Go Back  DoItYourself.com Community Forums > Electrical, AC & DC. Electronic Equipment and Computers > Electrical - AC & DC
Reload this Page >

94% efficient Server power supply pulling 90 percent more from wall than load

94% efficient Server power supply pulling 90 percent more from wall than load

Reply

  #1  
Old 10-09-17, 12:35 PM
W
Member
Thread Starter
Join Date: Feb 2017
Location: USA
Posts: 3
Received 0 Votes on 0 Posts
Post 94% efficient Server power supply pulling 90 percent more from wall than load

It is an HP Server power supply. its suppose to be 94 percent efficient and power up to 900 watts on 120 volts.

https://www.amazon.com/gp/product/B0...?ie=UTF8&psc=1
I have only 3 gpus hooked up to it rated at 130 watts a piece. What on earth could be the issue? Could the power supply unit be malfunctioning?

I am measuring electrical draw from each gpu on my computer through a program called GPU-z. this program is accurate by all means from the feedback I have received by other users that monitor power draw.

The psu Itself stays warm and the fan always runs. From what I have read this is par for the course for this particular PSU.
I cut the gpus off and the power being pulled by the power supply alone was 80 watts approximately.

Has anyone used this type of power supply before? If so, should I add more load in order to see if it increases the power draw by a marginal amount? Give up and just try to return it?
All of this came to my attention when I noticed the extension cord to the power supply was getting warm at what I thought was just 400 watts.


Many thanks
 
  #2  
Old 10-09-17, 12:47 PM
P
Group Moderator
Join Date: Mar 2003
Location: NC, USA
Posts: 24,460
Received 752 Votes on 691 Posts
Have you measured how much current the power supply is consuming with the GPU's online?
 
  #3  
Old 10-09-17, 12:48 PM
W
Member
Thread Starter
Join Date: Feb 2017
Location: USA
Posts: 3
Received 0 Votes on 0 Posts
yes, its consuming 780 watts. According to the watt meter.

Is this what you man? I mean that is the total consumption wattage wise. I could calculate the amperage of this.
 
  #4  
Old 10-09-17, 12:56 PM
PJmax's Avatar
Group Moderator
Join Date: Oct 2012
Location: Northern NJ - USA
Posts: 60,919
Received 1,344 Votes on 1,242 Posts
It would really be most effective to use an actual DC ampmeter to measure unit draw.
 
  #5  
Old 10-09-17, 01:02 PM
W
Member
Thread Starter
Join Date: Feb 2017
Location: USA
Posts: 3
Received 0 Votes on 0 Posts
good input. I do not know how to do that on this system. I assume I could use a voltmeter. but finding a way to get access to a live area would be the challenge. I think I am just going to return the power supply.

This is such a bummer. thanks for the feedback.
 
  #6  
Old 10-09-17, 04:44 PM
B
Member
Join Date: Oct 2008
Location: New England
Posts: 10,524
Received 37 Votes on 34 Posts
@ wrightpt, you said " If so, should I add more load in order to see if it increases the power draw by a marginal amount? "
My guess is that would be worth a try. If you add an extra load and the total power only increases by 110% of that amount then the unit is supplying the power at 90% efficiency (approximately).

Bud
 
  #7  
Old 10-10-17, 05:47 AM
P
Group Moderator
Join Date: Mar 2003
Location: NC, USA
Posts: 24,460
Received 752 Votes on 691 Posts
So, your basing your judgement on power consumption stated by the computer's software?
 
  #8  
Old 10-10-17, 08:15 AM
I
Forum Topic Moderator
Join Date: Feb 2005
Location: Near Lansing, Michigan
Posts: 10,943
Received 44 Votes on 42 Posts
Do you have the efficiency vs. load curve from the manufacturer? Power supplies always operate at different levels of efficiency for different amounts of load. You're probably just not at the sweet spot.
 
 

Thread Tools
Search this Thread
 
Ask a Question
Question Title:
Description:
Your question will be posted in: