Would a 650W PSU run hotter when powering a 484W computer than when powering
a 232W setup? Obviously case temperatures would affect things, but if we
could isolate that...
Does a PSU run hotter when under more load or is its heat output independent
of the amount of power drawn from it?
The heat from a power supply will increase as the load increases.
GT wrote:
>
> Would a 650W PSU run hotter when powering a 484W computer than when powering
> a 232W setup? Obviously case temperatures would affect things, but if we
> could isolate that...
>
> Does a PSU run hotter when under more load or is its heat output independent
> of the amount of power drawn from it?
"Mike Walsh" <spam_sucks@bellsouth.net> wrote in message
news:48AAD88D.AF17147@bellsouth.net...
>
> The heat from a power supply will increase as the load increases.
That is what I thought. So a significantly 'over-the-top' PSU for a PC would
tend to run cooler (due to larger cooling potential) than a
'pushed-to-the-limit' smaller PSU, which would have its little fan blasting
like mad to keep up.
GT wrote:
> "Mike Walsh" <spam_sucks@bellsouth.net> wrote in message
> news:48AAD88D.AF17147@bellsouth.net...
>> The heat from a power supply will increase as the load increases.
>
> That is what I thought. So a significantly 'over-the-top' PSU for a PC would
> tend to run cooler (due to larger cooling potential) than a
> 'pushed-to-the-limit' smaller PSU, which would have its little fan blasting
> like mad to keep up.
>
>
Take actual_load, and divide by the efficiency.
Say I have a 68% efficiency power supply and I draw 200W,
200W
---- = 294W , and 294W - 200W = 94W of heat
0.68
If I purchase an 85% efficiency power supply and I draw 200W,
200W
---- = 235W , and 235W - 200W = 35W of heat
0.85
In those examples, the actual load power is what determines the
heat. My load in both examples, was 200W at the outputs. If I
draw an actual 650W of power from each supply, then the
68% efficient one is going to get impressively hot. Heat is
proportional to output power. Draw more output power, and
get proportionally more waste heat.
A power supply that doesn't state the efficiency, could be as
low as 68%. The more expensive ones will be proud of their
efficiency, and will state they are "80+" or 85% or whatever.
Some reviews will attempt to measure the efficiency, to see
how close to the stated figure it gets.
In terms of operating temperature, you cannot make any
statement, because the fan can be slowed down to achieve
better noise performance. If I completely stop the fan on
the one dissipating 35W, I might be able to make it as
hot as the 94W one with the hurricane fan on it. But generally
speaking, you know the 85% efficient one is not going to be
allowed to get that hot. They'll leave the fan running to
some extent.
The efficiency varies a bit, with the load range. The
85% unit, might be 85% at a load of 200W to 300W. Perhaps
the efficiency drops to 80% at 100W. But it is likely
ahead of the unit that is 68% at 100W, so you're still
"winning".
The only time the 85% one is not a winner, is if the
design is marginal. For example, some early Seasonics,
had weak 3.3V or 5V rails, with excessive droop. On modern
motherboards, that would tend not to be an issue, as the
3.3V and 5V rails don't draw as much current as they used to.
But if you had a dual Athlon S462 motherboard, with all the
processor power coming from the +5V rail, and you combined
that motherboard with an early generation Seasonic, then
you may regret your purchase.
The best way to spot problems, is read the reviews on Newegg,
even if you aren't going to buy the supply from there.
On Tue, 19 Aug 2008 17:12:56 +0100, "GT"
<ContactGT_rem_ove_this_@hotmail.com> wrote:
>"Mike Walsh" <spam_sucks@bellsouth.net> wrote in message
>news:48AAD88D.AF17147@bellsouth.net...
>>
>> The heat from a power supply will increase as the load increases.
>
>That is what I thought. So a significantly 'over-the-top' PSU for a PC would
>tend to run cooler (due to larger cooling potential) than a
>'pushed-to-the-limit' smaller PSU, which would have its little fan blasting
>like mad to keep up.
>
In addition to what Paul wrote, the difference between a
truely higher capacity PSU and a lower one is that some (in
theory, all) of the heat generating parts are more robust
and able to shed their heat through larger volume (lower
power density) and/or better heatsinking. Therefore, a
better designed higher end PSU could in theory be producing
the same amount of heat as the lower end PSU, but the
individual parts susceptible to heat damage are staying at a
lower peak temp.
