How gets power usage measured in SSD tests?

[wolfgang123usa]

I would like to know the test setup in the SSD test to measure power consumption of the SSD drives. Amp and Volt meter in supply lines ?, or resistor in supply line and then measure voltage drop with scope on resistor and voltage on supply line....etc. Would appreciate any details. I assume a Watt meter for PSU on AC input would be very inaccurate......
regards, Wolfgang

[grantmeaname]

For our power consumption tests, we measured the voltage drop across a 0.1-ohm resistor placed in line with the 5V and 12V lines connected to each drive. We were able to calculate the power draw from each voltage rail and add them together for the total power draw of the drive.

They used to put that snippet in every article (the way they copy and paste other things, like their explanation of using FRAPS on the methods page), but they apparently stopped doing so like two years ago. I'm not an electrical engineer, so I don't know how they're measuring the voltage drop or where in the circuit. If the quote doesn't answer your question, you should email it to Jordan Drake (jdrake at techreport dot com), the podcast guy, for the staff to answer.

[Captain Ned]

I'd be surprised to find that the testing method has changed from the resistor-based test.

[grantmeaname]

Sorry. I didn't mean they stopped measuring it that way, I meant that they stopped including that detail in every SSD article.

[Dissonance]

We haven't changed our methods. Power consumption is still tested with the 0.1-ohm resistor. I'll make sure that detail's added to the test methods section of future storage reviews.

[wolfgang123usa]

Thanks for the info!
Wolfgang

[andy88]

Out of interest, what level of accuracy do you expect to get from this test setup? (given resistor and meter tolerance). Is the same fixture used to test all drives, to ensure any error is consistent?

I'm writing a short article about the different methods of power measurement, mostly from the perspective of testing in storage integrators/OEMs but I'm interested in how others do it as well...

Note: We (Quarch Technology) sell a unit designed specifically to measure SSD power consumption (rather than the breakout cables and scope+current probe) that would be needed normally in a test lab, so I must admit a motive behind this!

Thanks,
Andy

[MadManOriginal]

We haven't changed our methods. Power consumption is still tested with the 0.1-ohm resistor. I'll make sure that detail's added to the test methods section of future storage reviews.

You could publish it in every article. Alternatively, a 'methods' section of the website under the 'More...' menu would work too, then just link to the relevant methods section in each article. That would make for less repeated information to skip in each article.

[Dissonance]

Out of interest, what level of accuracy do you expect to get from this test setup? (given resistor and meter tolerance). Is the same fixture used to test all drives, to ensure any error is consistent?

We use the same resistor and meter for all our testing. The results should be reasonably accurate for our purposes, but I don't have a specific number to quote. When we first set this up, I remember looking at the accuracy of the resistor and the meter and thinking they were sufficient for comparative measurements.

[ludi]

Might be time to update the laboratory? Here's a true RMS meter with DC ammeter capability up to 10A and 4% maximum error band, for $40. A half-dozen of these would permit you to simultaneously measure current and voltage on all three positive voltage lines if necessary.

The in-line resistor can give approximate values but the resistance does change slightly with temperature. An actual ammeter will typically use a current transformer for measurements to avoid directly touching the power line.

[andy88]

Ditching the shunt resistor might well help the accuracy but trying to read 'simultaneous' measurements from several multi-meters at the same time (I assume to get power consumption) would be potentially rather inaccurate.

The quarch unit (http://www.quarch.com/products/qtl1455.html) runs multiple identical ADC units from the same clock so all measurement are taken at precisely the same time. This ensures that the power value is very accurate. You would also get a similar result from a decent 2 channel scope with a current probe on one channel.

I guess it really depends on what you are trying to do. A multi-meter will probably be fine for a basic estimate or quick comparisons. To get values accurate enough for qualifying drives to a specification or similar, you'll need a better setup!

Andy