Sign in with
Sign up | Sign in
Your question

From PC Power and Cooling: Power Supply Myths Exposed!

Last response: in Components
Share
December 26, 2006 8:50:14 PM

Re: PC Power and Cooling - "Power Supply Myths Exposed!"

I was looking through their site and read this "article". I found it interesting in that it contradicts some of the newer claims and benefits touted by other PSU vendors.

Look at:
1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)

I'd be interested in some of you opinions on these PCPC claims :?:
December 26, 2006 11:39:02 PM

No one interested enough or knowledgeable enough to comment :?:
December 27, 2006 12:00:24 AM

1) Wattage Claims can be BS. You can claim 1kilowatt but not have enough amps to run a vid card and 3 drives.

2) Modular plugs, they're good so you can disconnect what you don't need.

3) 120 mm fan moves more air than an 80mm fan.

4) I think I prefer single rails even though I have multi. I have a feeling my 2nd 12v rail is overburdened while my primary is under used. I like the idea of having all the amps on one rail.

5) I live in a cool climate country so 1 fan is good enough for me. I could see two fans being useful if your ps is getting too hot.
Related resources
December 27, 2006 12:14:31 AM

Did you read the article? Especially the part about modular plugs and 120 vs 80 mm fans - so I assume you disagree?
December 27, 2006 12:42:00 AM

As a general rule, I don't look too far into information provided by a company about their own products. It's very easy to paint things in a light that favors your own product while "discrediting" the virtues of others.

I'd rather get my info from an unbiased, impartial 3rd party.
December 27, 2006 12:45:10 AM

Nice Thread. There were quite a few thinks i didn´t know.
December 27, 2006 1:02:15 AM

Quote:
Nice Thread. There were quite a few thinks i didn´t know.

There are "guru's" that might not agree. I'm no expert, but I love a good, documented discussion ...
December 27, 2006 1:14:25 AM

Quote:


Look at:
1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)



1) I personally have never thought about temperature's affect on power, but from my knowledge of engines i would now assume that they are right on this subject.

2) They don't tell us why resistance is increased when less is plugged in???? So this seems bogus, modular plugs are very usefull if you care about teh aesthetics.

3) A well engineered PSU with 80mm fan would certainly work as well/better than a poorly designed 120mm fan in a PSU. But the fact is most PSUs arent well engineered at all with many "hot" components far away from their ideal positioning. So yeah, if your going for high power a better engineered PSU would be better than one that just has a bigger fan. But if they actually did a study on various PSUs I think they would find the temps of the heatsinks would be about the same from 80mm to 120mm.

4) I don't know enough about this...

5) Extremely idiotic statement... looks as though they looked at 2 PSUs and made a generalization. Two 80mm fans put out a great more amount of air than 1 80mm fan... The small amount of space you lose to the additional fan seems like it wouldn't hurt the layout as much as a 120mm fan would, so this dual 80mm fan seems like the best cooling method for PSUs that are equally well laid out.
December 27, 2006 1:17:14 AM

I've read this before. Now, I actually do agree a bit with some of what they say. Now, do take this with a grain of salt, I'm not completely versed in electronics... and this was written by them, so it is a bit biased, but consider this: how good are their PSUs? From what I know, the 850 and 1kW units are all tested at 50C for extended periods of time to make sure they'll be able to provide that full 1kW in a regular environment without any hiccups. Now modular cabling I don't like... simply because I'd lose all the cables.

Wattage claims? Just look at the Tagan turbojet 1000, they said 1100 watts max and yet they only got it around 800... if that. Personally, I use PC P+C, but that's just me.
December 27, 2006 1:55:17 AM

All the points they make are valid, but design and quality of individual PSUs is more important. As is the application.

Although their points on multiple rails are accurate, there are other considerations such as rail stability that cause me to favor multiple rails over single rails. If your CPU is on the same rail as your CD-ROM the CPU's voltage supply will fluxuate when you put a disk in. For most systems this shouldn't be a problem as the PSU shouldn't fluxate by much and the power regulators on the mobo should be able to handle it. For an OCed Over Volted system you need cleaner power to keep your components happy as they teeter on the edge of stability.
December 27, 2006 2:25:42 AM

Quote:

Look at:
1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)

I'd be interested in some of you opinions on these PCPC claims :?:


Ok, I used to work on electronics for the Air Force so I know a few things about wattage and so forth. Not everything by any means, but a few things.

1. Wattage claims are hard to follow, partly because there seems to be a lack of industry standards and/or advertised claims. But their point is lost until and unless they provide a charts showing their psus compared to other psus. Perhaps Tom's could do a specific study on this sometime.

2. There's nothing wrong with modular plugs. Used to use them all the time in the Air Force and still do in many places. They are (or were) very common in aircraft wiring. The power loss does need to be compensated for, but that isn't as big an issue as the article states. How many 2 foot lines do we have in a computer anyway? Beyond that, are the contacts gold, silver, copper, aluminum, or some other metal? What is the resistance of the wire itself? It makes a difference. Even further, think about the solder connections between wires, etc. Those have resistance also.

3. I really don't understand their beef with 120 mm fans. Ok, my present psu only has a 80 mm fan size, but why would putting a 120 mm fan on it make all the internal parts smaller? Unless you put the fan entirely inside the psu instead of on the surface, where part is inside and part is outside. Or are they implying that all psu's have the same external size? In reality, the 350wt psu in my oldest computer is about half the dimensions of the 680wt psu in my newest computer, so the 120 mm fan making the psu components smaller doesn't quite make sense to me, unless they're arguing the same external dimensions.

4. The idea of single vs multiple 12V rails looks like a red herring to me. Power doesn't get lost if all the wattage on a 12v rail isn't used. The power just isn't used. Multiple rails make sure that each component gets what it needs and no more. So they make one rail and split that rail to component parts. Why is that more efficient? I don't see an explanation. Another part, which I admittedly don't understand, is why they say in one sentence that "the maximum current from any one 12-volt rail of a multiple rail psu is limited to 20 amps (240v/12=20)", but for their psu, this is not a limitation. Ok, I may be lacking in my education here, but it looks strange to me.

