice to know my investment has more features than previously known.  I mainly got them for switching capability.

However, I would not mind have a new toy (meter) around the house to check the efficiency of new HW.  Thanks for the info on meters.  I'm researching them and pricing them now.

[dmwardjr]

Don't laugh as we all make the same rookie mistakes.  I have Dell 750W, HP 1000W, 2 EVGA G2 1300W and IBM 2K PSUs.  haha.  Now I'm debating if I should get the 2980....................................................

LOL...  I don't feel as bad now.

I must say, the new setup for the 2 x 2000 BB PSU's with voltage adjustment and digital volt meter is a nice set up as well.  Yes, the ability to choose fans to make it a quieter package is a plus [If noise is one's concern].  I say it all depends on what one plans to do and where they plan to do it.  If one wants less noise and the ability to over clock substantially, the 2 x 2000 BB setup with voltage adjustment is the way to go.  It would be my choice over the 2880 if noise was an issue and if I was not concerned about vertical headroom on my rack for more shelves.

My only concern with the voltage adjustment to over clock is "power efficiency."  Will the gains made from more bitcoin mined via over clocking with that setup trump the gains from a more power efficient PSU without overclocking?  The plus side from not over clocking is the potential for increased longevity of the blades in our rigs.  At least that's my opinion.  I'm open minded to any thing else someone would like to add.  I'm all about making the right decision.  I just feel this is important if one plans to have a lot of rigs powered by a lot of PSU's.  Making the correct decision on the right PSU for your long term plans is extremely important.

[dmwardjr]

I'll break this up to make it easier to follow. I'll leave the discussion alone after this post...

Hi Finksy,

Thank you very much for your constructive criticism  Your frankness is much appreciated.  I also appreciate your willingness to contribute HW in Prelude's efforts to determine power efficiency difference between PSU's.  You made very good points about what would probably be the proper way to test PSU efficiency for bitcoin mining applications.  Consistency in HOW the test is performed is extremely important... Making measurements with same meter at same location with same rigs over same period of time.

I hope I did not open up a can of worms.  That was not my intention at all.  I've made too many mistakes in the beginning of my bitcoin mining venture.  I've learned [The hard way] how important it is to make an educated decision about hardware used for mining.  I don't want to continue those mistakes anymore.  I'm hoping this venture will last for many years to come.  So, I'm wanting to make the wisest and most economical decision I can make while growing into the future.

All of the time and energy invested with engineering and design with breakout boards for various PSU's and the commitment to produce these on a relatively large scale for many people by you and J4bberwock is appreciated greatly, Sir!  Thank you both for all you have done and continue to do for the community!

[J4bberwock]

I'll break this up to make it easier to follow. I'll leave the discussion alone after this post...

Hi Finksy,

Thank you very much for your constructive criticism  Your frankness is much appreciated.  I also appreciate your willingness to contribute HW in Prelude's efforts to determine power efficiency difference between PSU's.  You made very good points about what would probably be the proper way to test PSU efficiency for bitcoin mining applications.  Consistency in HOW the test is performed is extremely important... Making measurements with same meter at same location with same rigs over same period of time.

I hope I did not open up a can of worms.  That was not my intention at all.  I've made too many mistakes in the beginning of my bitcoin mining venture.  I've learned [The hard way] how important it is to make an educated decision about hardware used for mining.  I don't want to continue those mistakes anymore.  I'm hoping this venture will last for many years to come.  So, I'm wanting to make the wisest and most economical decision I can make while growing into the future.

All of the time and energy invested with engineering and design with breakout boards for various PSU's and the commitment to produce these on a relatively large scale for many people by you and J4bberwock is appreciated greatly, Sir!  Thank you both for all you have done and continue to do for the community!

As far as I can tell, the easiest way we'll have to test efficiency will be to compare against a known efficiency PSU with either of the following options:
-Same voltage for both PSUs if the miner is using a string design
-Close voltage for both PSUs if the miner is using a buck converter (The S7 B8 and other with 135chips will work since it is an hybrid with the string fed by a buck converter similar to the design I tried for SFARDS chips)

Ideally, DPS2000 BB vs DPS2500BB, and 2880 vs 2980 with the same miners as load so we can cancel the effect of the built in fans.
I'll test the 2500 vs 2000 next week once I'm back home.

