Datacenter consumption calculation

[Fou-Lu]

Hi guys,
I'm needing to calculate some consumption numbers for our new datacenter (Kev, I'm looking you're way :P). So I've been going through and pulling some numbers off of the components in the racks we are planning to move, and where available cross-referencing them with user manuals. Best I can tell, these are all on 240V power connectors.
So, going through some numbers I pull off this from a blade PSU:
4 PSU @ 200V - 240V 13.5A. So, that should equate to 2700 - 3240W (VxA) consumption per PSU correct? There is a number on the back for each PSU indicating 2000W not anywhere near the numbers I'm calculating, so is it consuming 1240W over its output?

Also, I'm noticing that there are ranges for the amperes as well (I'm so clueless when it comes to electronics :o). I'm assuming that these are an indication of the amperes pushed when at the high and low voltages? For example, a FO switch lists:
100V - 240V, 2.5A - 1.3A. Does this provide a high and low watt rating of 250W - 312W (I assume it would be the left most and right most to multiply; 600W from a 240V*2.5A would seem quite high...)
Also, if I'm right at these assumptions, would I also be right in assuming that the blades with multiple PSU's need to be multiplied by this watt usage?

Man I'm terrible with electricity....

[oracleguy]

4 PSU @ 200V - 240V 13.5A. So, that should equate to 2700 - 3240W (VxA) consumption per PSU correct? There is a number on the back for each PSU indicating 2000W not anywhere near the numbers I'm calculating, so is it consuming 1240W over its output?

Depends where you got the numbers exactly (like which parts). It has been 4+ years since I've worked with that stuff since I can't say for certain off the top of my head.

100V - 240V, 2.5A - 1.3A

This means 2.5A @ 100V and 1.3A @ 240V. The reason for this is that at higher voltage it takes less current to do the same thing. (Ohm's law and all)

[Fou-Lu]

Ohm's law, thats what it was. I figured that the corresponding values were correct, so thats good; I figured less voltage is required to carry a higher amperage, and vice-versa. Nice, I like it when I remember the old physics stuff! Glad to know that the corresponding number range worked that way; I've never done range calculations (that I recall), so I wasn't sure.

As for the PSU, it just on the back of them. I figure its the PSU for the blade, there are four in each listing the same numbers. But, there is a unit that feeds it power, the unit itself has what appears to be a 240 connector, and four output cables that could just be 110. Also, I was googling some stuff, looks like there is a rating for the power that is never listed on the plates for the units and rarely listed in the technical specs. So that kinda sucks. Also, mention of phase variance was involved, best I could gather was that a Phase 2 variance caught every second wave in the AC current, so it effected the consumption too.
Another... discrepancy that I ran across. One of the fiber switch models I couldn't get any information off of so I had to IBM it up. I don't have the exact numbers, but I can tell you what I recall: listing 90 - 243V, 56W max power consumption. Interestingly enough it listed a 0.25kVA, which definitely doesn't jive with these (according to my initial calculations). I don't understand how it can have a 56W max power consumption and still list a 0.25kVA? Or maybe the kVA is its output you think?

[oracleguy]

Do you have any spare servers of the same hardware? You could just physically measure the power draw of the unit. Then you could take those numbers, add 10-15% and use it. That honestly would be the easiest way, lol.

That makes sense about the blades but depending on the design, the 240 power supply might not even be putting out AC to the sub power supplies.

Yeah things get messy with the phase angle, that is how the VA and the actual watts can be different, like on UPSes. It is really a pain in the butt truth be told. If you actually bother to calculate anything with phase angles, I recommend a calculator that can do Phasors (My choice being a TI-89).

[Fou-Lu]

Phasors, like the thing that they use in star trek :P
I talked to my boss, sounds like he is less interested in me doing the calculations since we can get one of the electrical engineers to calculate it out, and more interested in getting the numbers off the back (its hard to see them, but I'm a fairly small guy so I can get into small places :P). Since this is just the second day in my new position (yay!), and my boss wasn't exactly aware of my coming onto the team until I mentioned it on friday (I was interviewed by our department head; my boss was on extended vacation at the time). So I suspect that this may just be a way to busy myself until he finds something in specific for me to take care of.

Thanks for you're input mate, very helpful!