5. The 1 vs 2, or even 3 fans is kind of a "So what?" argument. One fan to exhaust out the back, perhaps one fan to exhaust out the top, maybe one to blow air in if 3 are used. If multiple fans do a better job at cooling the psu, what's wrong with that? It seems to me like someone saying that he only has one fan to cool his PC and that people who use more than one fan are stupid. Maybe I'm missing something here, but it seems like only an argument raised to defend their having 1 fan instead of multiple fans.

Overall, at least some of what they say seems to be a case of blowing their own horn without providing any real proof of their claims. Until and unless I see a third neutral party doing tests to verify all their claims, I remain sceptical of the whole of their claims. There was some good information for sure, but also some of what I think is hot air.
December 27, 2006 2:32:02 AM

I'll invite PCP&C in to defend/debate this "article" - it's a good debate.
December 27, 2006 3:51:50 AM

Quote:
Re: PC Power and Cooling - "Power Supply Myths Exposed!"

I was looking through their site and read this "article". I found it interesting in that it contradicts some of the newer claims and benefits touted by other PSU vendors.

Look at:
1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)

I'd be interested in some of you opinions on these PCPC claims :?:


First, this is obviously designed to promote PC& C, as it should be coming from their website... But lets go over the points that you've raised...

1. Wattage ... No real data (say from UL) presented, just their claims that PSU's fall off as they show on their charts.. So we are left to wonder what PSU's in general do under heat, and as well have to take PC&C's word that their TC 510 behaves as the graph shows.

2. Modular plugs. OK, I can buy that the plugs might introduce a few ohms extra resistance into the cabling. But if we followed the logic to its conclusion, all PSU's should be soldered in place. Personally, its a non-issue for me, as long as the modulars are well-made, good pins, etc., they MIGHT introduce a few mil ohms....

3. 120 MM fans. A top mounted fan for me will not work too well, as I generally have top mounted PSU's. Two 80's vs. one 80, or 1 one 120? I'd like two fans, as long as one was enough to do the job.

4. Single vs multi 12v rails. OK, maybe a valid point. Lose one rail, whether its a single or a multi, the PSU's most likely DED.

My 2p.
December 27, 2006 4:48:49 AM

Quote:
Look at:

2. Modular Plugs - A good thing?


If you really worry about this, then you need to get Cramolin and regularly clean your 24-pin ATX connection, device Molex connections, etc. I mean, all of your PS connections to devices are modular, right? So you already tolerate those, so why does increasing the number of modular connections by 30 percent (assuming 24-pin and CPU power is hard-wired) make such a critical difference?

Quote:
5. Single vs 2 fans (?)


When I tried my first Mushkin PS, I was worried about it having two 80mm fans, but actually they are quiet. Loud fans used to be the big beef against PCP&C (along with huge size and mass on their KW supplies), but they listened to the customer and found quieter fans.

Quote:
I'd be interested in some of you opinions on these PCPC claims :?:


The bottom line to me is that they offer a range of wattages, the rails are rock solid and the ripple is low. I'd like them better if they were less expensive and had modular cabling as an option, but then again, I'd like alot of things I can't have.
December 27, 2006 1:25:14 PM

Quote:
3. I really don't understand their beef with 120 mm fans. Ok, my present psu only has a 80 mm fan size, but why would putting a 120 mm fan on it make all the internal parts smaller? Unless you put the fan entirely inside the psu instead of on the surface, where part is inside and part is outside. Or are they implying that all psu's have the same external size? In reality, the 350wt psu in my oldest computer is about half the dimensions of the 680wt psu in my newest computer, so the 120 mm fan making the psu components smaller doesn't quite make sense to me, unless they're arguing the same external dimensions.


When your 350W is smaller than your 680W, that is in length, right? Like, it goes "deeper" in the system, closer to your optical drives? But, not in height, I imagine? All ATX PSUs are the same height. When you put a 120x120x25 PSU in the top, you are losing 25mm of space at the top, forcing you to make all your components a bit smaller. When you have a 80mm in the back, you are losing 25mm in the front of it, but you can make up that up by just making a longer at the back!

I never thought this was an issue, but..

I'd like them to defend it, see what they say..

~Ibrahim~
a b ) Power supply
December 27, 2006 3:28:59 PM

1. Agreed. The efficiency of the PSU does have a direct impact on power usage. 80%+ efficient PSU are more commonplace now than they were a year ago. That’s probably due to Energy Star’s strict 80+ program. Electronic devices must consume less power while idling and at load to receive an Energy Star sticker (to the best of my knowledge).

2. Agreed. Higher temperatures do tend to decrease efficiency. Now I under why Antec claims their PSU are tested in 50C – 52C ambient temps. That’s the internal temperature, not the temperature of the room. If the ambient temp in the room was 50C, then there is no way in hell I would turn on my PC.

3. I would have to agree their claim about modular PSUs. While convenient, connection points generally have higher resistance than the wire itself, which can lead to a slight power drop. However, no one has really done any analysis regarding how much power is lost. For the moment; I’ll say it’s personal preference. I’ll stick with a traditional PSU.

4. I generally do not read all the fine print. But marketing hype is true for all products, not just PSUs. I did spend about 3 months of researching before finally buy my Seasonic S12 500 PSU last year.