I can estimate, 87% for the 2880 including the fans will be between 90% and 92% without them depending on the load, so not too far away from the platinum efficiency measured on the 2980.
The 4kw may have a better global efficiency including the fans since we are using ultrakaze or similar fans, and only 2 of them for 4kw compared to 3x turbojet fans on the 2880/2980w

[philipma1957]

I GOT shipping info for the 2980 watt should come on tues or weds.

I plan to test the :

2x 2000

2880watt

2980watt

with a pair of avalon sixes  at 12.20 volts

then with a single avalon six at 12.20 volts


My s-7 is not available for testing my buddy that has an entire office floor rented got a new sublease deal..

This means he leases 80% and keep 20% for the floor.  He also pays 20% of the power  so the 13.3 cent full price power is only 2.66 cents a kwatt.

The limits are 1x 15 amp circuit  and keep the gear quiet. 

 So I hooked up 1 s-7 at freq 500  we are getting 4000th at 1000 watts with fan at 32%.    Sound is okay power use is okay.

 I ran s-3's in his place for 14 months I am hoping to run the s-7 past the ½ ing due to the 2.66 cent price.

Back to this testing.

I hope to get some good numbers and determine the difference between the 3 psu setups once and for all at about the 2000-2200 watt load

and the 1000- 1100 load

[VirosaGITS]

I GOT shipping info for the 2980 watt should come on tues or weds.

I plan to test the :

2x 2000

2880watt

2980watt

with a pair of avalon sixes  at 12.20 volts

then with a single avalon six at 12.20 volts


My s-7 is not available for testing my buddy that has an entire office floor rented got a new sublease deal..

This means he leases 80% and keep 20% for the floor.  He also pays 20% of the power  so the 13.3 cent full price power is only 2.66 cents a kwatt.

The limits are 1x 15 amp circuit  and keep the gear quiet. 

 So I hooked up 1 s-7 at freq 500  we are getting 4000th at 1000 watts with fan at 32%.    Sound is okay power use is okay.

 I ran s-3's in his place for 14 months I am hoping to run the s-7 past the ½ ing due to the 2.66 cent price.

Back to this testing.

I hope to get some good numbers and determine the difference between the 3 psu setups once and for all at about the 2000-2200 watt load

and the 1000- 1100 load

What batch miner? If its B8 i could pitch in the B8 data i get, running with ATX with known efficiency range at various load it could give a good benchmark to compare to. Though i'd be on 120v, the difference with 220-240v efficiency should be 2%.

[philipma1957]

I GOT shipping info for the 2980 watt should come on tues or weds.

I plan to test the :

2x 2000

2880watt

2980watt

with a pair of avalon sixes  at 12.20 volts

then with a single avalon six at 12.20 volts


My s-7 is not available for testing my buddy that has an entire office floor rented got a new sublease deal..

This means he leases 80% and keep 20% for the floor.  He also pays 20% of the power  so the 13.3 cent full price power is only 2.66 cents a kwatt.

The limits are 1x 15 amp circuit  and keep the gear quiet.  

 So I hooked up 1 s-7 at freq 500  we are getting 4000th at 1000 watts with fan at 32%.    Sound is okay power use is okay.

 I ran s-3's in his place for 14 months I am hoping to run the s-7 past the ½ ing due to the 2.66 cent price.

Back to this testing.

I hope to get some good numbers and determine the difference between the 3 psu setups once and for all at about the 2000-2200 watt load

and the 1000- 1100 load

What batch miner? If its B8 i could pitch in the B8 data i get, running with ATX with known efficiency range at various load it could give a good benchmark to compare to. Though i'd be on 120v, the difference with 220-240v efficiency should be 2%.

I had a batch 2 and it is now at my buddies office.

My testing on the psus with Avalon sixes will still work since the
Loads placed on the 2880, 2980, 4000 dual 2000 will be indentical.

One Avalon six and two Avalon sixes.  Volts will be as close as possible to 12.2 volts.

