Confirming validity of transfer events

[Tfj]

I started monitoring my UPS's about 3 months ago with apcupsd and email alerts.  In this time I have noticed several alerts that the UPS went to battery and came back a second later or less than a second later.

What techniques could I use to validate that my UPS is properly detecting problems with input power?

I have many systems that are not on ups and using the same input power and they are not disrupted. In general power outages here happen once a year or less.

I know that I can adjust the sensitivity on my APC upses. I did not ask how to do this in this thread

I want to know how to validate my UPS's observations of there being a power disruption.  Or how to substantiate my APC UPS's assertion that there ever was a problem with power

So far I have tried placing a camera on the UPS that can also see a light powered by the same circuit the power of the UPS. It records continuously at 30fps. I have experienced two events now where there was no discernible flicker of light during the supposed outage.

So I am actually wondering now if my APC UPS is faulty and detecting failures when there are no failures. Please help me disprove this.

Here's the email i got this morning, in case something sticks out to you...

Subject:. Hhghjn.org UPS Hhghjn.org Power Failure !!!

Hhghjn.org UPS Hhghjn.org Power Failure !!!

APC : 001,038,0999

DATE : 2022-08-13 07:02:28 -0500 

HOSTNAME : Hhghjn.org

VERSION : 3.14.14 (31 May 2016) redhat

UPSNAME : Hhghjn.org

CABLE : USB Cable

DRIVER : USB UPS Driver

UPSMODE : Stand Alone

STARTTIME: 2022-07-05 17:27:29 -0500 

MODEL : Back-UPS XS 1500 LCD

STATUS : ONBATT

LINEV : 124.0 Volts

LOADPCT : 11.0 Percent

BCHARGE : 100.0 Percent

TIMELEFT : 10.8 Minutes

MBATTCHG : 5 Percent

MINTIMEL : 0 Minutes

MAXTIME : 0 Seconds

SENSE : Medium

LOTRANS : 88.0 Volts

HITRANS : 139.0 Volts

ALARMDEL : 30 Seconds

BATTV : 27.2 Volts

LASTXFER : Low line voltage

NUMXFERS : 4

XONBATT : 2022-08-13 07:02:29 -0500 

TONBATT : 1 Seconds

CUMONBATT: 18 Seconds

XOFFBATT : 2022-08-06 17:50:14 -0500 

LASTSTEST: 2022-08-06 17:50:06 -0500 

SELFTEST : NO

STATFLAG : 0x05060010

SERIALNO : 3B0916X6937177 

BATTDATE : 2022-05-25

NOMINV : 120 Volts

NOMBATTV : 24.0 Volts

NOMPOWER : 865 Watts

FIRMWARE : 837.H7 .D USB FW:H7

END APC : 2022-08-13 07:02:30 -0500 

[BillP]

@Tfj 

If you want a scientific analyses you should use an oscilloscope to test the UPS input. If you know an electrician or electrical engineer you could ask to borrow one to run the test. 

[Tfj]

Okay, I'll borrow one a couple days before the next unexpected transfer to battery.

What's the command to see when the next time it will transfer?

[BillP]

@Tfj 

 There is no command to determine when the next power event will be. The oscilloscope will monitor and record the power. If the UPS switches to and from the battery, you will check to see what the oscilloscope recorded for the time. 

[Tfj]

So how would I know when to connect the oscilloscope?

Do you think I should get one of those isolators for measuring line voltage?  Or just wind a conductor around hot and another around neutral and infer voltage via the electrical noise it picks up?

There's a 4 channel that I can use so I might do that for hot, neutral, and ground (why not), and use the 4th to monitor an led that I'd keep in frame of video and blink a timestamps pattern so I can line up video and scope. I've got an Arduino I can use for the ledi guess.

[Tfj]

so i learned that those isolators are called 'differential probes' and they cost like $300 a piece.  Not doing that...

I think I read something about a 'potential divider'.  Basically taking 200 resistors, and connecting the first one between earth and the probe tip.  and clip probe ground to earth.  Then string the other 199 resistors along from the top to hot.

It sounds simple.  I'd expect to see one 200/th of the actual oltage on that channel.  But wouldn't the string of resistors become in efffect a short, or a space heater right quick?

I don't feel like going out and getting a tonne of resistors either.  And what ohm would I use..

Then I got to thinking, that big string of resistors is a lot like the big strings of Christmas lights that I have sitting int he basement anyway.  Each bulb is only rated for 3 or 5 volts.  So perhaps I could take two strands in series to keep that voltage under 2v.  And attach the probe one bulb away from ground.

Actually those would be plugging between hot and neutral normally.  using them between hot and ground would surely trip my gfci outlet.  Though i suppose I could find a circuit that's not GFCI protected

Or I could just attach the probe ground to neutral.  or leave it loose.  The resistance between the scope's ground and neutral should be pretty minimal.

@BillP  Do APC UPSes monitor or react to any potential between ground and neutral?  Or do they just react to the power between hot and neutral? (Easier to measure)