5. Agreed. Who cares what the wattage is if you do not know how many amps each rail can provide at constant power?

6. I prefer a 120mm fan because it is quieter than a 80mm fan. That’s one of the reasons why the Seasonic S12 series is so quiet. I really have to strain my ears to hear the fan. However, not all fans are the same. The better fans tend to have lower acoustic characteristics while still being able to move a good amount of air (CFM). I think the S12 series uses an ADDA fan. Generic fans tend to be noisy even under low voltage. The exhaust from the S12 500 is actually quite when my system is being stressed (i.e playing a FPS game and encoding video at the same time)

7. I generally avoid 2 fan PSUs because of noise.

8. I generally agree that too many 12v rails can be a bad thing. A single rail that’s able to provide massive amounts of power is good so you don’t need to worry about balancing the power drawn from each rail. Having two 12v rails seems like a good balance. Three, four or more can lead to wasted amps.
December 27, 2006 4:05:13 PM

owned

I happen to like my Hyper Type-R MODULAR PSU.
December 27, 2006 4:38:48 PM

Quote:


Look at:
1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)



1) I personally have never thought about temperature's affect on power, but from my knowledge of engines i would now assume that they are right on this subject.

2) They don't tell us why resistance is increased when less is plugged in???? So this seems bogus, modular plugs are very usefull if you care about teh aesthetics.

3) A well engineered PSU with 80mm fan would certainly work as well/better than a poorly designed 120mm fan in a PSU. But the fact is most PSUs arent well engineered at all with many "hot" components far away from their ideal positioning. So yeah, if your going for high power a better engineered PSU would be better than one that just has a bigger fan. But if they actually did a study on various PSUs I think they would find the temps of the heatsinks would be about the same from 80mm to 120mm.

4) I don't know enough about this...

5) Extremely idiotic statement... looks as though they looked at 2 PSUs and made a generalization. Two 80mm fans put out a great more amount of air than 1 80mm fan... The small amount of space you lose to the additional fan seems like it wouldn't hurt the layout as much as a 120mm fan would, so this dual 80mm fan seems like the best cooling method for PSUs that are equally well laid out.

2) I now see what they were talking about, i misunderstood. I thought they were saying that haveing less plugged in is worse. When in fact they meant the items that are plugged in have that additional connection which DOES increase resistance. To consumers i think the aesthetics are more important than the tiny amount of resistance added to the system.
December 27, 2006 4:41:39 PM

Quote:
If the ambient temp in the room was 50C, then there is no way in hell I would turn on my PC.


Sometimes you don't have a choice. About a year ago, my instrument was mounted in an aircraft that routinely had ambient temps in the high 40s for the first hour of the flight. At idle, the air in our controller PC would run in the mid-50s. It's a battle to keep stuff alive in that kind of environment but it can be done.
December 27, 2006 5:24:53 PM

Quote:

When your 350W is smaller than your 680W, that is in length, right? Like, it goes "deeper" in the system, closer to your optical drives? But, not in height, I imagine? All ATX PSUs are the same height. When you put a 120x120x25 PSU in the top, you are losing 25mm of space at the top, forcing you to make all your components a bit smaller. When you have a 80mm in the back, you are losing 25mm in the front of it, but you can make up that up by just making a longer at the back!

I never thought this was an issue, but..

I'd like them to defend it, see what they say..

~Ibrahim~


A couple things here. Maybe I should have said that the 350wt was REALLY old. None the less, it is smaller than the newer 680wt in all dimensions. A secondary thing is that there is a lot of room in my PC case for a fan to be sticking up high if need be. To help even more, I have the present psu mounted upside down, so if it stuck a bit farther into the case it wouldn't matter, assuming that the 120 mm fan was mounted on the outside of the psu.

My psu has two 80 mm fans and has no problem. I can't say by direct experience what a 120 mm fan would be like. But PC&C said they didn't even like two 80 mm fans, so my psu is bad by their standards.
December 27, 2006 5:25:17 PM

Quote:

What kills me is how people won't hesitate to spend $1100 on a quad-core CPU (despite nearly no software that takes full advantage of it yet), or even $1200-$1300 on a pair of 8800 GTX's, yet bitch about spending more than $60 on a power supply! When I hear that whining, I just pray that the person has nothing but trouble with their system...


The thing you are missing here is that if you spend $1100 on a CPU or GPU, you will notice a performance increase compared to say a $500 product, but if you spend $300 on a PSU vs a $50, your PC runs exactly the same. So people just aren't very motivated to put money into that. I think until you actually have a cheap PSU fry something, you don't know the value of an quality PSU.

That being said, I don't think I will ever buy a P&C PSU. They are way too expensive. Don't get me wrong here, quality is important, at the moment I think my PSU is more expensive than my processor (AMD 3700 / Seasonic M12), but I don't think that the 'top of the line' PSUs are really worth it.

As for the modular claims, they are probably true, but I can accept a mild loss of power to keep my case cleaner. You may even gain the power back with increased airflow and lower temps in the PSU because of this (doubtful, but I can rationalize this all I want :)  ).
a b ) Power supply
December 27, 2006 5:29:26 PM

Quote:
If the ambient temp in the room was 50C, then there is no way in hell I would turn on my PC.


Sometimes you don't have a choice. About a year ago, my instrument was mounted in an aircraft that routinely had ambient temps in the high 40s for the first hour of the flight. At idle, the air in our controller PC would run in the mid-50s. It's a battle to keep stuff alive in that kind of environment but it can be done.

Well, that's a special circumstance. But I agree, sometimes you don't really have a choice and you must manage heat as best as possible given the conditions.
December 27, 2006 5:39:15 PM

Quote:
The thing you are missing here is that if you spend $1100 on a CPU or GPU, you will notice a performance increase compared to say a $500 product, but if you spend $300 on a PSU vs a $50, your PC runs exactly the same. So people just aren't very motivated to put money into that. I think until you actually have a cheap PSU fry something, you don't know the value of an quality PSU.

That being said, I don't think I will ever buy a P&C PSU. They are way too expensive. Don't get me wrong here, quality is important, at the moment I think my PSU is more expensive than my processor (AMD 3700 / Seasonic M12), but I don't think that the 'top of the line' PSUs are really worth it.

As for the modular claims, they are probably true, but I can accept a mild loss of power to keep my case cleaner. You may even gain the power back with increased airflow and lower temps in the PSU because of this (doubtful, but I can rationalize this all I want :)  ).