[dmwardjr]

I had a batch 2 and it is now at my buddies office.

My testing on the psus with Avalon sixes will still work since the
Loads placed on the 2880, 2980, 4000 dual 2000 will be indentical.

One Avalon six and two Avalon sixes.  Volts will be as close as possible to 12.2 volts.

SweeeeeeeeeT   

Thank you very much for taking the time to do this, Phillip.

Were you also planning to get the new S7 batch 8 or 9?  I'm curious if it's efficiency would be the same as batch 6 if the batch 8 or 9 were under clocked from 700 MHz to 600 MHz?  It may be I would not have to be concerned with under clocking if the hash rate difference of 700 GH/s pays for the extra 100 watts consumed from the higher clock of 700 MHz in batch 8 & 9.  If the difference of the 700 GH/s more than pays for the extra power consumed to make extra profit, it is worth it.  If they simply offset, it would be better to under clock to try to increase the longevity of the rig.

[philipma1957]

I had a batch 2 and it is now at my buddies office.

My testing on the psus with Avalon sixes will still work since the
Loads placed on the 2880, 2980, 4000 dual 2000 will be indentical.

One Avalon six and two Avalon sixes.  Volts will be as close as possible to 12.2 volts.

SweeeeeeeeeT   

Thank you very much for taking the time to do this, Phillip.

Were you also planning to get the new S7 batch 8 or 9?  I'm curious if it's efficiency would be the same as batch 6 if the batch 8 or 9 were under clocked from 700 MHz to 600 MHz?  It may be I would not have to be concerned with under clocking if the hash rate difference of 700 GH/s pays for the extra 100 watts consumed from the higher clock of 700 MHz in batch 8 & 9.  If the difference of the 700 GH/s more than pays for the extra power consumed to make extra profit, it is worth it.  If they simply offset, it would be better to under clock to try to increase the longevity of the rig.

I got 4 batch 8 but they were purchased for a friend.

 I am a bit annoyed with bitmaintech  I will try to make no purchase from them for myself as  I feel they owe me about ½ a btc in compensation.  I got a poor batch 1 that only hashed around 4500gh and they were late on a batch two shipment.  So I will  not buy from them for personal ownership  until they send me some type of compensation.
 They just blow me off at this point when I mention  this which I have mentioned to them in more then 6 or 7 emails.
So if someone wants ask me to help them order an s-7 I do. But I do not order for myself anymore.

 I ordered 9 s-7's direct 2 for me and 7 for others. I ordered an hashnest s-7 and I ordered a lee group s-7.
So that is 11 s-7's  at this moment I have 1 s-7 at my buddies office and 1 s-7 mining in China with lee's group
I have 2 avalon sixes in house and most likely I will order at least one more avalon 6.
Since I have no hope of bitmaintech giving me coupons or compensation any that I order from them will be for friends not myself.
Sad since I have purchased more then 40k in gear from them.

I have to say one thing the batch 2 rated at freq 575 has been the best miner I ever purchased from them  it will go as high as freq 637 and 5150 gh. 

I have it at my friends office down clocked to freq 500 and 4000gh (need to keep it quiet) Since I pay 2.66 cents a kwatt that one will make money.

[dmwardjr]

I had a batch 2 and it is now at my buddies office.

My testing on the psus with Avalon sixes will still work since the
Loads placed on the 2880, 2980, 4000 dual 2000 will be indentical.

One Avalon six and two Avalon sixes.  Volts will be as close as possible to 12.2 volts.

SweeeeeeeeeT   

Thank you very much for taking the time to do this, Phillip.

Were you also planning to get the new S7 batch 8 or 9?  I'm curious if it's efficiency would be the same as batch 6 if the batch 8 or 9 were under clocked from 700 MHz to 600 MHz?  It may be I would not have to be concerned with under clocking if the hash rate difference of 700 GH/s pays for the extra 100 watts consumed from the higher clock of 700 MHz in batch 8 & 9.  If the difference of the 700 GH/s more than pays for the extra power consumed to make extra profit, it is worth it.  If they simply offset, it would be better to under clock to try to increase the longevity of the rig.

I got 4 batch 8 but they were purchased for a friend.

