Moved load CTs to feed from meter and Sense is really confused now

So yesterday afternoon I decided to move my load CTs from measuring only the load to the house, to measure everything behind the meter, which is where most users would have them. The CTs basically sit between the meter and the main breaker in the panel.

In my case, I have a fused disconnect immediately downstream from my meter base, that feeds both the house sub-panels as well as my shop sub-panel that contains my solar panels. So where is how it looks like now:

So the service comes in the top from the meter and at the bottom, one feeder goes to the shop building existing at the bottom, this is where the Solar CTs are connected. Made no changes there. The feeder to the house sub-panels are existing to the right. This is where the load CTs used to be connected and everything was working fine, but Sense was not able to pick up loads from the shop sub-panel.

After making the change, I did a solar re-calibration, which went fine despite it being later afternoon and a bit cloudy.

Last night the readings were all normal with Solar being close to 0W and my loads being what they normally are. But this morning before sun up, the power meter looked like this:

Later in the morning I checked the Sense Status, and it looked like this:

So it would appear that the solar production is directly impacting the reading of the load CTs in that the solar production pulls the load readings into the negative.

Here you can see that whenever solar production is higher than usage, I no longer get any usage reading on the power meter.

Another look from later in the afternoon:

And yes, I still have the issue where one of the Solar CTs drops out.

And later this afternoon after the sun set, things are starting to return to normal since the sun was no longer around to completely mess up the load readings.

So what’s going on here? Has my sense lost it completely?

I almost thought it might need another recalibration, but at the very least I really think you need to get those solar cts replaced if its dropping - that could be jacking up your calibration. One of my first posts here back in December was something similar - it required a replacement ct and they sent me a fresh device just to be sure - but even then there was a bug they had to fix to get the new one reading correctly. See here to see if any of this matches your symptoms: Failing CT or...? - #2 by NJHaley

I dropped a note to support with a link to this thread. Here’s another view from just now showing the transition from daytime with plenty of production with virtually no usage showing, except for the very brief moments when a cloud blocks the sun, and usage sneaks past production. You can see that at 1pm in the below image. Once the sun went down, the usage graph comes back, as usage is no longer fighting with production for visibility.

That -36 watts solar is exactly what I had before making yesterday’s changes and are due to my always on devices in the sub-panel at the shop where the solar breaker are. The 1,559 watts usage is also about what I have been pulling this time of night.

It is almost as if my Sense unit “remembers” how it was connected before, where the solar production was not impacting the current flow through the usage CTs. Since the change yesterday, solar production is flowing backwards through the usage CTs, since that is the only path back into the grid.

My guess is that during initial installation, Sense determines if solar production flows backwards through the usage CTs or not. In my setup that was not the case before, but is now, but I don’t have a means of changing that myself. I really wish Sense offered more user configurable options instead of everything being “automatic”.

I suppose I’ll try another solar calibration tomorrow early afternoon if I have not heard back from support by then, but I’m skeptical that is going to fix things.

I checked out your thread and I don’t think those are the same symptoms I have, but perhaps there is some correlation that I’m not seeing.

I do think you’re right about sense “remembering,” I remember vaguely someone from sense saying not to change your configuration for some reason or another, I’ll see if I can find where that was - either here somewhere or maybe on reddit.

Here’s a simplified diagram showing the before and after change I made to the Mains/Usage CTs:

It seems to me like you might have to tell it this is a brand new installation and start from scratch as what Sense is seeing down stream is totally different than what it was seeing before. I would send this new diagram to support directly and get their input.

Thanks I will do that. Power Meter from just now. It is very obvious what is going on. My sense needs to be told to subtract out the Solar CT readings from the Mains CTs readings. Right now it is just reading the Mains CTs without applying the needed correction factor.

Based on something I just did, try swapping your main CTs so that they’re on opposite legs. I wonder if when you swapped them you inadvertently put them on the reverse legs of AC.

