Power quality question - substantial voltage drops on one leg but not the other

Had some LED lights in the house begin flickering earlier this week, and at the same time several of the UPS protecting computers and home networking gear tripped multiple times. Since then, we’ve had the issue repeat and I started looking at the Sense labs data for power quality and we’ve got lots of trapped errors there. The relevant facts seem to be that the voltage drop is only on one leg - the low voltage leg happens multiple times a day but is NOT constant, with the low voltage dips often being around 100v with some excursions even lower. Scatter plot of voltages attached, you can see all sorts of drops but seems poorly correlated with the other leg – it doesnt’ look to me like we’re seeing a rise on leg0 when leg1 tumbles.

Thoughts? Raw CSV as exported by Sense labs here also…

power_quality_raw.csv (7.3 KB)


@ljwobker, here’s a different view of your power quality data from a little different perspective, against a few other users’ experiences. The chart shows the comparison of the offending min vs min or max vs max for the two legs. Yours is the first chart. And as you suggest, L0 is stable and and L1 is the sole deviant one, except for a couple points.

User 5 - both legs shows issues with 100% positive correlation - both go up and down together, a power company issue.

User 2 - both legs are negatively correlated, probably a neutral issue. One goes down when the other goes up, almost linearly.

User 1 - runs mostly on their own generator, solar and their own batteries.

Your issue looks closest to User 4 and User 3 (with legs reversed). Maybe I can check I with them to see if they ever diagnosed.

So this was a fun weekend - I got back home Sunday and the lights were flickering like a poltergeist was in the house… eventually I walked outside to check the utility meter and one of the lugs/socket connections was so loose that it had melted a hole through the acrylic housing of the meter itself.

A couple of emergency calls to the utility and a local electrician, a day with one leg of power (the utility capped the failing one but left us half the 120v side of the panel), and then another day of down while the work was being done… out $2200 but now I have a new utility side panel and a new meter and I’m pretty sure the house won’t burn down.

So… my question is… why didn’t I get an email from Sense about stuff like this? Looking back at the power quality data, it seems totally obvious that SOMETHING was wrong, and there had been plenty of suspicious power readings from at least a few days before things got really out of control.

I’m not trying to rant and rave here - I don’t pay Sense a subscription fee to monitor my power or whatever so I’m really trying not to be an entitled brat… but it DOES seem like this is exactly the point of having the power quality functionality in the system, right? If I can get an email whenever the power goes out and my Sense monitor goes offline, I would think I could get one that says “your power is WAAAAAY out of whack and you might want to think about looking at it…”


1 Like

@ljwobker , that’s a good suggestion because that’s the concept behind the Ting device that does require a 100$ upfront fee and as yearly subscription (in my case subsidized by my insurance company as part of a trial). It’s interesting because the Ting plugs into a 120V outlet so it really only monitors one half of a user’s electrical system.

So far it hasn’t alerted me to anything - not even the one dip out of range or a gap in the data. That plus the single leg monitoring makes me skeptical that it does anything better than Sense, yet.