 I am a bit annoyed with bitmaintech  I will try to make no purchase from them for myself as  I feel they owe me about ½ a btc in compensation.  I got a poor batch 1 that only hashed around 4500gh and they were late on a batch two shipment.  So I will  not buy from them for personal ownership  until they send me some type of compensation.
 They just blow me off at this point when I mention  this which I have mentioned to them in more then 6 or 7 emails.
So if someone wants ask me to help them order an s-7 I do. But I do not order for myself anymore.

 I ordered 9 s-7's direct 2 for me and 7 for others. I ordered an hashnest s-7 and I ordered a lee group s-7.
So that is 11 s-7's  at this moment I have 1 s-7 at my buddies office and 1 s-7 mining in China with lee's group
I have 2 avalon sixes in house and most likely I will order at least one more avalon 6.
Since I have no hope of bitmaintech giving me coupons or compensation any that I order from them will be for friends not myself.
Sad since I have purchased more then 40k in gear from them.

I have to say one thing the batch 2 rated at freq 575 has been the best miner I ever purchased from them  it will go as high as freq 637 and 5150 gh. 

I have it at my friends office down clocked to freq 500 and 4000gh (need to keep it quiet) Since I pay 2.66 cents a kwatt that one will make money.

Sorry to hear that.  I'm hoping my RMA doesn't take very long with bitmain.  I recently sent back 3 blades for replacement.  I'm hoping they get them some time next week and send the new one's back to me in the mail before next week is over.  I hate the waiting game.

By the way, how do you think the Avalon 6 will handle the heat this summer?  Did the firmware update come out yet that's supposed to do something to reduce the heat?

[philipma1957]

I had a batch 2 and it is now at my buddies office.

My testing on the psus with Avalon sixes will still work since the
Loads placed on the 2880, 2980, 4000 dual 2000 will be indentical.

One Avalon six and two Avalon sixes.  Volts will be as close as possible to 12.2 volts.

SweeeeeeeeeT   

Thank you very much for taking the time to do this, Phillip.

Were you also planning to get the new S7 batch 8 or 9?  I'm curious if it's efficiency would be the same as batch 6 if the batch 8 or 9 were under clocked from 700 MHz to 600 MHz?  It may be I would not have to be concerned with under clocking if the hash rate difference of 700 GH/s pays for the extra 100 watts consumed from the higher clock of 700 MHz in batch 8 & 9.  If the difference of the 700 GH/s more than pays for the extra power consumed to make extra profit, it is worth it.  If they simply offset, it would be better to under clock to try to increase the longevity of the rig.

I got 4 batch 8 but they were purchased for a friend.

 I am a bit annoyed with bitmaintech  I will try to make no purchase from them for myself as  I feel they owe me about ½ a btc in compensation.  I got a poor batch 1 that only hashed around 4500gh and they were late on a batch two shipment.  So I will  not buy from them for personal ownership  until they send me some type of compensation.
 They just blow me off at this point when I mention  this which I have mentioned to them in more then 6 or 7 emails.
So if someone wants ask me to help them order an s-7 I do. But I do not order for myself anymore.

 I ordered 9 s-7's direct 2 for me and 7 for others. I ordered an hashnest s-7 and I ordered a lee group s-7.
So that is 11 s-7's  at this moment I have 1 s-7 at my buddies office and 1 s-7 mining in China with lee's group
I have 2 avalon sixes in house and most likely I will order at least one more avalon 6.
Since I have no hope of bitmaintech giving me coupons or compensation any that I order from them will be for friends not myself.
Sad since I have purchased more then 40k in gear from them.

I have to say one thing the batch 2 rated at freq 575 has been the best miner I ever purchased from them  it will go as high as freq 637 and 5150 gh. 

I have it at my friends office down clocked to freq 500 and 4000gh (need to keep it quiet) Since I pay 2.66 cents a kwatt that one will make money.

Sorry to hear that.  I'm hoping my RMA doesn't take very long with bitmain.  I recently sent back 3 blades for replacement.  I'm hoping they get them some time next week and send the new one's back to me in the mail before next week is over.  I hate the waiting game.

