We have a HP T1500 G3 UPS which we noticed does several short switches to battery quite often during the day, every day. Nothing ever gets reported to its log (i.e. HP Power Protector Notifications), and, what's funnier, the output voltage gets reduced to ~90% of input voltage for ~2 seconds (it's working on buck transformer, since the output signal is not deformed).
This is how the voltage trend looks like (RMS values):
Green: Vin(RMS), Pink: Vout(RMS); Voltage is ~237VAC 50Hz.
Note how the input voltage signal (green) doesn't have any dips or swells, but the UPS reduces the output voltage by more than 10% (while operating on bypass).
This is how the waveform looks like for the start of the disturbance:
Green: Vin(RMS), Pink: Vout(RMS); Blue area shows inverter operation for 2 periods (40ms)
Short dip from the first chart is indicated with the blue area on the waveform chart, it's the point where UPS shortly switches to battery power (inverter) and the output signal (pink) is shortly decoupled from the input. At the same time, input signal (green) is normal.
But after the short switch to the inverter, it can be seen that the UPS switches back to bypass again, but this time the output signal is ~90% of the input signal at bypass, and this last for ~2 seconds, until it briefly switches to battery again and then everything returns to normal.
Servers don't seem to mind this. It also doesn't happen in regular intervals, and happens several times per day.
(Update)
From what we've seen, it seems like the UPS uses a buck/boost auto-transformer to lower or raise the voltage when it thinks it should. It switches shortly to the battery (inverter) while it adjusts the transformer, and then uses the transformer to slightly lower the input voltage (+/- 12.5%, from what I've seen) without the battery.
There are two reasons why this is strange:
Our UPS dip switches are set to 240VAC (manual says this represents the input range of 216–254 VAC), and this happens when voltage just millivolts over 240VAC. I would expect for it to wait until ~254VAC and then reduce the voltage to 240.
There are no entries in the UPS log (well, 1 of 10 of these events is registered in the log).
So, apparently, the device measures around 240 volts (sometimes it's 237, sometimes slightly over 240VAC), and decides to switch to the buck mode. At this moment, the voltage drops down to ~210 VAC (which is rather low), and then it presumably realizes it has done a mistake and returns back to the direct input. After that, it won't switch to buck anymore, although the input voltage stays above 240 volts for quite a while.
Is this considered "normal" operation? Or is it possible that the UPS measuring circuitry is misbehaving?