Interesting analysis and I agree for the most part. The thing is, you can be on the edge with a power supply and have a computer that will start and run low power consumption apps just fine but then choke when current demand goes up. For example, my son's game box ran fine at stock clock on the CPU + GPU and could even handle a mild CPU overclock as long as the app was not too demanding. But if the GPU was goosed up much at all, it could not handle demanding games like Oblivion. The less demanding apps were fine with even a big GPU overclock. So I stepped up his power supply to a unit with over 40 amps on the +12 and it now runs fine in all apps with a 500MHz OC on the CPU and 200MHz on the GPU. So his PC did not run the same after going to a higher priced power supply - it ran better.
December 27, 2006 5:40:15 PM

I too can't see why two 80mm fans wouldn't help at all. One intake, one exhaust, or both exhaust. More airflow is always better. I hope they're not saying that you lose heat-sink size due to adding more fans, because there are ways around that too.

I used to have a PSU that had 4 80mm fans. They were rather thin fans, but there was plenty of airflow and decent space for heatsinks. I replaced it because it didn't have the correct plugins for new video cards, and I didn't want to keep adding adapters here and there (thus cluttering my case even more). The PSU is currently in a friend's computer and running strong.

Their arguement that the fact people spot-cool their heatsinks means weak thermal management is completely stupid. IMO active cooling (fans) is far greater than passive cooling (bigger heatsinks), and having both big heatsinks PLUS a fan spot-cooling the hottest places would seem smart to me. I mean, doesn't that ultimately reduce temperature? Isn't that the purpose of thermal-management? And some of these PSUs have variable fan speeds which, in a way, simulates adding a larger heatsink. Aye?

You could totally go back and forth arguing every point they make. The only way to really prove something is to test it and document it.
December 27, 2006 5:41:18 PM

Quote:
The thing you are missing here is that if you spend $1100 on a CPU or GPU, you will notice a performance increase compared to say a $500 product, but if you spend $300 on a PSU vs a $50, your PC runs exactly the same. So people just aren't very motivated to put money into that. I think until you actually have a cheap PSU fry something, you don't know the value of an quality PSU.


I think you hit the nail on the head there. People don't want to upgrade something that doesn't cause a performance increase. And yes, most of us are reactive vs. proactive.

~Ibrahim~
December 27, 2006 5:45:32 PM

First, to clear up what they were saying about "trapped" amperage on the multiple 12v rails. They aren't talking about wasted electricity, they are talking about wasted potential. If you buy a quad rail PSU, max out two of the rails and leave the other two nearly idle you just paid for a quad-rail ~1000w PSU but you're using it as a dual-rail ~700w PSU and a single-rail 800w PSU might have been a better choice. A valid point but kind of a non-consideration; you should pick the PSU that will power your system best while meeting your budget and extra rails aren't going to hurt anything.

Quote:
I'll invite PCP&C in to defend/debate this "article" - it's a good debate.


There's not much to debate, they didn't many specific claims. Other than the one graph there's nothing about this article that is particularly contraversial. The contraversial part is left unspoken for the reader to assume: that PSUs that have these attributes are inherrantly inferior to their own for all applications. In the big picture there are two general counter points to their article:

1. You could break every single one of these rules, at the same time, and still make a better PSU but it would be a bit of an uphill battle.
2. There are some situations where it would be desirable to break some of these rules.

Really, they are encouraging people not to buy into marketing hype (AKA "Myths") but to look at the actual performance of the PSUs. This is very good. Unfortunately they are doing it in an article that is marketing hype and only gives one real example of actual performance (and one really vauge example comparing them to "Brand X")... Some people are too lazy to compare real performance. It looks like PCP&C is just trying to give them what they want so they can sell more product. This article could be quite useful to someone looking to get into DESIGNING PSUs (although I would hope they would have education and intelligence that would allow them to figure these things out for themselves before they thought they could engineer complex electrical components) but while quite accurate it's pretty useless to someone looking to merely BUY a PSU.
December 27, 2006 5:54:23 PM

Quote:
I too can't see why two 80mm fans wouldn't help at all. One intake, one exhaust, or both exhaust. More airflow is always better.


One intake and one exhaust does not give you significantly more airflow. The air can only leave the enclosure as fast as the exhaust fan can remove it. An additional intake fan can help improve pressure differences which will give a slight benefit to airflow at best. That was their point. It can give you much better spot cooling though and help to mitigate poor airflow design problems. If you design your enclosure so it has good airflow and doesn't need extra spot cooling you don't need a push/pull config and a single fan will be quieter, generate less heat, and use less electricity. But like you said, if extra spot cooling gives you more power without breaking other design goals: do it.

Unless the manufacturer put "exhaust" fans that blow into the case you can't put two 80mm exhuast fans on an ATX PSU, there isn't room.
December 27, 2006 6:28:37 PM

Its true that having a crappy PSU will often give you random reboots or even other weird glitches that are often hard to diagnose. When overclocking, having a quality and higher wattage PSU becomes more important due to the significant wattage increases.

For most people though (the ones who buy the cheap power supplies), if they build their PC and run 3dmark a couple of times with no issues, then their PSU is probably going to be fine assuming it doesn't blow.

I guess the main issue is that people upgrade their PCs without upgrading their PSU. I have a friend that has a three year old PC and is upgrading his GPU to a x1950pro AGP. His stock PSU is some crap 350W and I pretty much forced him to get a new PSU since he wants to get into overclocking (on his new card and his prescott).
December 27, 2006 7:18:25 PM

Quote:
I too can't see why two 80mm fans wouldn't help at all. One intake, one exhaust, or both exhaust. More airflow is always better.


One intake and one exhaust does not give you significantly more airflow. The air can only leave the enclosure as fast as the exhaust fan can remove it. An additional intake fan can help improve pressure differences which will give a slight benefit to airflow at best. That was their point. It can give you much better spot cooling though and help to mitigate poor airflow design problems. If you design your enclosure so it has good airflow and doesn't need extra spot cooling you don't need a push/pull config and a single fan will be quieter, generate less heat, and use less electricity. But like you said, if extra spot cooling gives you more power without breaking other design goals: do it.