However, some PSU these days have very good efficiency
ratings above some threshold like 80%, but if the PSU is
very oversized for the system then that PSU may never reach
that efficiency rating, always operating below it's design
target because the system didn't put enough of a load on it.
Even so, it is still a better margin to have than one
gaining a small percentage on efficiency but being shorter
lived by running closer to it's design limits... at least
that's my opinion, I'm sure some tree huggers will beg to
differ but then they don't want to talk about the waste of
resources or power to produce a 2nd PSU or parts to replace
one that failed prematurely.
> Would a 650W PSU run hotter when powering a 484W computer than when powering
> a 232W setup? Obviously case temperatures would affect things, but if we
> could isolate that...
>
> Does a PSU run hotter when under more load or is its heat output independent
> of the amount of power drawn from it?
You cannot consume more power without generating more heat. More power
means more heat.
Somewhere on teh intarweb "VanguardLH" typed:
> GT wrote:
>
>> Would a 650W PSU run hotter when powering a 484W computer than when
>> powering a 232W setup? Obviously case temperatures would affect
>> things, but if we could isolate that...
>>
>> Does a PSU run hotter when under more load or is its heat output
>> independent of the amount of power drawn from it?
>
> You cannot consume more power without generating more heat. More
> power means more heat.
>
> 1000 Watts = 0.948 BTU
Indeed. However, the PSU doesn't 'consume' the power, it supplies (most of)
it to other devices which consume it.
The question was really about PSU efficiency at different loads. The answer
(near enough) is that a certain amount of the output is lost as heat in the
switching process. A good PSU is rated at 80+ efficiency meaning that, if
it's putting 80W out it's using less than 100W with the difference being
lost as heat from the PSU.
Therefore, assuming that efficiency is linear (it's not but it's often
close) a 650W 80% effecient PSU would produce 121W of heat producing the
first load mentioned, 484W, and 58W of heat when delivering the power for
the 232W 'setup'.
Good link:
"Everything You Need to Know About Power Supplies"
DISCLAIMER: If you find a posting or message from me
offensive, inappropriate, or disruptive, please ignore it.
If you don't know how to ignore a posting, complain to
me and I will be only too happy to demonstrate... ;-)
> Somewhere on teh intarweb "VanguardLH" typed:
>> GT wrote:
>>
>>> Would a 650W PSU run hotter when powering a 484W computer than when
>>> powering a 232W setup? Obviously case temperatures would affect
>>> things, but if we could isolate that...
>>>
>>> Does a PSU run hotter when under more load or is its heat output
>>> independent of the amount of power drawn from it?
>>
>> You cannot consume more power without generating more heat. More
>> power means more heat.
>>
>> 1000 Watts = 0.948 BTU
>
> Indeed. However, the PSU doesn't 'consume' the power, it supplies (most of)
> it to other devices which consume it.
>
> The question was really about PSU efficiency at different loads.
True, but since efficiency is never 100% then the more power consumed by
the target devices (mobo, memory, cards, drives, etc.) means the more
power also consumed by the PSU. For a 600W PSU, if efficiency were 70%
(a measure in difference between input and output power at some load), a
load of 300W means the PSU is consuming 75W (300/.8 - 300, or 375W input
power minus the 300W load). For a 400W load, the PSU eats 100W (
400W/.8 - 400W, or 500W input power minus the 400W load). As the PSU
consumes more power under a higher load, the more heat is produced by
the PSU itself. Of course, assuming 80% efficiency is being generous
since, as you said, that's typical of "good" power supplies.
So back to the OP's question of will a PSU run hotter with a 484W load
than with a 232W load, and whether efficiency were 70% or 80% for the
PSU, my answer still applies. More power consumed means more heat. No
matter what PSU the OP uses and irrelevant of the PSU's maximum capacity
(i.e., max load), you cannot *consume* more power without generating
more heat because. Efficiency is never 100% for the PSU.
