Measuring power usage of computer

[raden]

I'm not sure that this will work

If the voltage goes outside the design limits, the PSU will go into foldback and cut out.

i.e. all you can do is keep increasing the load until the o/p voltage suddenly drops to zero and measure the resistance at the foldback point From the resistance, you can deduce the current in the knowledge that the voltage won't vary by more than a small amount over the linear range

You could always measure the current using a Hall Effect sensor, but unless you know somewhere that's got one, they're not cheap to buy and the above will give you just as accurate a result

[Ian Stirling]

Err, no. You'r using the wire as a current shunt, and measuring the voltage drop across it, which is proportional to current.

[raden]

But if you're just using a load at the termination, then what are you measuring? You know the nominal output voltage, if you used a normal resistive load (not a halogen bulb whose resistance changes with temperature)

All you can prove is that the PSU will deliver the current or not.

The PSU will deliver a constant voltage up to it's current foldback limit (not strictly true as the total power output of the combined supplies might be the limiting factor (by e.g. saturation of the transformer)

... or am I missing something obvious here?

[Andrew Gabriel]

If you are referring to the Brennenstuhl device, it's off by about a factor of 3 for switched mode power supplies like those in PCs and compact fluorescent lamps. I suspect it's only notion of power factor is that which is due to phase shift, which doesn't apply to switched mode power supplies.

[Ian Stirling]

You'r measuring the resistance of the cable using a load which draws a known current.

You plug in the lead to the PC, and measure the voltage in use.

[Ian Stirling]

Some of them get some switched mode power suppies right.

However, many harmonic corrected supplies have very odd current waveshapes. For the switchers that simply have a diode bridge feeding a big cap, I found they work OK.

For the more modern switchers, you'd need a bandwidth of better than 3Khz to measure some of them, the waveshapes are strange.

[raden]

In message , Ian Stirling writes

Which is going to achieve what exactly ?

It will deliver the required voltage into a load up to the point where foldback occurs and the PSU shuts down

See above

The voltage doesn't decrease in proportion to the load, it remains constant (hopefully within the spec) until the current limit is reached then, the voltage collapses.

If there is a short, then the PSU shuts down preventing damage to the system

again, am I missing something here?

[Andrew Gabriel]

Oh it's certainly possible to do -- I made one some ~20 years ago which is still going fine. I've not seen an accurate true power meter which will handle this type of load in the However, many harmonic corrected supplies have very odd current waveshapes.

That's the type this one gets wrong by a factor of 3.

[Ian Stirling]

It lets you know the resistance of the cable.

You plug in the lead to the PC, and measure the voltage drop along the cable, with the probes inserted into the cable. You then use the resistance you have previously found by using a known current draw, in order to calculate the actual current flowing.

It does not rely on the impedance of the power supply, or anything. It's just a means of measuring the current in the wire directly.

[Dorothy Bradbury]

That name rings a bell, I'm sure that is the one sold by Maplin? o If so, they aren't very good - ok on some SMPS, useless on others

---- eg, it rated a 280W-loaded (Fluke) IBM SMPS at 65W (a bit out :-) o Fluke do some good ones, you can hire meters ok

---- not that expensive altho perhaps not appropriate here o Never tried the cheap clamp-meters on a L cable

You can get plug in adapters which run to a meter with a dial. They used to be advertised in a few places - like utility co dials. Hopefully someone can give them their proper name :-)

[Pete C]

Hi,

Is this the sort of thing you want?

I reckon you should be OK unless you use a high power CPU or fan cooled video card, try asking on the forums at

for more info.

If not you could always fit a quiet ATX psu into an external box and use that to supply the PC.

cheers, Pete.

[raden]

In message , Ian Stirling writes

Yes, we're talking about different things aren't we.

Being a bit thick there wasn't I

[IMM]

Maxie, of course you were. Don't you feel ashamed?

[Ian Stirling]

We all have days when we feel a bit imm.

[raden]

Or DIMM - pref. the bit just around the jugular

[IMM]

Maxie you are due for another R&R. This Dimm Lin the far eastern luscious lovely is going to your head.

[raden]

In message , IMM writes

Your fantasies are running out of control again - take control and grow up