Power quality opinions

Hi all,

I installed a Sense in the home we bought last year and noticed that our frequency of dips & spikes is quite high. It seems that something is amiss, similar to pchase [0], DanielC’s [1], and brianf’s [2] concerns in previous posts.
[0] Resolved My Power Quality Issue with a New Transformer on the Pole
[1] [Power Quality] - 1000x Dips and Spikes in 3 Days
[2] Power Quality - 350 events in 5-6 days

I’ve attached screenshots & the raw data here.

For background, the house was built in the 1950s, has 100amp service, a 4 ton central AC, electric dryer, electric range, and a pool pump as large loads. There were no major concerns raised by the home inspector, apart from some double tapped breakers. The lights dim slightly when the AC or dyer turn on, but there are no other noticeable symptoms.

Any thoughts or feedback on potential issues would be appreciated.


power_quality_raw.csv (78.5 KB)

Hi! Cant believe my post is helping people! So I am glad to help with this. Glad you tagged the post otherwise I wouldnt have knonw.

Our issues are good now but we had a bunch of issues that compiled over time. That we tried to fix ourselves or through friends of friends. We actually had lots of audio and video equipment fry from all of those spikes and dips.

  1. We did not have a proper ground
  2. Our Neutral in panel
  3. The utility neutral.
  4. Old AF service drop

You gotta get a good electrician in. Not a friend or DIY. Straight up show the data and say others had bad (or missing) water grounds, and bad neutrals. It cost us around $2,500 all in.


You posted your data so I plotted using a little routine I developed for seeing in a little different way. I plot the mins (dips), then the maxes (spikes) agains on another to see if they are happening at the same time of separated and on which day, so you can look for patterns. Looks like you are mostly seeing dips, mostly on L0 (or L1 on the chart).

Smaller number of spikes on L1 (L2), but some.

This distribution chart just shows the same stuff in a different way.


From looking at your graph, I think the problem is with your HVAC. The dips are more frequent during the heat of the day, not so many during the nighttime hours. Re-read this old forum post (posted above) of other users with the same problem.

You may need to replace your Run/Start capacitor or add an additional Hard Start kit. I installed a brand new HVAC system and had this problem. I think my power cables to my HVAC are undersized based on the distance from the meter panel to the outside compressor. They meet NEC code, but they are not large enough for the power demand on ‘starting watts’.

1 Like

Very cool, thanks for running your routine on the data! :+1:t2:

Thanks @Dcdyer, I’ve attached the graph of usage during the same time period and it does look like the AC starts match up with the voltage dips.

Is there a reason this would manifest as voltage dips on L1 exclusively? My understanding is that the 240v circuit for the AC is L1/L2 out of phase, but admittedly I don’t fully understand the mechanics of large inrush currents & how the capacitor affects L1/L2 voltage, etc.

More broadly, would you say it’s a cause for concern beyond the minor annoyance of dim lights on startup?

In any case, I’ll continue to monitor & report back w/ data on days when the AC is off.

I’d actually replaced the dual run/start capacitor two weeks ago during an annual HVAC service simply because I had a spare on hand and the previous one was on the older side.

I’m thinking it’s your AC as well, though I noticed one more thing that makes it a little weird. Usually when you put a turn on a big 240V load (like an AC) it drops your voltage on both legs. It looks like when you have your dip / spike events, your legs are going the opposite way ! Just to confirm that observations with numbers I did two things - analysis to understand what type most of the events are, then looking at the relationship between those two legs at those events.

There are 4 possibilities for each leg when you have a dip / spike event in the log, because there is a minimum and a maximum measurement for each leg in that moment in time (we can discuss more on why 2 later)

  • Both In - the leg voltages are in between the minimum and below the maximum limit - OK voltage, so the other leg caused the event.
  • Both Out - the leg voltages are outside the minimum and above the maximum limit on bot ends - unlikely, but we have to check for this anyway.
  • Below - the leg minimum is below the minimum limit
  • Above - the leg maximum is above the maximum limit

When I look at a count of the kinds of events by both L0 (vertical) and L1 (horizontal) below, I can see a few interesting things. First off, good news - there are NO “Both Out” events ! Second, the most common event by far is a Below on L0 happening 71+19+827 times = 917 out of 1000 events. When that happens, most of the time L1 is either Both In, of the L1_maximum is Above the limits.

You might want to check with your electrician to see why voltages might be moving in opposite directions when a big load jumps on.

1 Like

To: @chestnut

  1. Take a very up-close detailed expanded wattage graph (screenshot using your phone APP) of a single event or several single events. Post those pictures on the forum. Look at the ‘max peak starting watts’ versus the ‘running wattage’. My starting peak watts was 6500 watts. After installing a ‘Hard Start kit’, my starting watts was reduced to 3500 watts. I had a huge argument with my HVAC installer. He insisted that there was no problem when he tested using his cheap voltmeter. You need screenshots and lots of them. I was able to get the HVAC manufacturer (American Standard) to upgrade and install my Hard Start kits for free.

  2. As for dips on L1 versus L2: Maybe when your compressor kicks in, so does the outside fan. The fan runs on 120Vac. Maybe the problem is more with the fan. Do you hear a squeal when it initially starts up? Are there leaves or debris on the fan blades? Does the fan bearing need oil? You should see two distinct starting peaks when your outside AC kicks in compressor, then fan.

  3. Is this a concern? Maybe not, but you might be stressing other sensitive electrical equipment and setting yourself up for premature failure in an area you don’t suspect.

  4. If you have a home generator and it is capable of running your AC, a high peak load on HVAC start-up is going to tax your generator.

  5. My “Run/Start capacitor” and my additional “Hard start kit” use a mechanical relay. (Check, some have an electrical chip that sense startup. Don’t oil those.) When mine were new, they were stiff. They tripped, but not fast enough. The SENSE monitor still saw a high starting wattage. I used CRC-26 electrical spray and some WD-40 to lubricate the mechanical relay. I still had problems for about a week, then the trip became fast enough that the problem went away. Maybe the lube needed time to work.

Hard start kits are $100-$250 depending on brand. What is your AC manufacturer and brand? When was it installed? When was it last serviced? What is the length of wire from your breaker panel to your outside HVAC? What is the wire size? Wire Gauge to HVAC compressor?

PS. I repaired my problem. Now I see the same problem on my neighbor’s HVAC. We are connected to the same bus bars inside our ground transformer. I now have new voltage dips due to his failing capacitors.


@chestnut , did you ever get resolution on the power quality effects you were seeing ?? After looking at your L1 and L2 numbers, I kind of determined that it’s hard to figure out what’s going on by just looking at the values outside the limits. I decided to look at my Home Assistant sampled L1 and L2 over 3 weeks (sampling from Sense to HA is every 2 minutes - not nearly as detailed as Sense data.

Just looking at the general case it looks like there is fairly extensive positive correlation between L1 and L2, though hI think the greatest info is revealed by the spikes / dips, though not just the out of spec spikes and dips.

This what correlation is all about:

1 Like

@chestnut, finally noticed a pattern a little like what you saw. Here’s my leg voltages today vs. my Total usage on the bottom. It looks like some device usage in the morning (10AM to 11AM) caused my legs to separate substatially, though not take them out of spec.

If I look at that 1 hour slot of time, I see negative correlation between my voltage legs.

But if I look at the data for the rest of the day, it’s mostly positive correlation, though not completely.