Unless the manufacturer put "exhaust" fans that blow into the case you can't put two 80mm exhuast fans on an ATX PSU, there isn't room.



I know some that do, incidentally. In fact, when you have only a single 120mm fan on a PSU with no exhaust fan on the back (mesh design), you effectively have hot air being blown back in to the case through the sides of the PSU. This is with every PSU design though as they all have vents around the enclosure. I also think that even though, like you said, adding an intake and exhaust will only equal out the pressure and give you marginal gains, that the intake fan can hit a nice hot-spot and help with cooling better than just pulling air through the PSU. There's a lot of potential for pockets of hot air being trapped with a single fan design unless it's a 120mm fan that blows directly on top the components, shooting the hot air out the sides. That's where PCB/HS designs must be done well. If you have a single 80mm fan sucking air out the back, you're probably missing a lot of the hot stuff behind the heatsinks unless there's something to force it out, i.e. an intake fan. It's my understanding that the best thermal management is to let the air glide across the heatsinks rather than having the airflow in opposite directions with the heatsink being nothing more than a blockage.

If designed right, a 2-fan design will outperform a single-fan design IMO.
December 27, 2006 7:27:21 PM

I personally think that a 120mm intake and an 80mm exhaust is probably a better cooling solution that the front / back push / pull and better than any single fan.
December 27, 2006 7:30:17 PM

Yup I agree. My PSU has a 120mm intake on the bottom and an 80mm exhaust, and a mesh shell. The exhaust fan doesn't move too fast (and is thus nearly silent), but it's enough to direct most of the heat out the back. It also stays on for 5min after I shut off the PC to keep it nice and cool :) 
December 27, 2006 7:43:20 PM

Quote:
leo2kp wrote:
I too can't see why two 80mm fans wouldn't help at all. One intake, one exhaust, or both exhaust. More airflow is always better.



One intake and one exhaust does not give you significantly more airflow. The air can only leave the enclosure as fast as the exhaust fan can remove it. An additional intake fan can help improve pressure differences which will give a slight benefit to airflow at best. That was their point. It can give you much better spot cooling though and help to mitigate poor airflow design problems. If you design your enclosure so it has good airflow and doesn't need extra spot cooling you don't need a push/pull config and a single fan will be quieter, generate less heat, and use less electricity. But like you said, if extra spot cooling gives you more power without breaking other design goals: do it.


In power supplies it is a bit different, usually there is a massive heat sink rising up in the middle of the supply, pulling air from one end helps airflow move around this heat sink. Adding another fan to blow air into the PSU only bounces air off this heat sink while the other sucks air off of one side of the heat sink.

Power PC and cooling was talking about a top-mounted 120mm fan in there myths exposed which is correct in limiting the vertical height of components, which limits their ampage rating.

Also modular plugs do have some resistance at the connector which robs you of power. Using a 12V rail as an example, if a modualr connector had 1 ohm of resistance then using ohm's law E2/R=W or 144/1=144 watts of power is being used by that connector alone! Even if it is 0.5 ohms that is still 88 watts! Plus the fact that modualr connectors introduce another failure point.

I have been using Power PC and Cooling supplies for 4 years and have yet to have any problems whatsoever from their supplies. But plenty of other problems from other hardware. Plus it only takes once to have a power supply fail and take with it your video card and hard drive to teach you a valuble lesson. You need a good noise free PSU in todays power hungry hardware.
December 28, 2006 12:29:36 AM

Quote:
I personally think that a 120mm intake and an 80mm exhaust is probably a better cooling solution that the front / back push / pull and better than any single fan.


Just got my new power supply. It really goes against the "advice" from PC&C Power and Cooling. Its a 700wt Thermaltake W0106RU with modular cables, one 140 mm fan, multiple rails, and only rated for up to 85% efficiency. Hey, if a 120 mm fan is bad, what's a 140 mm fan like? And I thought Thermaltake was supposed to be a good psu. Oh well.
December 28, 2006 2:10:59 AM

how about you take a little time and research your possible candidate for purchase. I spent two weeks researching 4 PSU's before i settled on the Ultra X-Finity 600w, of the 4 i was looking at it held consisten amperage across the 12v rails, and it actually was withing +-1v of the 3v 5v and 12v rails.

Take some time and look up 3rd party reviews and read what people have to say, and pick one that fits your needs, and check your video cards required amperage as the x1900 needs 30amps which is insane.
December 28, 2006 2:14:22 AM

Quote:

1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)


1) I study Physics at college and it's a fact that Resistance is Directly Proportional to Heat. Ie More heat more resistance and therefore lower voltage/current/wattage

2) don't know about these but if the materials connecting are different maybe the resistance in them are different.

3) I've seen my 120mm fan and its less than an inch high. Unless you got a teeeny psu or a real high 120mm fan then this is nonsense.

4) dont know

5) 2 fans will always push more air than one fan, unless they are running at lower speeds than the one fan.
December 28, 2006 3:58:20 AM

Quote:
Re: PC Power and Cooling - "Power Supply Myths Exposed!"
I was looking through their site and read this "article". I found it interesting in that it contradicts some of the newer claims and benefits touted by other PSU vendors.
Look at:
1. Wattage Claims
2. Modular Plugs - A good thing?
3. 120 mm fans (?)
4. Single -vs- Multi 12 volt rails.
5. Single vs 2 fans (?)

It looks to me like the editorial constraints of cramming the Monthly Myth into one page, has short-changed the points covered in the subject matter.

For example, purchasing a power supply that alleges to be more efficient MAY pay off no matter what, but the best payoff by far results from choosing a correctly sized unit so that typical power draw is within the most efficient range. IOW, don't buy a 1000W supply for a system that draws 200W at idle at 400W at full throttle.

