Lars-Erik Østerud wrote:
Anyoneout there have a clue about calculation power consumption?
Before I buy a new PSU I had a look at my old one.
It's a Enermax 350W and the label states these specs:
Enermax EG365P-VE
+3.3V +5V +12V -5V -12V +5VSb +3.3&5V Total
32A 32A 26A 1A 1A 2,2A 185W 350W
105W 160W 312W 5W 12W 11W 185W 350W
The last line is my calculated Watts for the lines,
strange that the sum of the individual lines is much
higher than the total (it's not on new PSUs). Is this
something fishy, or what does that really tell you?
As you see it has only one 12V line, and this today powers:
- Asus P4PE mainboard (with Promise S-ATA raid ctrlr)
- Intel P4 2.54 Ghz (478 socket) with stock fan
- Creative SB Live 1024 PCI sound card
- Plextor PX-716A DVD-writer
- Samsung Spinpoint 120 250GB S-ATA harddisk
- Asus V8440 Nvidia Ti-4200 video card
Now, if I replace the video card with a 7600GS
will this work, or will my PSU give up on me.
How much Amps will the other stuff in my rig use?
It's the CPU that takes most juice here, right?
Is there any tables or any experiences that would
help me calculate this? Or do I need a new PSU?
Pentium 4 Processor 2.53 GHz 61.5W
http://processorfinder.intel.com/det...px?sSpec=SL6PD
From 12V rail, this is (61.5W/12V) * (1/0.90) = 5.7A at 90$ efficiency.
DVD-writer 12V @ 1.5A max
Hard drive 12V @ 0.6A idle
7600GS 27.4W, 27.4W / 12V = 2.28A
http://www.xbitlabs.com/articles/vid...r-noise_6.html
Cooling fans - 0.5A estimated
Total 12V == (5.7A + 1.5A + 0.6A + 2.28A + 0.5A) = 10.58A
The 3.3V and 5V consumption cannot be calculated exactly, but
allocating 50W should be enough. Total power for system approx:
(10.58 * 12) + 50W = 177W, and that power may arise when gaming
with a "key disk" inserted in the optical drive.
Neither the 12V consumption, nor the total power limit, are
being exceeded.
Power supplies are constrained by two limits (at least as
seen on their labels). The entire power supply has a thermal
limit, such that, if all outputs go to the max, the inside of
the chassis would get too hot. The primary circuit driving the
transformer, presumably has some limits. Those help define the
total power limit.
Any individual output, stressed to the maximum by itself, has
a higher limit, than when all are being stressed collectively.
If the neighbouring circuits are cool, one individual output can
give more output, as thermally conditions allow it.
Some outputs can share a transformer. You may see coupling between
output limits in that case (like if 3.3V and 5V shared resources).
Since outputs may share a transformer, the output voltages are
established by the turns ratio of the transformer, for each output.
The outputs are rectified after the transformer. When one output is
loaded, in a case where outputs share a common primary, it is not
possible to accurately regulate all outputs at the same time.
And that is why the outputs will use the full range of variation
allowed by the ATX spec. Only a few power supplies have completely
independent circuits and regulation, as it costs more to make.
Xbitlabs plots cross regulation for power supplies.
On this one, the 3.3V shows colored bands, meaning when the supply
is at full load, the 3.3V cannot keep up:
http://www.xbitlabs.com/articles/oth...x-psu5_11.html
On this one, the 3.3V is tightly regulated, while the other two
outputs "flap in the breeze".
http://www.xbitlabs.com/articles/oth...atx-psu_6.html
This supply design was set so that the 5V always had a high output.
Which is why the 5V plot is mostly red. The Xbitlabs cross
regulation plots show plus/minus variation, and either a
+5% error, or a -5% error, would give a red color in the
plot. So these plots don't show whether the 5.0V was 5.25V or
4.75V. Merely that the output has hit the 5% tolerance mark.
http://www.xbitlabs.com/articles/oth...tx-psu_13.html
The Xbitlabs plots consist of three images. One image shows the
3.3V output. The second is 5V. The third is 12V. The legend in
the upper right hand corner, labels which rail is being viewed.
The X and Y axis, show the level of power being extracted from
the rails. The "combined 3.3V and 5V" on the Y axis, is an
assumption by their reviewer, that the supply actually shares
resources for 3.3V and 5V, which is not always true. So
there are limits to the amount of intelligence contained in
their compact representation of regulation performance. Still,
having these plots is an excellent way to display how the
supply functions. It beats the hell out of web site reviews,
where all they do is compliment the manufacturer, on the color
of LEDs used inside the power supply :-)
Paul