 pnjunctionTeksavvy ExtremePremium
join:2008-01-24
Toronto, ON
kudos:1 | reply to srr2
Re: I'm looking for a new quiet running PSU....Is yours quiet? I see your point many people probably go out and get 600W+ without even considering their system when 400-500W would be more than adequate. Fine if they want to upgrade later, but if they're never going past mid-range video chances are it will never be used.
I haven't done measurements, although I do want to pick up a simple meter like Kill-a-watt or something at some point. I just go by reviews.
On a power note, I'm looking forward to upgrading my 4870 to a 5870 or 5850. Preliminary review showed the idle power being 42 watts lower. With my system running 24/7 it will save me $5/month on my power bill! |
|
srr2
join:2001-12-20
Bethlehem, PA Reviews:
 ·RCN CABLE
| said by pnjunction:On a power note, I'm looking forward to upgrading my 4870 to a 5870 or 5850. Preliminary review showed the idle power being 42 watts lower. With my system running 24/7 it will save me $5/month on my power bill! Yeah. It's about time too that the video card people make some effort to minimize the power that those things eat. They all sit there burning up while the screen is blank! There's really no excuse for that. There hasn't been for years, except for their lack of consideration for what it costs to run them.
Most people don't realize how fast that power adds up and what it really costs them. If they did, they wouldn't be so giddy about running any of those distributed computing apps. For what it costs in electricity to "use up those unused clock cycles" (which may have been somewhat true back in the P3 days) they could make a cash contribution for the organization to buy some time on a real computer which is 1000x more cost-effective.
I'm hoping to save 25W on the worst of the machines that has to run 24/7. At the end of this year, power is gonna cost $0.15/KWH roughly. .025KW * 8760 = 219KWH = $32.85 So the payback is about 15 months.
42 Watts saved on the video card itself is a huge amount. Add to that the factor for power supply efficiency, let's say 80%, and that translates to 53Watts at the wall. Wow. You are almost in the range of justifying the purchase of the card on power saving alone. A payback of two years or less is a good deal for this kind of equipment.
I guess we've strayed a bit off the topic of silence here. But keep in mind that less power dissipated in the box means less air movement and less fan noise. Get the dissipation low enough and you *might* be able to get by on the power supply fan alone. That's where I'm headed with at least two of the three applications where these three new supplies are going. |
|
| reply to srr2
said by srr2:I have yet to see a "power estimator" that doesn't grossly overestimate. Among other flaws, they make the assumption that every component will simultaneously draw its worst-case power, which is both logically and statistically highly improbable, and frankly, not borne out in the hundreds of power measurements I've made. PCs draw a lot less, on average, than either the manufacturers or the power supply manufacturers would have you think. said by srr2:My point is that *everyone* overestimates the power that their machine draws and buys a supply that's much too large. Really. Just look at the typical power supply discussion in the Hardware Review forum here. It's positively ridiculous. Worse, virtually none of the rabid participants in those discussions knows what the hell they're talking about because none of them have made real measurements. Given that:
a) All power supply estimators are wrong, and
b) No one here knows what the hell they're talking about,
perhaps you could enlighten us with your method of PSU sizing. I'd like to hear more about these "hundreds of power measurements" that you've made.
--
The music business is a cruel and shallow money trench, a long plastic hallway where thieves and pimps run free, and good men die like dogs. There's also a negative side.
- Hunter S. Thompson (1937 - 2005) |
|
srr2
join:2001-12-20
Bethlehem, PA Reviews:
·RCN CABLE
| What do you want to know? I did, or had done for me, power surveys for HVAC sizing. The results of that belong to the clients that paid for the work. In summary though, the actual power requirements came out to be far below manufacturers' specifications in all cases. That's not unexpected because they don't want to be accused of underestimating consumption and throwing off HVAC estimates that choose to rely on them.
You will also note that I said that in the case of high power video cards, I take those requirements literally. Very few -- you could probably count the number on one hand -- of the machines I surveyed had video cards of that type.
I'd like to see a post from someone who has one of those massive systems and who actually measured the power it consumes. There may be one; I don't claim to have read every post. I'd like to see an actual working system that genuinely needs a 1KW power supply.