It's true, wattage claims are not always perfectly trustworthy. I've seen many reviews that have fully loaded power supplies to manufacturer's claims, just to see them fail in minutes or hours. Even for trusted manufacturers I still like to have 100W of "headroom".

Modular plugs do introduce resistance. You're actually adding three extra electrical joints to each wire: the crimp at the connector inside the PSU, the press-fit joint between the two mating connectors, and the crimp on the exterior wire leading out to the devices. But in the big scheme of things, there's also at least one solder joint from the PSU rail, a crimp at the terminating Molex or whatever, the termination joint, and a solder joint in the device. So 7 vs. 4 joints along the path from point A to point B. And since most PCs are in fairly dry, vibration -free environments I would not worry too much about loosening and corrosion.

A good set of joints should add up to about .03 Ohms at those extra three connections. At 18A (a number that I chose because it's a typical max rail rating in these 12V supplies), that's a .6V drop in what reaches the device, which is within the tolerance of most 12V equipment.

OTOH, unless the PSU manufacturer mentions whose connectors it's using and you check their their specs for duty cycle rating, aging, retention, etc., I would try not to abuse those connectors too much.

I'm a little suspicious of that ExtremeTech link. Nothing personal. I mean, it's impressive that they though of that "rundll32.exe advapi32.dll,ProcessIdleTasks" command and of manually setting the clock rates for their benchmarks. On the other hand I get the impression that the ASUS board was the one they were having PS issues with, and they left PEG-link mode on! Maybe I'm making a mountain out of a molehill but I gotta wonder.

I'm awfully suspicious of the logic criticizing 120mm fans, which by the way are bottom-mounted, not top, no?? Smaller components are not necessarily less durable, to begin with. And less volume inside the power supply means higher air velocity through the unit, which should offset the decreased heat sink area, if the area is in fact decreased. OTOH they do ignore the scattered air flow problem caused by a big fan that aims ninety degrees away from the air outlet. And all PSUs with variable speed fans will ramp up the fan speed at 80% load. Solution: over-spec your supply and eat the trade-off in efficiency. Finally, a larger fan will always move the same volume of air with lower noise, all other design and quality factors being equal.

The issue of "rails" could be a Myth of the Month all by itself. The problem is that there is no solid definition of what a rail is, so all kinds of crap arrangements are referred to as multiple rails when in reality some offer little benefit beyond relieving the need for #6 wire on the 12V lines. That loose definition of "rail" has a direct bearing on the accuracy of the claim about under-utilization trapping available power delivery potential.

Separating the current really is important by the way. Let me put this into perspective for you... an 800W PSU might allow 650W on the 12V rails, which is over 54A. Does that number sound familiar? That's cranking current for an automobile. Now pop open your car's hood and look at the size of the wires connected to the battery. Get my drift here? Of course there's some hyperbole here, as those wires are rated by a lot more current than that, but there really is a practicality issue when delivering huge currents.

-Brad
December 28, 2006 4:49:29 AM

Quote:
Also modular plugs do have some resistance at the connector which robs you of power. Using a 12V rail as an example, if a modualr connector had 1 ohm of resistance then using ohm's law E2/R=W or 144/1=144 watts of power is being used by that connector alone! Even if it is 0.5 ohms that is still 88 watts! Plus the fact that modualr connectors introduce another failure point.

Your argument is completely fallacious. In order to keep it simple let's assume a series circuit. Your 1 ohm resistance of the connector is the total resistance of the circuit. This generates a current of 12 amps I=E/R or 12=12/1. Obviously there is going to be additional resistance in the circuit. Again for the sake of simplicity lets assume an additional 999 ohms of resistance in our fictional series circuit. Current, Which remains the same through the entire series circuit, is I=E/R or 12/1000 or .012 amps. The wattage of the total circuit is now 12x12/1000 or .144 watts. The calculation to determine the voltage drop, across the 1 ohm connector, is E=IxR or .012x1 or .012Volts. Wattage is ExI or ExE/R or .012x.012/1= .000144 watts consumed by the 1 ohm resistance in this fictitious circuit. By the way 144/.5 is 288 watts not 88 watts. Do you see how ridiculous your statements are? Understand, as previously stated, this is a totally fictitious oversimplified series circuit and has really nothing to do with the power supply discussion. (No hijack intended). But, before you start spouting off, you should take a refresher in Ohm’s law.
December 28, 2006 5:02:39 AM

Thanks... Much better said than most, the points better laid out and covered in exquisite detail with thoughtful logic.
December 28, 2006 5:52:15 AM

bberson:
Please! Automotive starter current is usually in the 200 to 450 amp range!!!

Others: As an electrical engineer with 30+ years experience I applaud all of the discussion here. There are a few issues I have an opinion about and I would like to express.

Having worked with computer power supplies since 1973 (for Interdata Computer Systems and several of the suceeding companies), I prefer dual fan power supplies. If they are thermastatically controlled, they do not usually make more noise than tollerable and I have empirically found they usually cool the PS sufficiently. My real bone to pick is with air re-entering the back of cases when it should enter from the bottom and flow the normal "heat rises" path to the top. Otherwise you are fighting the laws of Physics.

I have also experienced several instances where too much airflow is worse than not enough. Too much air flow will not allow the air to be in contact with the heat sink long enough to transfer the heat away from the metal to the air. It is simply a balancing act.

Dual +12 volt rails are ALWAYS preferable to one. Split the load and you will suffer less voltage loss. (simple OHM's Law)

To me the power supply is usually the weakest part of any computer system. Skimp on the PS and you are looking for trouble every time.

To finish up, here is a thought nobody seems to ever discuss in any PS thread I have seen. When you turn off your PS, always wait a minimum of 15 seconds before you turn it back on or you run the risk of breaking the welds on the plates of the primary input capacitors! It took us almost 15 months to diagnose this problem back in the late 1970's. As a result, we advised our customers to follow this rule and the RMA returns fell drastically! The primary PS circuit designs have not changed. The same rule applies now!
December 28, 2006 1:00:10 PM

Quote:
Please! Automotive starter current is usually in the 200 to 450 amp range!!!

