• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Measuring Current Draw

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

markodude

Member
Joined
Jul 15, 2002
Location
Europe
I want to know how many amps my system is using, particularly the CPU - but the main thing is to determine if the dual 14A +12v rails in my PSU are being maxxed out.
I have been looking at Dual Core power consumption - the CPU in my system could chew up 200w or even more when overclocked to 4.xghz - thats nearly 17A if my calculations are correct?
Add 2 volt modded 7800GTs at around 100w each? Then theres the mobo itself and its components, 5 hard drives, 9 fans, etc etc
My dual 14A rails could probably both be maxxed.
How can I work it out - I have a multimeter?
I know basic electrical equations!
Thanks
 
There are logical problems inherent in this undertaking. Sure you can measure the actual current consumption of the CPU. But what do you compare it to? Unless you further measure the actual maximum sustained output of the power supply, there is no valid figure to compare it to.

Output ratings are just that--ratings. They are not measurements of output. At some point numbers poorly quantify the reality, especially when you are lacking such a critical variable in your proposed equation. In truth, about the only real way to ascertain whether a given supply is limiting the OC of a given system is to try one that is known to be more capable and note the result. In all truthfulness a person can simply ascertain the make and model of your supply, the load you are placing on it, and guess if it's limiting you with at least as high of an accuracy level as can be achieved with numerical analysis.

Another problem inherent here is that you don't have any way of knowing at which point the load on the supply can hurt the OC, even if that load be less than the finite maximum. Does a load that consumes 90% of the supply's output test it severly enough to hurt the OC? 80%? 95? There just isn't an established body of knowledge in the area sufficient choose one over the other.

It's all rather like using a dynamometer to assess a racing engine's performance. In the end, you don't race dynos. Those that invest enough time generating enough dyno results are usually at some point able to correlate their dyno figures with the reality of speed, but until that correlation is established the numbers are meaningless.
 
The cheapest way I know of is to cut the power cables of a PSU so you can hook your amp meter in series with each set of wires, but this requires a meter that can read at least 20A, and it's possible that the connections with the meter will have enough resistance to cause the voltages reaching the mobo to be too low. Another way is to solder shunt resistors in series with the wiring and measure the voltages across them. If you build your own shunts, steel wire may be better because it has a lot more resistance than copper does.

The easiest way to measure current is with a clamp-on amp probe, but those for DC cost more than the AC-only ones. The current issues of Silicon Chip (www.siliconchip.com.au) and Everyday Practical Electronics (www.epemag.com) have articles about building one from a UGN3503 Hall effect sensor and a toroid core cut in half.
 
Wow that sounds complicated! But thanks for the inputs both of you.
I did some investigating and came to the following results...

Say my CPU is about 150w at stock given that it is a 3.4ghz, Xbitlabs report the EE is 156w or so.

http://www.benchtest.com/calc.html

According to this:

Default – 150w, 1.268v, 3400mhz

4000Mhz @ 1.3v = 175w = 14.58A – I am running at this speed now prime stable with the case closed and in the normal position under the desk – my PSU is 14A +12v

4250Mhz @ 1.3v = 186w = 15.5A

4500Mhz@ 1.4v = 228w = 19A

5000Mhz @ 1.5v = 291w = 24.25A

So I am going to go for the PCP&C 510 SLI - 34A on +12v single rail. Hopefully the rest of my components wont chew more than 10A (5 HDDs, 7 fans, DVDRW, Mobo, Memory, SLI GTs)
Hmm maybe I will still struggle!
 
markodude said:
So I am going to go for the PCP&C 510 SLI - 34A on +12v single rail. Hopefully the rest of my components wont chew more than 10A (5 HDDs, 7 fans, DVDRW, Mobo, Memory, SLI GTs)
Hmm maybe I will still struggle!
I doubt it. The PCP&C 510 is one of the strongest supplies made, and drives most anything you can throw at it. I think your choice is rational and it is exactly the unit I would have recommended assuming the absolute lowest noise floor isn't the first priority.
 
markodude said:
Hmm maybe I will still struggle!

I think you'll be ok too - I doubt the CPU will go much past 200W myself if it draws even that much. Zippy has a 700W unit with a 45A 12v rail (PSL-6701P) if you're worried about it, but it costs a fair bit more than the PC P&C 510W which should already have no problem powering things at all.

Another alternative though would be the Silverstone Zeus 560W - it's a little cheaper and more powerful than the PC P&C, but doesn't quite regulate as well from early reports. Same level of build quality though.
 
Back