SMT Range UPS Battery discharge % behaviour

The SMT range has a quirk which results in many questions like the below. Its worthwhile making the information + explanation web-visible.

Dear Seller!

The UPS arrived yesterday, and I tested it with 100% charged batteries at 21% load, and the charge indicator dropped below 50% within a few seconds. After charging the batteries, I would have done the test again today, but there was no need to do so, because after turning on the "selftest" and completing the startup, the 100% charge immediately dropped to 57% (by the time I took the picture, it had increased to 58%). I think this is not right. What do you think about this?

This is due to the UPS' internal battery constant being inaccurate.
Though the UPS' batteries have been replaced and are new, the internal battery constant is still set for the old cells.

For prior generations we could do an engineer reset via serial to manually change the battery constant to a new value (different for each model of UPS). For the SMT and later generations this is no longer possible. The LCD screen system is using the serial bus to poll for its values, and we cannot interrupt that.

The constant will trim upwards on the fortnightly UPS self test.

APC have a procedure that they say also resets the internal constant. We have found this works only 2 in 5 cases (and this is when we have 10 identical UPS on a bench all doing the same reset). The process is to fully charge the UPS, apply a ~70% constant load, pull the mains power and let it discharge fully, including waiting 10 minutes after power to the load is lost. Even after doing that we've found the UPS still have poor runtime estimations, with large variations between identical units, after identical reset process.

The UPS will still run for the same duration, they will provide power for as long as the batteries have chemical energy.
This is in spite of the UPS battery % estimation dropping from 100% to 50% within seconds of the mains power being removed. They often spend a long time sat at 10%.

Our general advise is that you do not need to do anything, let the UPS do its fortnightly self tests, self update its internal battery constant and trim towards more accurate (or perhaps less inacurate) battery % and runtime predictions.

If you are concerned to perform a full runtime test, with a test load, to power off and compare this to expected runtimes. This is the proof in the pudding method, and often casts a shadow on the UPS's estimates.

There was something to be said for the old sytem of 5 LED. Still not perfect, but closer to accurate than the LCD sytem %. The system simply is not accurate enough to warrant 100 degrees of graduation.

NB-23 Member Warrington

NB-23 Member Warrington

Report Post

Write your reason