Warm starting current for today's 4-bangers can be closer to 100 amps but HEY, I did say there was some hyperbole there.

Quote:
Dual +12 volt rails are ALWAYS preferable to one. Split the load and you will suffer less voltage loss. (simple OHM's Law)

I think this forum should have a test question before you can access it, like "E=I*?".

Quote:
always wait a minimum of 15 seconds before you turn it back on or you run the risk of breaking the welds on the plates of the primary input capacitors

When I see "primary input" I can't help think of the primary side of the transformer. Which is of course ridiculous. Is that really what those caps are being called??

-Brad
December 28, 2006 2:47:20 PM

Quote:
I personally think that a 120mm intake and an 80mm exhaust is probably a better cooling solution that the front / back push / pull and better than any single fan.


Just got my new power supply. It really goes against the "advice" from PC&C Power and Cooling. Its a 700wt Thermaltake W0106RU with modular cables, one 140 mm fan, multiple rails, and only rated for up to 85% efficiency. Hey, if a 120 mm fan is bad, what's a 140 mm fan like? And I thought Thermaltake was supposed to be a good psu. Oh well.

I think the TT units are pretty decent. One thing to keep in mind about fans for power supplies is that many units are very congested - the air volume inside is proportionally small. For example, thing about the air flow to volume ratio of a CPU HSF. In a PS, you have bulky items that take up a bunch of space and they do not have a high surface area to volume ratio like capacitors, tformers, etc. Heat sink volume in many power supplies is nearly nonexistent, so one advantage of a push-pull is that it generates turbulence at two spots and reduces hot spots. But if a unit has two 8cm fans, they better be quiet fans or they could be the major sound source for your PC.
December 28, 2006 3:34:14 PM

I guess I don't know much about electronics like you guys do. I got the basics down enough to understand. I bought this power supply a few months ago because some cheapo thing that came with my case I had been using for a few years died on me and took my mobo with it...grrrr. Anyway, here is what I got http://www.newegg.com/Product/Product.asp?Item=N82E1681... . I was wondering if I chose wisely or did I choose poorly? I know its not the most efficient thing or the best but I don't have a lot of money either, so I was kind of limited. It runs WAY WAY better than my last one did and its way quieter and it actually lowered my system temps about 5 degrees due to the 135mm fan in the bottom. My voltages are way more stable than before and there are more amps on the 12v rails plus there are 2 12v rails instead of one and my last one even had a higher wattage rating but this one is lower and has more amps. Kind of off topic but let me know what you think so that in the future I can be less ignorant.

Thanks!
December 28, 2006 4:30:57 PM

Quote:
I personally think that a 120mm intake and an 80mm exhaust is probably a better cooling solution that the front / back push / pull and better than any single fan.


Just got my new power supply. It really goes against the "advice" from PC&C Power and Cooling. Its a 700wt Thermaltake W0106RU with modular cables, one 140 mm fan, multiple rails, and only rated for up to 85% efficiency. Hey, if a 120 mm fan is bad, what's a 140 mm fan like? And I thought Thermaltake was supposed to be a good psu. Oh well.

I think the TT units are pretty decent. One thing to keep in mind about fans for power supplies is that many units are very congested - the air volume inside is proportionally small. For example, thing about the air flow to volume ratio of a CPU HSF. In a PS, you have bulky items that take up a bunch of space and they do not have a high surface area to volume ratio like capacitors, tformers, etc. Heat sink volume in many power supplies is nearly nonexistent, so one advantage of a push-pull is that it generates turbulence at two spots and reduces hot spots. But if a unit has two 8cm fans, they better be quiet fans or they could be the major sound source for your PC.

I was being a bit facetious in my post. I do think the Thermaltake psu is pretty good, or I would have bought something else. My intent was that it seems to fly in the face of PC P&C' "Myths".
December 28, 2006 4:34:40 PM

Quote:

4. The idea of single vs multiple 12V rails looks like a red herring to me. Power doesn't get lost if all the wattage on a 12v rail isn't used. The power just isn't used. Multiple rails make sure that each component gets what it needs and no more. So they make one rail and split that rail to component parts. Why is that more efficient? I don't see an explanation. Another part, which I admittedly don't understand, is why they say in one sentence that "the maximum current from any one 12-volt rail of a multiple rail psu is limited to 20 amps (240v/12=20)", but for their psu, this is not a limitation. Ok, I may be lacking in my education here, but it looks strange to me.


That's because they obscured their answer in marketing speak, to geek up teenaged buyers. Here's a better explaination:

Say you need 20A for your graphics cards and 10 A for your CPU, but your power supply only puts out 18A per rail across two rails. What you end up with is 2A too little power going to your graphics cards, and 8A available power to your CPU that's not being used.

So you "wasted" 8A of capacity and your system still doesn't work right because you need 2A more for your cards.

Because of that, a single 36A rail supplying the cards and the CPU would be better than two 18A rails.
December 28, 2006 4:37:11 PM

Quote:
Your argument is completely fallacious. In order to keep it simple let's assume a series circuit. Your 1 ohm resistance of the connector is the total resistance of the circuit. This generates a current of 12 amps I=E/R or 12=12/1. Obviously there is going to be additional resistance in the circuit. Again for the sake of simplicity lets assume an additional 999 ohms of resistance in our fictional series circuit. Current, Which remains the same through the entire series circuit, is I=E/R or 12/1000 or .012 amps. The wattage of the total circuit is now 12x12/1000 or .144 watts. The calculation to determine the voltage drop, across the 1 ohm connector, is E=IxR or .012x1 or .012Volts. Wattage is ExI or ExE/R or .012x.012/1= .000144 watts consumed by the 1 ohm resistance in this fictitious circuit. By the way 144/.5 is 288 watts not 88 watts. Do you see how ridiculous your statements are? Understand, as previously stated, this is a totally fictitious oversimplified series circuit and has really nothing to do with the power supply discussion. (No hijack intended). But, before you start spouting off, you should take a refresher in Ohm’s law.


