How to measure power consumption on a SSD

[defuse]

Hi, I want to run a test to verify how much power my SSD is using during different scenarios, does such a program exist?

 

[Oldie]

There's no program that could read something like that. You would need a hardware meter.

 

[sub.mesa]

Best would be a direct DC meter; i've seen some meters that are like almost a full circle and the wires of the device you'd like to test go through the circle; anything that goes through is measured. Don't know what the exact name is for such a device, but i'm sure it will serve your purpose.

Generally SSD is 0.1-0.5W idle and 1-2W when stressed with I/O.

 

[ToddW2]

That's a multi-meter. And you need to grab 1 wire at a time, so you would have to cut-up your cable going to the SSD.

They require so little power I think that is one item you could leave alone for power consumption

To get exact measurements due to power level you would need to actually connect/prong the wires, and use a quality multi-meter... a cheapy won't be too accurate at those levels imho.

 

[defaultluser]

Best would be a direct DC meter; i've seen some meters that are like almost a full circle and the wires of the device you'd like to test go through the circle; anything that goes through is measured. Don't know what the exact name is for such a device, but i'm sure it will serve your purpose.

Generally SSD is 0.1-0.5W idle and 1-2W when stressed with I/O.

Those are current loops (hall effect), and they would be a bad choice for this test, as they typically can't measure currents accurately below 200mA.

My testing from my own 60GB OCZ Agility on my multimeter agrees with your general power numbers: half a watt at idle, and 2.5w at full load. IF you add more flash chips, or more complex controllers, the power will go up.

 

[sub.mesa]

Hmm i would have expected that a DC meter would be more accurate than a wall-socket AC meter, especially since that would include your whole system and fluctuates alot; so makes it very hard to measure power used only by the SSD. What would be a good measuring device for his purpose?

There are some nice power consumption comparisons of SSDs on the web; the only real spikes are SSDs like OCZ Apex which uses several JMicron controllers in RAID0. High power consumption and still piss poor performance.

Some reviews disagree on the idle power usage of SSDs. Some quote the Intel as 0.6W idle others as 0.1W.

 

[defaultluser]

Hmm i would have expected that a DC meter would be more accurate than a wall-socket AC meter, especially since that would include your whole system and fluctuates alot; so makes it very hard to measure power used only by the SSD. What would be a good measuring device for his purpose?

There are some nice power consumption comparisons of SSDs on the web; the only real spikes are SSDs like OCZ Apex which uses several JMicron controllers in RAID0. High power consumption and still piss poor performance.

Some reviews disagree on the idle power usage of SSDs. Some quote the Intel as 0.6W idle others as 0.1W.

Well, a DC multimeter is exactly what you need, you just don't want one that uses a current clamp. That is what you described in this part of your post:

i've seen some meters that are like almost a full circle and the wires of the device you'd like to test go through the circle; anything that goes through is measured. Don't know what the exact name is for such a device, but i'm sure it will serve your purpose.

A regular multimeter just measures the current by placing itself in-series in the circuit, and thus does not rely on the hall effect.

Current clamp: immensely convenient, but bad at measuring smaller currents (unless you spend the big bucks).
series ammeter: less convenient (you have to be able to break your circuit and insert the ammeter), but more accurate at lower currents.

 

[Metaluna]

Current clamp: immensely convenient, but bad at measuring smaller currents (unless you spend the big bucks).
series ammeter: less convenient (you have to be able to break your circuit and insert the ammeter), but more accurate at lower currents.

Another problem with either approach is that most lower end multimeters aren't going to be very good at recording the effect of fast current transients, though I don't know how "bursty" the power demands are on an SSD. If the power draw is relatively steady this might not matter much.