Somewhere on teh intarweb "VanguardLH" typed:
> ~misfit~ wrote:
>
>> Somewhere on teh intarweb "VanguardLH" typed:
>>> GT wrote:
>>>
>>>> Would a 650W PSU run hotter when powering a 484W computer than when
>>>> powering a 232W setup? Obviously case temperatures would affect
>>>> things, but if we could isolate that...
>>>>
>>>> Does a PSU run hotter when under more load or is its heat output
>>>> independent of the amount of power drawn from it?
>>>
>>> You cannot consume more power without generating more heat. More
>>> power means more heat.
>>>
>>> 1000 Watts = 0.948 BTU
>>
>> Indeed. However, the PSU doesn't 'consume' the power, it supplies
>> (most of) it to other devices which consume it.
>>
>> The question was really about PSU efficiency at different loads.
>
> True, but since efficiency is never 100% then the more power consumed
> by the target devices (mobo, memory, cards, drives, etc.) means the
> more power also consumed by the PSU. For a 600W PSU, if efficiency
> were 70% (a measure in difference between input and output power at
> some load), a load of 300W means the PSU is consuming 75W (300/.8 -
> 300, or 375W input power minus the 300W load). For a 400W load, the
> PSU eats 100W ( 400W/.8 - 400W, or 500W input power minus the 400W
> load). As the PSU consumes more power under a higher load, the more
> heat is produced by the PSU itself. Of course, assuming 80%
> efficiency is being generous since, as you said, that's typical of
> "good" power supplies.
>
> So back to the OP's question of will a PSU run hotter with a 484W load
> than with a 232W load, and whether efficiency were 70% or 80% for the
> PSU, my answer still applies. More power consumed means more heat.
> No matter what PSU the OP uses and irrelevant of the PSU's maximum
> capacity (i.e., max load), you cannot *consume* more power without
> generating more heat because. Efficiency is never 100% for the PSU.
All true (except the 70% typo, that threw me for a while <g>).
I was simply elucidating to answer the OP's second question; "is its heat
output independent of the amount of power _drawn_ from it?" You answered
with a reply talking about *consuming* power, I just wanted to cover
*supplying* power.
Cheers,
--
Shaun.
DISCLAIMER: If you find a posting or message from me
offensive, inappropriate, or disruptive, please ignore it.
If you don't know how to ignore a posting, complain to
me and I will be only too happy to demonstrate... ;-)
> I was simply elucidating to answer the OP's second question; "is its heat
> output independent of the amount of power _drawn_ from it?" You answered
> with a reply talking about *consuming* power, I just wanted to cover
> *supplying* power.
>
> Cheers,
I must be too muzzy-eyed this morn. I still see no independence. The
more power _drawn_, the more power *consumed* by the PSU. At 80%
efficiency, a 100W load (drawn) would have the PSU consume 25W, and a
200W load (drawn) would have the PSU consume 50W, and a 300W load would
have the PSU consume 75W. As the load (drawn power) increases, so does
the power consumed by the PSU. No independence. More power consumed
(by the PSU), the more heat generated (by the PSU).
Somewhere on teh intarweb "VanguardLH" typed:
> ~misfit~ wrote:
>
>> I was simply elucidating to answer the OP's second question; "is its
>> heat output independent of the amount of power _drawn_ from it?" You
>> answered with a reply talking about *consuming* power, I just wanted
>> to cover *supplying* power.
>>
>> Cheers,
>
> I must be too muzzy-eyed this morn. I still see no independence. The
> more power _drawn_, the more power *consumed* by the PSU. At 80%
> efficiency, a 100W load (drawn) would have the PSU consume 25W, and a
> 200W load (drawn) would have the PSU consume 50W, and a 300W load
> would have the PSU consume 75W. As the load (drawn power) increases,
> so does the power consumed by the PSU. No independence. More power
> consumed (by the PSU), the more heat generated (by the PSU).
No, you're right, there *is* no independance. However, the OP asked about
the relationship power *drawn* from the PSU and heat and you went on to talk
about power *consumed*. You didn't initially go into the fact that they are
linear blah blah blah...
Simply put, you didn't answer the question.
:-)
--
Shaun.
DISCLAIMER: If you find a posting or message from me
offensive, inappropriate, or disruptive, please ignore it.
If you don't know how to ignore a posting, complain to
me and I will be only too happy to demonstrate... ;-)
PSU temperatures 
Linear Mode