By the way, how do you think the Avalon 6 will handle the heat this summer?  Did the firmware update come out yet that's supposed to do something to reduce the heat?

i will use the dual 4000 watt breakout board set volts at 11.80 then down clock to 3000 gh i may also need to jack up fans to 75

[kilo17]

Just wanted to quiet the 2880 down a bit and came up with a 8640 watt beast - 

But it is as quiet as a mouse with a 180cm fan pushing about 176 cfm- more quiet than a single Avalon6

[nhando]

That is so awesome Kilo!!  How high is the entire combo and where did you get that FAN?

[philipma1957]

That is so awesome Kilo!!  How high is the entire combo and where did you get that FAN?

Assuming it's the fan I think it is, it's a Sliverstone - here's the black version:

httvVCc ---------I disabled your link

They're nice fans - really quiet and has a couple speed settings which are nice.  I only wish they had a PWM version...

do not give hidden referrals to buy at amazon.

also that is not the fast one.
the one below is not a referral

and it is the fast one 2000 rpm

http://www.newegg.com/Product/Product.aspx?Item=9SIA4TZ2T50267&cm_re=silverstone_180mm_fan-_-35-220-048-_-Product


looking for a better price then newegg.


the better fan but still not cheap:

http://www.amazon.com/Silverstone-Penetrator-2000rpm-Channeling-AP182/dp/B008VQ2Y2M/ref=sr_1_1?s=electronics&ie=UTF8&qid=1451845806&sr=1-1&keywords=silverstone+sst-ap182


I will try ebay:

http://www.ebay.com/itm/Silverstone-AP182-180mmx180mx32mm-Wide-Blade-Air-Channeling-Speed-Control-Fan/271036161367?


now a costly product for the fast one   forty bucks.

but 3 sets of oem server fans burn a shit ton of watts  this fan does 3 2880watt or a mix of 2980 and 2880 watt psu's


15.6 watts to cool 3 2880 watt psu's !!!

[judypug1956]

....

the one below is not a referral

and it is the fast one 2000 rpm

http://www.newegg.com/Product/Product.aspx?Item=9SIA4TZ2T50267&cm_re=silverstone_180mm_fan-_-35-220-048-_-Product


looking for a better price then newegg.


the better fan but still not cheap:

http://www.amazon.com/Silverstone-Penetrator-2000rpm-Channeling-AP182/dp/B008VQ2Y2M/ref=sr_1_1?s=electronics&ie=UTF8&qid=1451845806&sr=1-1&keywords=silverstone+sst-ap182


I will try ebay:

http://www.ebay.com/itm/Silverstone-AP182-180mmx180mx32mm-Wide-Blade-Air-Channeling-Speed-Control-Fan/271036161367?


now a costly product for the fast one   forty bucks.

but 3 sets of oem server fans burn a shit ton of watts  this fan does 3 2880watt or a mix of 2980 and 2880 watt psu's


15.6 watts to cool 3 2880 watt psu's !!!

  I used my secondary account to quote myself  this is  just the info on the correct model fan and links.  yeah it is 40 bucks but 15.6 watts vs 150 watts  to cool 3 2880 watt psu's

[philipma1957]

I am going to buy the silverstone  ap 182 and use it to show  watts saved and sound level.


@ kilo17  you are going to cost me money


found a review on the fast model:




http://thermalbench.com/2015/06/17/silverstone-ap182-180mm-fan/

and found a better usa price. no referral link here it is.


https://www.performance-pcs.com

http://www.performance-pcs.com/silverstone-sst-ap182-air-penetrator-180mm-case-fan-w-adustable-speed-control.html 

 28 bucks and 10 to ship via usps

[kilo17]

That is so awesome Kilo!!  How high is the entire combo and where did you get that FAN?

Assuming it's the fan I think it is, it's a Sliverstone - here's the black version:

http://amzn.to/1OEvVCc

They're nice fans - really quiet and has a couple speed settings which are nice.  I only wish they had a PWM version...

Here is the fan- it is a step up from the 181

http://www.amazon.com/gp/product/B008VQ2Y2M?psc=1&redirect=true&ref_=oh_aui_detailpage_o01_s00