BITMAIN AntMiner APW3-12-1600 PSU Series

[generalt]

The first 2 I got from Bitmain have 12 PCIe connectors. Also no provision for DC on/off unless there is a header inside. Now they all come with 10 PCIe and a DC on/off pigtail  plug w/jumper on it.

Thanks for the clarification.

[NotFuzzyWarm]

np. Think my first 2 were from late batch 2 or early 3. I was luckily cautious enough to sit out the first round of them.
Since the s7's came out they are all I use. Most of my HP 1200w PSU's got sent to sidehack to run the s5's I have hosted with him and the my IBM PSU's are in use here.

Just got simpler to use something premade for the job. So far zero problems with now over a dozen of them.
I really have to update pix of the main farm. This is pre-s7 days http://phluph.imgur.com/all/

The area looks downright barren now but has more hash power yet slightly less kw used than what it took pre s7's.

[notlist3d]

The first 2 I got from Bitmain have 12 PCIe connectors. Also no provision for DC on/off unless there is a header inside. Now they all come with 10 PCIe and a DC on/off pigtail  plug w/jumper on it.

I have 3 originals from S5+ day's with 12 connectors.  It did not need more then  the the 10 but it did have it.   

Not sure what caused them to remove it.  Although they might have decided 10 is  all they will use and that is quite a bit.

[Prelude]

Can anyone verify the power factor on one of these? I'm wondering if the one I tested is defective, or if they're all crap with a PF of 0.93.

[philipma1957]

Can anyone verify the power factor on one of these? I'm wondering if the one I tested is defective, or if they're all crap with a PF of 0.93.

I always struggle with pf vs efficiency

 These are supposed to be 93% efficient platinum.

So what are they?

I know the 2880 IBM psu is gold at best
Th 2980 IBM psu is a platinum.   

Both my test on them were done with stock fans and with my fan mods

Stock screamers use 60 watts, my mods use 25 watts.

So does pf = eff?

[NotFuzzyWarm]

Can anyone verify the power factor on one of these? I'm wondering if the one I tested is defective, or if they're all crap with a PF of 0.93.

I always struggle with pf vs efficiency

 These are supposed to be 93% efficient platinum.

So what are they?

I know the 2880 IBM psu is gold at best
Th 2980 IBM psu is a platinum.  

Both my test on them were done with stock fans and with my fan mods

Stock screamers use 60 watts, my mods use 25 watts.

So does pf = eff?

No 2 different things although PF does impact efficiency.

Power Factor is how well the load compares to a purely resistive load which is a PF of 1.0. A purely resistive load has current draw exactly in-phase with the voltage.

Toss in capacitance or inductance eg. the cap inputs to a switching supply or a filter choke and current then either leads or lags behind voltage. Normally not an issue for users but when you start running heavy loads like several 10's of kw -- then the power grid starts to complain because the out of phase current draw will start to become much higher that what the average current reflects.

PF can either be passively or actively corrected in a PSU. Passive correction works at the expense of overall efficiency, active control is more efficient but costs more to implement.

EDIT: All about PF http://powerelectronics.com/power-electronics-systems/back-basics-power-factor-and-why-we-correct-it

[dannygroove]

What power cord do you recommend to use with this PSU?

[HagssFIN]

At least 3x1.0mm2 power cord with C13 connector in other end should do.

[philipma1957]

What power cord do you recommend to use with this PSU?

if you are usa based 16awg  is thick enough since it is 240 volts.

I like 14 awg

[HagssFIN]

I noticed that there is a new version of the APW3 PSU at cnshop.bitmain.com

Bitmain APW3+ -12-1600 Series
https://cnshop.bitmain.com/productDetail.htm?pid=00020160331093405522rt9ve3Ig06CB




More info via Google Translate: https://translate.google.com/translate?hl=fi&sl=zh-CN&tl=en&u=https%3A%2F%2Fcnshop.bitmain.com%2FproductDetail.htm%3Fpid%3D00020160331093405522rt9ve3Ig06CB&sandbox=1

[grimjaw]

Hi, i just got this psu today.
Initially it run for about 20 mins, then it shuts down and i am not able to get it back up. Things i have tried as below:

1) Used different power cord.
2) Used different PSU to make sure it's not the miner that faulty.

[gt_addict]

You're using it on 205v or more I assume?

[grimjaw]

yes my country standard wall socket is 220-240v.

It works initially for about 20 minutes. How do i troubleshoot this? any internal fuse or something?

[NotFuzzyWarm]

yes my country standard wall socket is 220-240v.

It works initially for about 20 minutes. How do i troubleshoot this? any internal fuse or something?

If you are not experienced with servicing electronics DO NOT OPEN IT OR ATTEMPT TO SERVICE ANY PSU! The line power side can at best give a heavy poke and at worst KILL YOU.

I assume the fan runs and it does not get warm/hot? If it simply shuts down and is not being overloaded then time to start an RMA with Bitmain.

[grimjaw]

yes my country standard wall socket is 220-240v.

It works initially for about 20 minutes. How do i troubleshoot this? any internal fuse or something?

If you are not experienced with servicing electronics DO NOT OPEN IT OR ATTEMPT TO SERVICE ANY PSU! The line power side can at best give a heavy poke and at worst KILL YOU.

I assume the fan runs and it does not get warm/hot? If it simply shuts down and is not being overloaded then time to start an RMA with Bitmain.

The fan run and the miner hashing for only 20 minutes then it stopped completely and i am not able to turn it ever since. Currently using other psu(not bitmain) to run my s7

[NotFuzzyWarm]

Again, file an RMA with bitmain. I have dozens of the supplies and many are several years old -- zero failures. Once Bitmain fixed the problems with the batch-1 supplies they have been pretty bullet proof with very few reported failures.

[MrKaizerSoze]

The high power efficiency translates to big cost savings -- if one uses a secondhand general use PSU running at 83% efficiency, when compared to the APW3-12-1600 running at 93.8% efficiency, assuming an electricity cost of $0.10 per kWh, a user could save more than $150 USD per year in power fees.

Just to be clear in how this was calculate:
In 1 hour = 1.6 kW
In 1 month (744 hours) = 1,190.4 kW

1 PSU in 1 month @ USD$0.10 per kWh
1,190.4 x 0.10 = USD$119.04

Are my assumptions correct?

[LevinSwe]

I'm searching for an pdf or som specs on the U10 chip:
8233BD
1534DFL0DU

I think it is the switchmode controller in the APW3-12-1600-B2 PSU?.. 

One of my psu died when the temperature in my mining house dropped to -10c.. is there any drawing of the board and components? The DC bus is ok, the 12v low current supply is working, mosfets and diods seems to be ok.. 

[mustangy]

i can't get the differences between those two.
one them is 5 pairs pci-e other 9 pairs
all numbers are same max output etc
5 pairs one is for s9 l3 and d3
9 pairs one is for l3 only ??
any idea prices are $105 $115
http://imgur.com/a/739mK
http://imgur.com/a/IUUkB

[HagssFIN]

i can't get the differences between those two.
one them is 5 pairs pci-e other 9 pairs
all numbers are same max output etc
5 pairs one is for s9 l3 and d3
9 pairs one is for l3 only ??
any idea prices are $105 $115

More cables, more price.

They are both the same power supply.

But you must understand that the one with 5 pairs cables (10x cables) is good to power ONE S9 or L3+ or D3.
The one with 9 pairs cables (18x cables) is good to power TWO L3+ or ONE S9 or ONE D3.