OT: Computer power supplies and you guys being smarter than I...

Page 2 of 2  
On Monday 18 March 2013 01:58 Metspitzer wrote in alt.home.repair:

Fairly sure - at least for the EU.
I had a small computer room at a previous job - 200-odd servers talking around 70kW total. This particular room had an industrial meter wired into the main incoming 3 phase circuit. Meter showed PF, amps and volts. PF was 0.95
The CE approval (for the EU) requires a PF >= 0.95 for >00W (lower soon, if it hasn't already). Given that most far eastern PC PSUs these days are wide input 100-240V so designed to ship anywhere, I'd be very surprised if there was much available in the US that did not conform - unless there are domestic PSU manufacturers and local regulation is not as strong?
--
Tim Watts Personal Blog: http://squiddy.blog.dionic.net/

http://www.sensorly.com/ Crowd mapping of 2G/3G/4G mobile signal coverage
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:

longer than a "right sized" or borderline supply. I almost always use 600 watt replacements - they generally cost me about $5 more than a 250.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Are you making the processor do a lot of shit ? Look at CPU usage.
Greg
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Sunday, March 17, 2013 2:36:00 PM UTC-7, Hench wrote:

The original Kill-A-Watt is rated 0.2% basic accuracy, and one person who tested it against an old Simpson analog wattmeter said the two units gave identical readings. I believe the Simpson is rated 2% accuracy, but I don't know if that's 2% of the reading or 2% of full scale. Another person who built an adapter for reading watts from the wall outlet (Electronics Design magazine, maybe a Bob Pease column) said the same. Also I tried plugging a P-4400 Kill-A-Watt into another P-4400 Kill-A-Watt (different internal designs, despite same model number) and got readings that matched within 1-2 counts. However a person who writes reviews of PC power supplies said he doesn't give efficiency numbers because he doesn't think the Kill-A-Watt is quite accurate enough, but that may simply mean its accuracy is in the range of 1% to 2% while the rest of his other instruments are way better.
Apparently Kill-A-Watt type devices are really good at measuring nonlinear type waveforms because I got about the same watt numbers from a cheap backup power supply, whether it was plugged into the AC outlet and putting out real sine waves or running off battery and putting out stepped square waves.
The 90W - 130W readings you get are probably accurate, while power supply estimation worksheets can be way off. XbitLabs.com and SilentPCreview.com measured power consumption and found that few systems with only one video card ever consumed over 500W:
http://www.silentpcreview.com/Power_Distribution_in_Three_PCs
http://www.xbitlabs.com/articles/cases/print/system-wattage.html
Even old 7200 RPM hard disks didn't consume more than 9W while spinning, maybe 15-20W while seeking. Here's what some contemporary drives consume:
http://www.xbitlabs.com/images/storage/1tb-hdd-roundup-4/Power-Consumption,-Idle,-Overall.png
http://www.xbitlabs.com/images/storage/1tb-hdd-roundup-4/Power-Consumption,-Random-Read,-Overall.png
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Mar 18, 9:15 am, snipped-for-privacy@my-deja.com wrote:

That helps confirm what I thought. The process of computing the power isn't complicated, it's simple. You just have to multiply voltage times current and do it fast enough that you are accurately sampling the waveform. It doesn't take a very complicated chip to do that. And we've had cheap consumer electronics for decades that does a lot more, eg calculators.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Mon, 18 Mar 2013 06:29:47 -0700 (PDT), " snipped-for-privacy@optonline.net"

Correct. There are some subtleties but that's basically it. Kill-A-Watts, and similar, are amazingly good at what they do for the price charged.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Mon, 18 Mar 2013 06:15:34 -0700 (PDT), snipped-for-privacy@my-deja.com
Re: OT: Computer power supplies and you guys being smarter than I...:

What is a "nonlinear" vs. "linear" type waveform, if you don't mind?
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Re

The terms may not be correct, but normal power follows a sine wave curve going from 0 volts to a plus voltage and back to 0. It then follows the same sine wave curve to a maximum negative voltage and back to 0. For what is often called 120 volts at 60 hz really has a peak voltage of around 170 volts. Most old analog meters were designed to 'average' this value to what is called a RMS value. Sometimes this has been the same voltage an equal voltage of DC would heat a resistive load.
Newer devices often take the power in pulses so it does not use the whole part of the sine wave curve, but just spikes of voltage or portions of the sine wave. It is often difficult to measuer this voltage/power with simple meters.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Mon, 18 Mar 2013 06:15:34 -0700 (PDT), snipped-for-privacy@my-deja.com wrote:

Good links. They push high watt PS's because they're more profitable. Kind of like a 500hp car that can go 180mph. But never does. Just a while ago I became aware of how much overpowered PS's are in an internet forum discussion. The guy who knew all about it caught hell. Somebody was doing a build and he recommended 400w. I thought that was low, and said so. He responded with facts and cites, so I reversed my position. He still caught hell from a bunch of others with overkill PS's. Some of it was his attitude. Wasn't shy being about being knowledgeable. Bottom line as add up peak component watt usage, and possible future additional watts, and buy quality at those watts. Most people have more wattage than they need, just like many have more hp than they need in their cars. There's an efficiency penalty, but with a PS it doesn't add up to much. I've got a 600w Corsair and a similar setup as the OP. I added a second GPU this year to get good PhysX in a game, and noted that by having the 600w PS I didn't give it a second thought. That was only reason for disagreeing with his lowball watt suggestion. He said 400w would have worked for me as well, so I really looked at all the numbers. He was right. But I don't regret having the 600w. It's overkill. Not a big deal. The 750w Corsair is overkill, but not a big deal unless you want to make it so.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Sunday, March 17, 2013 5:36:00 PM UTC-4, Hench wrote:

ow accurate these meters are when it comes to power supplies? Most power su pply sizing webpages imply that I'd be using 500-600 watts for my pc. I ass ume this is under 75 or 90% load or somesuch. Yet when I measure with a kil l-a-watt knockoff meter it's around 90-130 watts. Power supply in question is a corsair TX750. PC in question is an I7 3770K with 1 ssd, 9 sata harddr ives (all low RPM), 5 120mm fans, nVidia 650ti 1 gb vid card, 4 sticks of r am, 2 pci cards, a few usb 2 devices.
It's going to draw a lot more if you ut all those drives to work.
You don't mention the drive sizes but 9 is a lot. Why low speed? I'd go 7 200 or 10k myself for a media server. But the network is probably your bot tleneck anyway.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 3/17/2013 5:36 PM, Hench wrote:

Great power supply. I have several of those running our machines. The actual wattage that the machine uses should always be less than the capacity of the supply. It's certainly not hurting anything or costing anything to have an oversized supply
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 3/17/2013 5:36 PM, Hench wrote:
Will be sticking with the supposed oversized PSU.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Related Threads

    HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.