I was careful to label my L1 legs with Blue tape and my L2 legs with Red tape when I installed everything. I did not reverse the CTs when I moved them.

I did another Solar re-calibration at noon today, and lo and behold, it seems that took care of the issue.

So either doing a re-cal just before sun down (even though I knew I would produce in excess of 1,000 for the duration of the process), or I just needed to let Sense get used to the new configuration and then do another re-cal. I suspect the latter was likely the cure. Fingers crossed that the L2 solar CT won’t drop out again.

@peter It seems like Sense polarity issues is being done manually by technicians, where I believe it could easily be done programmatically during solar setup. See my post here Reverse power detection

I have been following that thread as well. Like I said above, I’m 99% sure that I did not reverse my solar CTs when I moved them from being on the feed to the house panels, to being on the feeder to my whole site.

I do still have the issue where one of my Solar CTs drop out.

You can clearly see where one leg dropped 100% out at 1:35, and 50% out at 2:05 and then 100% again at 2:37.

I have already re-seated the CTs several times and made sure they are both fully closed. That ain’t it.

Does anyone know if there is a support number to call at sense or is interaction 100% via email tickets to support? I ask because often days or weeks go by before I get a response.

100% email unfortunately. I’m surprised someone hasn’t interjected here yet, I’ve posted a couple @Maarja-Liis call-outs already today. They’re not quite like genies, it seems :frowning:

I wonder if you’d be up for something? I’ve got a second pair of cts on the way - we can’t figure out why my solar is under-reporting, and I wanted to see if a different pair of cts would behave any differently. The curves are all perfect, they just read about 6% less than my inverter and meter is telling me. Maybe I could send you my old cts and you could see what happens with your monitor? Maybe they do the same - under-reporting 6%, maybe they do whatever yours did previously on a perfect, functional day? What do you think?

Sure, I’d be up for that. My POCO installed a new meter at my place last week, so I now get daily DE (kWh delivered to me), RE (kWh they receive from me) and NET (overall usage) reports from my meter. So I can compare Sense production numbers to what my RE readings are on a daily basis.

POCO is still working working with their billing vendor to resolve a bug that is preventing their online portal from displaying the DE and RE numbers, but they are emailing me a weekly summary of the daily values until that gets resolved. It sure beats going out and reading the values on the meter at midnight each day. :slight_smile:

Curious how you get Net usage, maybe the question is what you mean by net usage. Don’t see how this would be possible unless the utility was measure your solar output rather than the amount of solar you deliver to them. My utility measures power flow in and out on a peak and off peak basis. They have no clue how much solar I produce that I consume, thus there is no way to come up with a consumption number. The net number is just power flow. I have a separate meter on my solar production and the only way to derive consumption is to take what they say the power flow in is and then adding to it my solar production number less what they claim is the solar out flow to them. That is one of the main reasons I installed Sense as it does that (or should be doing that) in real-time.

[quote=“Howard, post:15, topic:379”]
Curious how you get Net usage, maybe the question is what you mean by net usage.[/quote]

In my case, my utility meter reads the net value, but my solar production meter shows the entire solar produced. That solar production number is transmitted by internet to indicate how much I produced, for SRECs and the like. So the utilities do know how much I produce and my net usage.

Sorry yes, I meant that the POCO can only see they they deliver and receive from me. Since getting the new meter installed and getting the daily readings, I have been maintaining the following spreadsheet:

The sunMAX Production column is the actual power produced by the panels each day.

The Sense Production is what Sense sees the production being for each day. These numbers are always lower than actual because I have equipment in my shop that is connected to the same sub-panel. There is also some voltage loss from the microinverters to the Sense solar CTs located some 200’ away.

The Sense NET is just the next of the Sense readings each day.

The orange columns are the POCO meter readings. The daily values represents the calculated values from the previous day.

The last column shows the variance between the Sense NET reading and the POCO meter NET reading. I’m not sure why there is as much variance as their is, or if I should be doing something different to attempt to reconcile the readings.

