Looking through current prices for these drives I stumbled upon this review on Amazon by a EE who works as a power supply designer.  According to his measurements peak +12V current is ~600mA during spin-up and drops to 240mA for normal write/read/erase.  He also notes that the +5V rail appears to drive the head servo.  He measures peak +5V current to be 500-600mA during spin-up and 400mA during head movement activity.

 

I also found it interesting that he found total power at idle was 3.5-3.7W while the datasheet quotes 2.5W.  The numbers he gives are +12V @ 0.24A and +5V @ 0.13A (he then calls this 4.2W, but even giving him the benefit of rounding error it would only be 3.7W).  It seems like the WD datasheet power values might be reporting only +12V power (12V*0.24A = 2.88W)

Is my reading and 'rithmatic correct that I need 7.2 watts (say 10 if I'm being conservative) for each drive in my array if I want to support simultaneous spinup of the drives?

 

Then again, maybe I should just watch my wall meter the next time I spin up my Unraid server...

You need 7.2W on the 12V rail. There is another 3W needed on the 5V rail. This is still much lower than the 2A or 24W commonly recommended for each green drive.

 

Peter

Holy crap, a quarter of an amp for spinup of a 2TB EARS drive?!?! Hitachi lists 1.5A for their 5K3000 drive on spinup. I wonder if the OP enabled the super low spinup current draw with that utility.

 

If other drives are this low I won' have to shell out for such a beefy PSU.

Holy crap, a quarter of an amp for spinup of a 2TB EARS drive?!?! Hitachi lists 1.5A for their 5K3000 drive on spinup.

That was idle current, not spin-up current.

 

I would still suspect the .6 Amp peak.  You cannot cheat physics.  Either the actual peak is higher, and not captured in his test configuration because it is very short, or, the spin-up time is very long.  Putting the "head seek" current load on the 5Volt supply make it easier on the power supply, since most will have plenty of capacity there.

 

Joe L.

Quote from review;

 

"When the drives are first powered up, +5V draws an almost constant 500-600 mA. +12V current is a ramp, starting at 200 mA and works its way up to 600 mA after about 10 seconds, when the drive is then up to speed."

 

Sure sounds like the peak on the 12V rail is 600mA. Would you consider 10 second spin-up to be long?

 

Peter

 

I'd certainly trade off a fast spinup for a longer spinup that draws much less power. Once all drives are spun up the PSU will likely fall out of efficiency unless it's an 80Plus PSU which is always >80% efficient.

unless it's an 80 plus PSU which is always >80% efficient.

Sorry but that's not quite right.  An 80 Plus PSU must have greater than 80% efficiency at 20%, 50%, and 100% load.  Many 80 plus PSUs provide less than 80% efficiency at less than 20% load.  But your point is valid in that many users who select a large capacity PSU (i.e. >550-600w) to power a small server (<5-6 disks) are going to suffer from poor PSU efficiency when their server is idle/spundown.

Sorry but that's not quite right.  An 80 Plus PSU must have greater than 80% efficiency at 20%, 50%, and 100% load.  Many 80 plus PSUs provide less than 80% efficiency at less than 20% load.  But your point is valid in that many users who select a large capacity PSU (i.e. >550-600w) to power a small server (<5-6 disks) are going to suffer from poor PSU efficiency when their server is idle/spundown.

 

Thanks for bringing that to my attention. I'd be willing to bet that many active PFC PSU's meet these specs, but the manufacturers don't want to spend the money to have them 80Plus certified.

^^You might be right.  I usually don't rely on the mfg specs instead I find a good online review (i.e. slientpcreview, anandtech, etc.) where they tested the PSU under different loads and measured its efficiency.