First off my example is an exaggeration because I don't think that modular plugs introduce 1 ohm of resistance. It was just an example, second you cannot assume a series circuit. Since there are many ground paths to the device connected it would actually be a parallel circuit (think of leakage circuit).

My attempt was to get someone to think about plugging a 18A rail into a CPU board that requires 17A. You can come dangerously close to the borderline from leak circuits, temp, and other factors. It would probably work starting out OK but what happens when your playing a game, or over clocking and you start to get weird failures like ram, or video problems, or shutdowns. Nobody wants to spend frustrating days troubleshooting only to find out a better PSU could have saved them both time and money.
December 28, 2006 4:46:56 PM

I haven't had any direct experience with Rosewill psu's, but looking at the reviews on Newegg, the psu doesn't stand out overall as a good one. Only 71% of the people rated it as a "5". Further, it has an efficiency rating of only >70%. If it is working better than your last psu, then good, but it doesn't look like the best. I may be wrong, but I wouldn't look to any long term reliability from this psu.
December 28, 2006 4:59:22 PM

Quote:

4. The idea of single vs multiple 12V rails looks like a red herring to me. Power doesn't get lost if all the wattage on a 12v rail isn't used. The power just isn't used. Multiple rails make sure that each component gets what it needs and no more. So they make one rail and split that rail to component parts. Why is that more efficient? I don't see an explanation. Another part, which I admittedly don't understand, is why they say in one sentence that "the maximum current from any one 12-volt rail of a multiple rail psu is limited to 20 amps (240v/12=20)", but for their psu, this is not a limitation. Ok, I may be lacking in my education here, but it looks strange to me.


That's because they obscured their answer in marketing speak, to geek up teenaged buyers. Here's a better explaination:

Say you need 20A for your graphics cards and 10 A for your CPU, but your power supply only puts out 18A per rail across two rails. What you end up with is 2A too little power going to your graphics cards, and 8A available power to your CPU that's not being used.

So you "wasted" 8A of capacity and your system still doesn't work right because you need 2A more for your cards.

Because of that, a single 36A rail supplying the cards and the CPU would be better than two 18A rails.

Now that makes a lot better sense to me.
December 28, 2006 6:11:45 PM

Quote:
Your argument is completely fallacious. In order to keep it simple let's assume a series circuit. Your 1 ohm resistance of the connector is the total resistance of the circuit. This generates a current of 12 amps I=E/R or 12=12/1. Obviously there is going to be additional resistance in the circuit. Again for the sake of simplicity lets assume an additional 999 ohms of resistance in our fictional series circuit. Current, Which remains the same through the entire series circuit, is I=E/R or 12/1000 or .012 amps. The wattage of the total circuit is now 12x12/1000 or .144 watts. The calculation to determine the voltage drop, across the 1 ohm connector, is E=IxR or .012x1 or .012Volts. Wattage is ExI or ExE/R or .012x.012/1= .000144 watts consumed by the 1 ohm resistance in this fictitious circuit. By the way 144/.5 is 288 watts not 88 watts. Do you see how ridiculous your statements are? Understand, as previously stated, this is a totally fictitious oversimplified series circuit and has really nothing to do with the power supply discussion. (No hijack intended). But, before you start spouting off, you should take a refresher in Ohm’s law.


First off my example is an exaggeration because I don't think that modular plugs introduce 1 ohm of resistance. It was just an example, second you cannot assume a series circuit. Since there are many ground paths to the device connected it would actually be a parallel circuit (think of leakage circuit).

My attempt was to get someone to think about plugging a 18A rail into a CPU board that requires 17A. You can come dangerously close to the borderline from leak circuits, temp, and other factors. It would probably work starting out OK but what happens when your playing a game, or over clocking and you start to get weird failures like ram, or video problems, or shutdowns. Nobody wants to spend frustrating days troubleshooting only to find out a better PSU could have saved them both time and money.


The series circuit example was to keep the explanation as simple as possible. Whether the resistance was 1 ohm or not has nothing to do with it. Nor does “leakage” WTF currents. Your understanding of ohm’s law is non-existent. Please read my post again. I do not want to argue with you because it is pointless. Good day sir.
December 28, 2006 6:33:31 PM

bberson:

i worked at one time with dual +5volt, 500 amp rails on a computer system. the attempted use of a single 1000 amp (buss bar) rail was disasterous! the voltage drop was way too much, reguardless of the buss bar size. i am just going by this and many other experiences when i say dual busses are better. Ohm's law has never been appealed.

if i used a term that offends you, i'm sorry. i just try to make my descriptions precise so everyone will understand my meaning.

would it suprise you to know that YES there is a transformer inside a switching PS? it is a torroidal transformer normally operating in the 40Khz to 200Khz range. it is much lighter in weight, cheaper, and more efficient than the old 50/60Hz transformers. because of the feedback necessary to control the switching transistors (FETS?) on the primary side of the transformer, the frequency can be adjusted to more efficiently control the output, depending on the changing load on the secondary side of the torroidal transformer.

to everyone: have a GREAT DAY!!! bobvc99
December 28, 2006 6:48:54 PM

Quote:
dual busses are better. Ohm's law has never been appealed.

Indeed. Problem though, is that many supplies just split the buss and you still end up with potentially large losses at the split. Multiple regulator circuits would be an improvement, no? But I don't think we can count on that, which is why you see such confusing "2+2=3" -variety mathematics on the specification stickers on these power supplies.

I'm not taking anything personally, just wanted to make sure you didn't think I was a fool.

I'm not that familiar with switching designs. Reading between the lines here, you're rectifying and filtering the AC before the toroid and then using semiconductors to create pulsed DC for the toroid? Throw me a pointer to a schematic so I can give myself a refresher...

-Brad
!