ACTUAL UPS overhead (Watts) of different Smart-UPSes?

[Anonymous user]

I have calculated the power loss of some UPS-models from the official efficiency charts (charts are not available for all old SUA-models) and its clear that the overhead/loss in the UPS is constant over the load range. I compared the results with the specified "Online thermal dissipation" in BTU/hr (converted to W). These numbers should match, but they don't. Why?

Source of all data is for 230V models available in Sweden: [http://www.apc.com/products/family/index.cfm?id=165]

Column M is Watts directly calculated from stated BTU/hr and column N is W calculated from the efficiency charts (done in column S-AI) assuming "100% load" is 100% of the specified maximum Watts. Column P is assuming that "100% load" is max VA (which should not be possible with a resistive load).

Only SMT750RMI2U and SMT1000I are somewhat near/ok.
I have measured SUA2200RMI2U and it consumes only 40W despite a BTU/hr rating equal to 88W so perhaps the BTU/hr is overrated for a lot of models.

Message was edited by: marin849

Message was edited me: Changed topic to reflect the answer.

[BillP]

i can't say i know one way or another here what the accuracy is but a few thoughts that come to mind are:

# did you factor in online versus on battery operating modes?
# what about the battery being fully depleted and the UPS charging it with full charging power?

[Anonymous user]

1. The efficiency charts are measured during "normal operation" which I interpret as "fully charged batteries and connected to mains power". Otherwise the charts would make no sense. The BTU/hr rating (thermal dissipation during operation) is to low to be valid for the UPS thermal dissipation under full load when on batteries so I assumed that it is valid for the same condition as the efficiency chart: fully charged batteries and connected to mains input.

2. See 1. 🙂

[Anonymous user]

I have results for the SUA1000I.
UPS overhead ranges from 16.6 W (on, no NMC, no load) to 26-30W (with NMC, full load). The rating of 100 BTU/hr is therefore correct for this unit (worst case in standby-mode with full batteries). The indicated load percentage is 8% off (0 percent is equal to 50 W, 100% is 670 W, it shows 0% up to a bit below 100W).

I still would like to know why the successor (with green technology) SMT1000I also has a rating of 100 BTU/hr (29W) while its efficiency chart gives that the overhead is 8.9 W (idle) to 9.6W (full load). My guess is that the efficiency chart is valid for full batteries and no AVR trim/boost. And that the BTU/hr rating includes AVR trim/boost being used so that the transformer is active and generating heat (SMT models can disconnect the transformer to save power - thats where the power saving comes from).

The charts are valid for SUA1000I with AP9617 installed. AP9617 adds 2.5-3.0 W.

[BillP]

sorry for the delay.

I checked with one of our engineers and here is his response.

Thermal dissipation data is provided for the sole purpose of ensuring that adequate cooling is provided to the UPS. Thermal dissipation numbers are often worst case values (full load, maximum charging power, AVR transformers in use, etc.) and are not representative of typical usage conditions. Efficiency data on the other hand, is representative of steady state use and assumes nominal conditions as described in the notes which accompany the curves. If the real desire here is to compare our newest products to those that we are obsoleting, or to understand how UPS losses vary as a function of load, then the UPS Efficiency Comparison Calculator (found at http://www.apc.com/tool/?tt=5 ) should be used.

More specifically, by selecting the UPSs you want to compare, and setting the electricity cost to 0.1142, and clearing currency symbol, the electricity cost curve in the results pane becomes a loss vs. load curve as shown in the following attachment.

[Anonymous user]

Nice, then my conclusion in the previous post was correct. 🙂
Thanks for the answer. That tool also looks very useable, but it doesn't have data for SUA1000I (makes sense, since there is no efficiency chart available either).

[Anonymous user · Accepted solution]

I now also have a SUA750XLI and I did the same measurements on it (it also lacks an efficiency chart on the APC-website). Efficiency seems identical to SUA1000I.