Question on actual load on PSU

  • Thread starter Deleted whining member 223597
  • Start date
D

Deleted whining member 223597

Guest
I just finished setting up my miner in my old x58 system again and my kill-a-watt is reading ~750-770w at the wall. My PSU is a one of the Strider Gold series(not the newest one, but it was the flagship model around a year ago I believe). It's rated for 750w. I thought I remember reading that what is being pulled at the wall is different from what the actual PSU is using. I just wanna make sure my PSU isn't gonna die on me.

I'm guessing it's actually using around the mid to high 600's?


EDIT: Looks like it shut off while I was posting this. Windows said there was a blue screen so not sure if that as due to the PSU being overloaded or some other funky stuff going on.
 
Last edited by a moderator:
probably 87% efficient at max load, so 750w psu would pull 862 from the wall theoretically before failure.

newer psus can usually handle peak load on the 12v rail only, so you may get more then 862.

of course, your kill a watt is not exactly a precision instrument.
 
depending on % of load and the certification(if it actually measures up to it) will say how much it is actually supplying, example
Fraction of Rated Load
20% 50% 100%
80 PLUS 80% 80% 80%
80 PLUS Bronze 82% 85% 82%
80 PLUS Silver 85% 88% 85%
80 PLUS Gold 87% 90% 87%
80 PLUS Platinum 90% 94% 91%

so if you have a gold rated and from the wall was 750-770 that would be ~650-669w supplied power.
Depending on the power supply, usually a "good" rated power supply can supply +/- ~5% max of this level of power so if its rated for 750 it can supply 715 or so, and will draw from the wall ~860w though there are "safeties" built in to prevent it from shorting out and overdrawing for to long form heat issues and the like.

In your case make sure you have it set so you can see what the blue screens were/are. Just an FYI, I had dual 7870 running on an HX750 with my Phenom II 955 all overclocked and it certainly was not happy, she was getting quite hot and I 100% guarantee the power draw on your rig is quite a bit more then mine can be considering the cpu you have in an overclocked/volted state.
 
Looks like it was just my one card having a bad overclock setting. I did under volt and down clock them a bit though, so it's reading around 700 watts now. Thanks for the info though!

This x58 system is pretty power hungry though compared to my Z77. Never buy micro-atx unless you really need it :(
 
how is x58 microATX? I thought they were all ATX and/or E-ATX? and which cpu are you using in it, yes it can be power hungry but they are 130w tdp rated chips except for one of them that works in x58 if I am right, so of course it will draw more power, they are a higher end chip, more memory channels, more cores etc etc.
 
dragonstongue said:
how is x58 microATX? I thought they were all ATX and/or E-ATX? and which cpu are you using in it, yes it can be power hungry but they are 130w tdp rated chips except for one of them that works in x58 if I am right, so of course it will draw more power, they are a higher end chip, more memory channels, more cores etc etc.
Click to expand...

There is an Asus and also an EVGA I believe, but I meant my Z77 system is micro atx. I moved back to my x58 since I need to be able to run 3 GPU's off of it with a riser(my Gene V doesn't work properly with my setup with risers).

It's a Xeon W3520 with a Classified E760. I'm using all the power saving methods it allows for, but I just left pretty much everything on auto. I could probably work a bit with the voltage, but I'm not sure it would do that much. If HWinfo can be trusted, it says it's fluctuating between 0.98ish volts and 1.258.
 
770W x 85% efficiency = 655W.

Did the PSU shut down, or did the motherboard shut down the PSU? The latter could be caused by the CPU not making good contact with its heatsink.
 
larrymoencurly said:
770W x 85% efficiency = 655W.

Did the PSU shut down, or did the motherboard shut down the PSU? The latter could be caused by the CPU not making good contact with its heatsink.
Click to expand...

I have no idea actually, I just came back(I have it in my garage for mining) and it had restarted. But my 7950 showed a hardware error and the driver crashed, so I assume it was due to that.
 
Kill-A-Watts are incredibly inaccurate and unreliable. I wouldn't sweat it when Kill-A-Watts are +/-40% accurate in the same batch number.
 
+/- 40% is huge variance. I know Thermaltake makes power supply testers but I do not belive they can test when it is running like a kill-a-watt can.

You would be surprised how much adjusting voltage can help.

Just food for thought, my Phenom II 955 stock clocks and voltage uses ~118w loaded idles ~80w(this is with the older C2 revision not the better C3 which I do have) stock voltage is 1.325-1.39v depending this makes it sit at idle/load 32-48c its low clock is 800Mhz voltage at .980 which brings power down to ~68w. I have dropped the top voltage to 1.183v low voltage .945 this drops power consumption down to ~98w mid range ~74w low clock ~56w idle/load temps 24-35c I used a kill-a-watt myself but am mostly going by a review toms and other sites did on adjusting stock voltages and how much power consumption difference it can make, temps wise I can see myself obviously.

My gpu
stock clocks are 1050/1200 idle 150/300 at 1.219/.950 volts temps it gets are 35c idle at 56% fan and load 64-67c at 72% fan.
I overclock it via afterburner so it automatically makes wonky numbers until I reboot and it "resets"
Anyways I can run it at 980/1375 1.165 loaded ~56c 62% fan, Idle 150/150 at .865v by hitting reset, fan at 53% and temps 25c-30c. FYI I use CGminer version 2.11.4 vectors 2 helps me a ton.

There is usually a heck of a lot of room to devolt if you tinker with it long enough and the temps if anything else speak for themselves. You do not always need alot of volts if its fine-tuned enough, and in many cases it can actually hit a bit more speed cause temps/voltages do not hinder it as much.

Don't forget many motherboards/chip makers(asus, intel amd etc) and such the auto settings 9/10 feed a good chunk more volts then they given they do not have/take the time to really tinker with them hence the variance for the voltage and tdp ranges its an average not a best case.

GPU wise 9/10 AMD and Nvidia both(though Nvidia seems not quite as often) they have a flat line voltage for their cards, for example for the 7870s the average is 1.218-1.22v stock, most cards I have seen people tinkering with the voltage they can drop this 1.195-1.212v stock clocks and if they declock just a bit like I have, can drop the voltage a massive amount like to the tune of 1.13-1.175 let alone if they game in a declocked state which surprisingly does not hurt performance as much as one would think but massively reduces power/temps, its all about the fine tuning and getting a nice roll in the silicon lottery :)
 
dragonstongue said:
+/- 40% is huge variance. I know Thermaltake makes power supply testers but I do not belive they can test when it is running like a kill-a-watt can.
Click to expand...
Those PSU testers are basically just voltage readers that applies a little load. If you have a multi-meter as well as a HDD or a fan to connect to the PSU, you basically have a PSU tester right there.

But yeah, 40% variance has indeed been confirmed:
http://hardforum.com/showpost.php?p=1032190998&postcount=7
 
Ok I don't understand anything anymore. I decreased my Twin Frozr's voltage down to 0.800V I believe, 100+ watts less on load(lost about 140 kh/s with that). When I did the bios mod for my Gigabyte 7950 from 1.250V to 1.090V. 0 watts less power usage. I'm honestly getting sick of all of this. It makes no fucking sense. :mad:
 
You must log in or register to reply here.
Back
Top