Belkin F6C1500TWRK 1500VA Battery Backup UPS

Load Testing & Battery Runtime:

As mentioned before, the Belkin UPS uses a pair of 12V / 7Ah batteries, this is the size you typically find in the inexpensive home / small office UPSes. My 1500VA APC Smart-UPS uses a pair of 12V / 18Ah batteries, however my 700VA APC Smart-UPS uses the 12V / 7Ah batteries (note the VA rating is less than half the Belkin though). Obviously the larger the battery capacity the longer the runtime, something to consider when making a purchasing decision. Belkin states a runtime of anywhere between 1.5 to 120 minutes... Yeah, a little bit more detail on those figures would of been helpful to the consumer.

First I let the UPS charge up overnight, just to be sure the batteries were at full capacity. For the most basic test I hooked up my spare Linux box to the UPS, powered it on and let it idle. The Bulldog Plus software reported a 10% load, and I proceeded to unplug the UPS from the wall. Within a couple seconds the battery level dropped to 86%, and a couple seconds after that it dropped to 70% capacity! I let it run on battery power for a minute or so and the number seemed to stay around 67-70% capacity, so I plugged the power back in. Once the UPS switched back to line power the batteries reported they were up to 95% capacity! I could already tell this was not going to be good. I shutdown my Linux box and let the batteries charge back up.

To put a constant and reproducible load on the UPS I pulled out the trusty PSU load tester and found an old power supply. I ran approximately a 325VA / 185W load (remember power factor is an issue here), and the UPS reported a 20-21% load. Doing the math, 325VA is 21.6% of 1500VA, so the hardware/software was pretty close. After I pulled the plug from the wall the battery level immediately dropped down to 69%, and within 30 seconds dropped down to 56%. From that point on the percentage seemed to stabilize and would slowly drop over time. When the battery got real low the UPS starts continually beeping, letting the user know they better do something.

Final run time for a constant 20-21% (325VA) load... 22 minutes (rounded to the nearest minute). For comparison, my dual-core Opteron system and 22" Samsung LCD monitor only report a 18% load on the UPS when idle at the desktop.

Once the UPS reached (what it thought was) 0% battery power it turned itself off (killed the power at the sockets). However the monitoring software continued to work (odd, I know). The battery level jumped back to 43% (even though it was still unplugged). When I plugged the power back in, the UPS turned itself back on and the battery level jumped up to 50%! Curious I unplugged the UPS once again and the battery level immediately dropped to 8% and within 30 seconds reached critical and shut off the power again.

However, this UPS is classified for network / workgroup type use, so lets assume we have two computers hooked up. After letting the UPS charge up over night I ran a second test. This time I ran approximately a 600VA / 360W load, and the UPS reported a 40% load. Doing the math, 600VA is 40% of 1500VA, so again the software is correct. I recorded the time and pulled the plug from the wall. Again the battery level dropped considerably in the first few seconds, but then tapered off and dropped more regularly.

Final run time for a constant 40% (600VA) load... 10 minutes, almost to the exact second.

So what load level would allow a 120 minute runtime? By doing some simple math, that would be only a 5% / 75VA load. That's quite a bit less than what the average home computer consumes, not to mention a home user would probably also hook up their DSL modem, maybe a broadband router, and who knows what else. I find that "120 minute backup time" statement on the front of the box a little misleading.