APC UPS switches to battery 50 times a day

Submitted ⁨⁨1⁩ ⁨year⁩ ago⁩ by ⁨dosse91@lemmy.trippy.pizza⁩ to ⁨selfhosted@lemmy.world⁩

Hopefully this is the right place to ask.

I have an APC Back-UPS XS 1400U that I use to keep my home server running 24/7.

It was purchased in 2015, batteries replaced around 2020, everything was fine until around June 2023 when it started randomly switching to battery for a few seconds for no apparent reason once or twice a day.

The UPS is connected to my home server via USB so I can get some readouts. It says “Unacceptable line voltage changes”, but it’s configured to switch when it’s outside the 160-280v range and it gets nowhere near those thresholds, the voltage fluctuates in the 224-234 range.

I connected an oscilloscope to the mains to see if there were transients when the problem occurred but I don’t see anything out of the ordinary and the problem has been getting worse, now it switches an average of 50 times a day.

The UPS still works, it can keep the server up for hours if I unplug the power, so the batteries should be good. What’s going on?

source

Comments

Sort:hotnew top

glimse@lemmy.world ⁨1⁩ ⁨year⁩ ago
Oh man I went through this a few years ago…though I doubt it’s the same problem.

I was a custom integrator and one of my clients had a beeping UPS. Unstable incoming voltage. Ok, not the first one of these to kick the bucket so I do the RMA.

Later that same week I’m at a different client for the same thing. Call my boss and tell him we should start looking at other manufacturers, he agrees.

RMA arrived and I install it. Later that day the client calls…the beeping is back. I call my boss and he’s at yet another house dealing with the same thing.

“Wait…you’re at [client]'s? That’s a few blocks away. And so are [other 2 clients]. Are we sure it’s the UPS…?”

TURNS OUT IT WASN’T! There was something fucky with the neighborhood grid (the transformer? node? I don’t know, I’m not an electrician). We talked to ComEd, they and a few weeks later they did some repairs. No more problems.

source
- UnlimitedRumination@sh.itjust.works ⁨1⁩ ⁨year⁩ ago
  How did you identify the problem and get ComEd to look into it? My UPS is switching inconsistently, sometimes a couple times a day, and I’ve seen some signs that makes me think it’s not just that UPS/outlet/breaker. I wasn’t at the point where I was going to do anything yet but since you’re talking about the same provider I have I figured I’d ask.
  
  source
  - glimse@lemmy.world ⁨1⁩ ⁨year⁩ ago
    We didn’t actually identify the problem at all! My boss had the electrical knowledge and couldn’t find anything wrong with the line. But as for contacting ComEd, I never had to call them myself but I’m sure he has some special industry phone number he can call for that sort of stuff. Or at least knew someone who did, he has tons of friends in every trade
    
    source
Rivalarrival@lemmy.today ⁨1⁩ ⁨year⁩ ago

and it gets nowhere near those thresholds, the voltage fluctuates in the 224-234 range.

I’m assuming you measured hot to neutral, and didn’t find any transients.

Did you put your oscilloscope on hot to ground as well? A faulty ground could cause what you are seeing.

Do you have any CFL bulbs anywhere in your house? I once had a dead CFL bulb in a fixture in my basement. Anytime I turned on my basement light, every LED lamp in the house would strobe.

source
- dosse91@lemmy.trippy.pizza ⁨1⁩ ⁨year⁩ ago
  I tested live to ground, live to neutral, both in and out of the UPS, and I couldn’t see anything out of the ordinary.
  
  The input signal isn’t a clean sine wave but it’s not dirty either, I’d say it’s sine with some extra harmonics and a little bit of noise. There were no sudden peaks or meaningful variations even while the UPS was switching furiously.
  
  I don’t have CFL bulbs, only LEDs. I can’t think of anything else that could be causing interference. I’ll try contacting the power company as suggested by @glimse@lemmy.world
  
  Thanks for all the replies.
  
  source
  - Rivalarrival@lemmy.today ⁨1⁩ ⁨year⁩ ago
    Just for shits and grind, how about neutral to ground? They should be bonded together in the breaker box (and nowhere else), but if there is a fault, you might see some significant potential across them.
    
    You’ve tried different outlets?
    
    source
    -> View More Comments
walden@sub.wetshaving.social ⁨1⁩ ⁨year⁩ ago
I had one that switched on constantly. I changed some sort of sensitivity setting and now it only switches on with actual power loss.

source
Techranger@infosec.pub ⁨1⁩ ⁨year⁩ ago
I would wonder about shared loads on the mains causing undesirable power factors or other electrical noise. I’d check that the mains frequency remains within specification as well: 50/60Hz ± 3Hz according to APC.

source
PrettyLights@lemmy.world ⁨1⁩ ⁨year⁩ ago
My laser printer being on the same circuit causes my UPS to trip unexpectedly even though they aren’t connected. I’ve had to move it to another breaker.

source
thelittleblackbird@lemmy.world ⁨1⁩ ⁨year⁩ ago
It could also be noise on the line, try introducing a ferromagnetic filter (low pass filter) to see if the situation improves.

Check also that the occurrences are not linked with the activation of an electric motor. I one bought a meat grinder that evertyime was on the analogical radio got only noise. And they even were not connected to the same circuit.

source
CarbonatedPastaSauce@lemmy.world ⁨1⁩ ⁨year⁩ ago
Don’t know for sure but the first thing I’d do is get a new battery. 3 years ago means it’s overdue, and if the battery is damaged or not putting out proper power levels it could be the cause. I always know it’s time for a swap when my UPSes start tripping for no good reason, although I’ve never had one go nuts like yours is. They usually just stop working if I let them go way too long between swaps.

source
Overland@lemmy.world ⁨1⁩ ⁨year⁩ ago
I’m dealing with something similar but not as bad.

Our best guess is brownouts for our instance, but it doesn’t sound like what you’re dealing with. Good luck. Let me know if you’re able to find out. Hope the hardware isn’t failing.

source