PSU usage

Does a PSU with higher power use that higher power all the time? Or
merely on demand?

If two computers were idling, one with small PSU, one with large one,
would they consume the same wattage?

Put another way, should you limit the PSU max. to your MB max?

Ed

On 05/08/2018 7:37 AM, Ed Cryer wrote:
Does a PSU with higher power use that higher power all the time? Or
merely on demand?

If two computers were idling, one with small PSU, one with large one,
would they consume the same wattage?

Put another way, should you limit the PSU max. to your MB max?

Ed


Hi Ed, Yes they both will draw about the same, For instance, even tough
I have a 750 watt PSU in my desktop it only draws 118 watts when at idle.

Rene

Ed Cryer wrote:
Does a PSU with higher power use that higher power all the time? Or
merely on demand?

If two computers were idling, one with small PSU, one with large one,
would they consume the same wattage?

Put another way, should you limit the PSU max. to your MB max?

Ed


A 500W power supply does not draw 500W continuously.

From the wall, it draws just enough to supply the load.

If the CPU draws 100W, the PSU draws 100W from the wall.

That is the "first order approximation" picture.

*******

In the real world, the PSU pays a slight tax. It
still follows thw above rules, but some laws of
physics are still present.

Imagine there is a "supervisor" in the power supply.
It's keeping watch. It needs a bit of power to operate.
Let's say that's a "fixed" power loss. A static figure
of some amount, no matter how much load is on the supply.

Further, let's use Ohms Law. The more current that flows
out of the supply, to the motherboard, the more V=R*I
voltage drop there is, and the more P=V^2/R heat loss
that results.

Using our vivid imagination, the supply is less than
100% efficient. Some of the losses are always present.
Some of the losses are load dependent.

When shopping now, we look for "80+" power supplies.
These are color-coded, silver, gold, titanium, and
so on. We didn't always care about efficiency, as
the older supplies had a fair flow of hot air and
a fan running at high speed. The efficiency has
improved over the years.

*******

This is the latest affectation. A power supply with
Gallium Nitride (GaN) MOSFETs. Already sold out, at
$500 a unit. This is basically a technology demo.

https://www.anandtech.com/show/12645...ed-performance

At idle, where you and I would be using it, it is 88% efficient.
At 800W, it is 96.5% efficient.

The second graph on page 4, converts the efficiency
numbers into "power loss". So now we can look at the
"taxation curve" and see what kind of taxes must be
paid in watts. This is easier for consumers to understand,
than doing the reverse calc to convert eff to watts.

At idle, it dumps about 10W of heat into the air.
At 800W load, it dumps about 27W of heat into the air.

The fan only starts to spin on the power supply,
when the load surpasses 600W. That means in a lot
of situations, the PSU fan will be off. And in the
real world. when the computer case has an exhaust
fan, the small amount of air that funnels through
the supply by accident, is likely to keep the fan off
quite a bit of the time.

So when I say

CPU draws 100W, the PSU draws 100W from the wall

in actual fact, using that review as an example

CPU draws 100W, the PSU draws 110W from the wall

So what numbers might I see in actual usage.

My newest PC:

Running 7ZIP, 650W supply provides ~200W
Idle 650W supply provides ~100W
S3 Sleep 7.5W
S5 Soft-Off 0.5W
Pull plug from wall :-) 0W

My peak power isn't very impressive, since I don't
have a gamer video card. There is at least one
video card out there, that draws 300W max (and
that one works best, with water cooling).

HTH,
Paul

On 05/08/2018 07:37 AM, Ed Cryer wrote:
Does a PSU with higher power use that higher power all the time? Or
merely on demand?

If two computers were idling, one with small PSU, one with large one,
would they consume the same wattage?

Put another way, should you limit the PSU max. to your MB max?

Ed

I often see similar posts, and am aware how little some people know
about electricity. A PSU doesn't "use" that power but supplies it, and
there's no way it can supply it without a high power load present.

"Sam E" wrote in message
news
I often see similar posts, and am aware how little some people know about
electricity. A PSU doesn't "use" that power but supplies it, and there's
no way it can supply it without a high power load present.

On the other hand, it *could* be that higher-powered PSUs are less efficient
than lower-powered ones and therefore consume more power, wasting the
difference between input and output as heat. It's like a car with a larger,
more powerful engine: does it use more fuel than a smaller (possible
higher-revving) engine, if both are driving at a steady speed? I can
understand people wondering whether this may be the case, even if in fact a
larger PSU/engine doesn't use more energy for similar output power
consumption.

Assuming that a "larger" PSU is still about the same efficiency as a
"smaller" one, for an idling PC, there is presumably no disadvantage (other
than cost) in fitting a larger PSU than strictly necessary.

In article ,
says...
GG here with no expletives; That's a first here!!!!!!!!!!!!

Congrats! ( :-) )