AGP 6800GS really drawing 45amps???

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
Just watching the hardware monitor in RivaTuner and while running NFS MW the
Voltage Regulator Current jumped up to over 45amps!!!

Is this card REALLY drawing 45amps?

Re: AGP 6800GS really drawing 45amps???

Quoted text here. Click to load it

Depends on where the measurement is being done.

For example, I was working out someone's processor current tonight.

I took an Intel processor rated at 103W.

Measuring the current as it comes from the power supply, it would be:

(103W/12V)*(1/0.90) = 9.5A from the 12V supply (90% efficient Vcore).

That is the current measured feeding into the Vcore circuit.

On the output side of Vcore, the processor VID might be 1.3 volts.
We know the power is 103W.

103W/1.3V = 79 amps flows into the processor pins.

That is the current flowing from the MOSFET area of
the motherboard, into the processor.

The power is a constant. If the voltage at a point in the circuit
is low, then the amperage must be high. If the voltage is high,
then the amperage must be low. Thus, in my Vcore example above,
we have 12V @ 9.5A in one part of the circuit, and 1.3 @ 79A
in another part of the circuit.

On a video card, the GPU can be fed from +12V. There is a switching
regulator on the video card, just like the motherboard. The GPU
core uses a low voltage. It is quite possible for the current
feeding the GPU, to have a high value. Some GPUs are running
around 60W+, but I don't know what voltage value they use for
their cores.

The 45 amp value, suggests a current is being measured in a
low voltage area of circuitry.

So, I don't know what part of the circuit your software tool
is looking at, but a numerical value of 45 amps is not impossible.
Consider a 130W processor running at 1.3V - that would be around
100 amps.


Re: AGP 6800GS really drawing 45amps???

On Tue, 07 Mar 2006 00:02:24 GMT, "Noozer"

Quoted text here. Click to load it

Not drawing, but possibly producing.  It has a step-down
voltage regulation circuit on it, that might be (guessing,
since I don't have the data for this specific
implementation) about 88% efficient so at a peak of 45A out
and 1.5V vGPU, that could mean around 6.5A 12V (IF it were
all drawn from 12V rail which I don't know).

Or, the software reading could just be wrong... I doubt they
used high precision parts (more expensive) for this so it's
subject to some error, perhaps a lot of error with low
voltages used.  If you can find data on the current
consumption of the card (like via Google), that might help
determine if the 45A is even theoretically possible as an
onboard state.

Site Timeline