And yeah... there are a LOT of posters who throw around opinions like confetti without knowing a damn thing about the subject, especially where power supplies are concerned. Want to start counting wannabe engineers that think you can multiply volts times amps AC and come up with watts? Oh, and then there are the MTBF discussions. Some of them are so ridiculous that it's hard to stop laughing. I have yet to see anyone who actually knows how MTBF should be interpreted. Do you? Not to mention that the numbers that the manufacturers quote are patently ridiculous and/or are so loaded with unrealistic assumptions as to be laughable. Think anyone realizes that without specifying the methodology (including the underlying reliability data of components, interconnects, hazard rate, stress levels, and so on) by which MTBF is calculated, the numbers are completely meaningless? I wonder who'd spend the money to develop such a component database. It exists for MIL-SPEC equipment because the US Taxpayer paid for it. It's a hugely expensive and complex process to maintain it too. In the world of consumer products, there is no equivalent. So where, exactly, do you suppose those numbers that get thrown around come from? Field data? Hahahah. Just look at a power supply at random at Newegg, say. Look at the reviews and make a rough count of how many DOAs are reported. Then look at the claimed MTBF. See any inconsistency?
What do you find surprising about those power estimators that uniformly overestimate? They want to sell you the largest, most expensive supply that they can. Amazing, huh? For instance, they're perfectly happy to add 20W-25W for each of four disk drives, as if anyone's going to be doing butterfly tests on all four at once. 15W-20W for an optical drive. etc. If they want to assume absolute worst case conditions, fine. But there's no reason to choose a power supply on the worst case since statistically, the chances of it occurring are vanishingly small. And the data I've seen bear that out.
There's so much baloney in this subject area, and so many self-proclaimed experts, that it reminds me of the Hi-Fi Power business back in the 60s and 70s before the FTC stepped in and made them behave. You'd have companies selling amplifiers with a power transformer the size of a tennis ball, and push-pull output made of two 2n3055s, claiming that they can produce 800watts of "peak music power". ROFL! Uh-huh. Same thing now with Monster Cables, come to think of it. So you have people swearing that replacing their 6' generic line cord with some oxygen-free silver-plated $250 Monster Cable (or whatever) line cord "improves the sound stage incredibly". It all gets pretty silly after a while.
I suppose that makes me one of the resident cynics. |
|
| said by srr2:What do you want to know? ... I'd like to know your method of determining PSU size in desktop computers. You can take the rest of your smokescreen to Rants 'n' Raves. |
|
srr2
join:2001-12-20
Bethlehem, PA | My advice to you, for free. Buy the hugest, most expensive supply you can find. Make sure it has lots of pretty lights in it.
Other than that, you can send me a RFP for the work. |
|
| reply to srr2
said by srr2:There's so much baloney in this subject area, and so many self-proclaimed experts... It all gets pretty silly after a while. Well you've certainly proven that hypothesis from your evasive and rambling replies here. You have nothing to add but FUD.
Bye-Bye. |
|
 djr777Premium
join:2005-01-25
Pacific Grove, CA | reply to srr2
What do you mean by "modular" PSU? Removable cables or something else?
I was referring to removable cables.
I have not messed with a PSU and I was not sure if I was going to have to hard wire every connecting wire from the PSU to each component.
You prefer a PSU that is already harnessed with less chance of connection errors. That makes since to me.
--
...there will be an answer. let it be |
|
|
|
djr777Premium
join:2005-01-25
Pacific Grove, CA | reply to pnjunction
Thanks for those measurements.
I checked mine it is 5.9"x3.4"x 5.5". There is room for a 5.9" as well. Now it is time to find and order one!
Here is how it is boxed in.
I hope the wires all come out on the left side as well.
Thanks
--
...there will be an answer. let it be |
|
djr777Premium
join:2005-01-25
Pacific Grove, CA | reply to djr777
I just wanted to give an update:
I ended buying a Corsair TX650W...what a difference. My computer use to run fast most of time with my old Bestec 400W. I use this computer in the living room with either my 23" monitor or the 50" plasma. If we wanted to watch something on the plasma we had to compensate by turning up the volume.....not anymore!
Thanks for all your suggestions.
--
...there will be an answer. let it be |
|
.JPG/thumb.jpg "100621 bytes")
.JPG/thumb.jpg "87571 bytes")
.JPG/thumb.jpg "57679 bytes")