Definitely one problem is that you’ve got that CT that’s dropping off. It makes it really hard to say how much you’re actually producing through Sense - you have to keep track of when and how frequently it’s dropping off, which in my case was really unpredictable and could be most of the day or just a few minutes. I’d imagine if that CT was working as it should be, your variation would be much more consistent.

My new CT arrives tomorrow - I’ll post up what I find. It may turn out to give the exact same readings as the old CTs had, in which case the problem lies somewhere in the calculations that the Sense software is doing with the solar CT data. It may turn out that the readings become more in line with what my meter is telling me, in which case the problem lies physically with the CT (production defect, etc.). If the latter is the case, I wouldn’t want to ship you a bum CT to replace your bum CT :slight_smile: If the former is the case…well, I still think you really need to get a replacement from Sense so they’ll be responsible for any issues that may come up. But we’ll see :slight_smile:

About my Solar CT that is dropping off, it looks like I have another issue going on. I’m having issues with groups of microinverters randomly dropping out with an “AC_OVER_FREQUENCY” error during times when the sun is high in the sky. On the Sense Power Meter, it looks exactly the same as when a Solar CT drops out, or shows a lower than expected reading. The microinverter vendor is shipping me an AC power analyzer that will monitor and record cycle-by-cycle metrics to hopefully get to the bottom of what’s going on. So on my end I’ll get that knocked out first. There definitely was an issue with one of my Solar CTs, but re-seating the connector and double checking it was closed all the way may have taken care of the issue. Won’t know for sure until I get this other issue resolved.

On a side note, I bet the sense would make a kick ass AC analyzer with the right software. The sampling rate is certainly high enough. Adding a voltage “Power Meter” and a frequency “Power Meter” would be pretty straight forward I would think. Having a real-time “oscilloscope” of the AC waveform would be pretty cool too, especially if it had the ability to hold, store and zoom. Perhaps that would be an add-on software option down the road? I know this isn’t really the focus of the Sense device, but I would think the hardware is fully capable.

I must say I’m very disappointed in the response time from Sense support.

I opened a ticket on Tuesday the 7th of Feb.

First response came on Friday the 10th where it was suggested that I re-seat my Solar CTs. I replied immediately and got a 2nd response back that same day.

I then tried various things on my own based on this thread, and replied back to Sense on Thursday the 16th about me moving the Solar CTs to their new location upstream from the Safety Disconnect switch. Didn’t get a response, so I replied back on the 17th with some additional information, including the simplified diagram I posted above.

I finally got a response on Thursday the 23rd, asking if it would be ok to factory reset my device to recalibrate since the CTs were moved. I asked if that would mean loosing all my existing stats and devices.

I got a second response back that same day stating that:

“Since the CT’s were moved the channel map and solar calibration is invalid. That will cause the data going forward to be inaccurate. Unfortunately, this is the best option that we have to ensure that the data will be correct in the future.”

So they didn’t really answer my question, but I’m assuming I would loose all my history and devices. After thinking about it for a while, I decided to go ahead with the factory reset and basically start over (assuming that would be the effect of a factory reset).

So I replied back, agreeing to move forward with the reset. No response since.

This is very frustrating. No phone support and email support response time is measured in days and weeks. And there has been no direct response from sense in this Technical Support thread either.

I have been involved with customer support for a Service Provide several times throughout my career, and it is pretty much industry standard to strive to respond to support tickets within 24 hours. Sense is falling way short of that standard.

I understand that Sense is a small startup with limited resources, but I started an ISP back in 1995 with 4 employees, and we had support via email and phone from the start and always responded to customers within 24 hours (typically within 1 hour of the ticket being opened). The ISP grew to over 20,000 customers and 30 employees over the next 5 years and we never broke the rule of responding to customers within 24 hours, even if that meant working 80 hours a week, if needed.

1 Like