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

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

To help answer your question about resetting - yes, your data will be wiped. The good news is that, when I had to start over at least, device discovery seemed to come a little quicker.

I also hear you on support - the response time is what is absolutely hurting Sense. They might have the best product in the world with the best models, but if customers aren’t getting the support they need, a company is in big trouble. Note that this isn’t something that just started, the customer service speed was one of the first things that made me leary of buying Sense. It’s been an obvious weak link to anyone paying attention since their early launch 6-8mos ago. I chalk it up still to growing pains of a small company but it still should be top priority with all this feedback. The customers are trying to tell Sense how to improve the product, and Sense would be wise to prioritize their efforts based on what their customers are telling them. We’re always told “Sense is a small company,” etc. No better time to get all this sorted out than now, before Sense turns into an even bigger unresponsive company (like one of their direct competitors - Neurio - has seemed to have done). Those are who don’t learn from history are doomed to repeat it.

So it seems that Sense support DID reset my unit yesterday at 3PM, but never sent me an email reply to let me know.

It would have been nice to know. For example, I probably would have cancelled my weekly generator exercise that runs for 45 minutes every Saturday at noon, Since that exercise is under load, it means that Sense sees zero usage during that period.

I would have also liked to capture my daily usage and update my tracker spreadsheet, but all that data is gone now.

I do hope their support department improves over time as it most certainly leaves a lot to be desired at the moment.

So this morning my Sense said it had finished checking signals and everything was looking good.

I was down in my shop and made the air compressor come on. It pulls right around 8,000 watt, which was about how much solar production I was making a little after 4pm. Here’s the power graph:

So the solar production did drop to near 0 W as expected, but usage did NOT increase to around 9,600 watt as expected.

The whole point of me moving the usage CTs to just after the POCO meter, was so that I could begin capture all usage, including what the shop consumes…

Something’s off - your solar production shouldn’t change, it’s not like the sun turned off. I’d have to think more on what could be going on, but right off the bat that’s odd…

Keep in mind that my Solar CTs are on the feeder going to the shop sub-panel, that the shop equipment is connected to. So I would expect to see the air compressor “eat into” the solar production. But what I don’t get is why the load CTs, which now monitor everything my entire household consumes, did not pick up the ~ 8,000 watt load increase that occurred when the compressor turned on.

I guess the Sense cannot deal with ANY loads connected to the solar CTs…

I see now. I think the problem is that your main cts are measuring incoming from the grid - note the usage did bump up a little when the compressor came on, but since the bulk of the power for the shop was coming from solar nothing much changed at the mains. The amount of energy your house is using isn’t directly measured with the solar set-up, it’s calculated. This is the issue I’m seeing where it’s improperly calculating how much we’re using during the day because it’s improperly measuring how much we’re producing.

I’d be curious to see what happens if you do the same thing when it’s dark out, or when you’re not producing much solar…I bet you see the mains compensating :wink:

Edit: rethinking this, your mains should’ve been seeing your solar leaving before and after you turned the compressor on, and that should’ve been reflected when the solar stopped leaving as a concomitant rise in usage. Your solar is going out through the main cts, right? May be time to bring in an engineer who could boil down how things are being measured and calculated. If only… Sigh :confused:

So this morning it is overcast and I’m only producing about 1,600 watts. So I went out into the shop, turned on all the overhead lights (12 x 250 watts MH), the compressor (8,000 watts) and car lift (6,000 watts). Despite all my effort, it didn’t make a dent in my usage graph, it only made the solar production disappear.

Not sure what that reverse spike of 3,000 watts is all about.

Here are the actual power readings while the shop loads were on:

So assuming production was 1,619 watts during the shop loads, the usage number should have increased by 11,473 watts (the difference between 1,619 positive and 9,854 negative), for a total of 21,324 watts. Instead, it only increased by 428 watts, and that could have been from something else in the house.

On different but likely related note, here’s my graph from yesterday:

Note how my usage from 9 AM to 11 AM is trending down as my solar production is ramping up. That’s my “daytime dip” as you call it.

So do I need to disable my solar CTs in order for the load CTs to properly see all my devices and report on them correctly?

Having production overlaid on the same graph and having trending is nice, but if it means my usage is being incompletely reported, I’d rather just reply on my inverters and Poco Net Meter readings for my solar.

I don’t suppose it is possible to have one Sense dedicated to usage only, and a second one dedicated to Solar and then have the 2 units both report to the same “cloud account” and get accurate readings that way?

My Solar is over 200’ from my main meter, so there is no way I can hook up my Solar CTs in such away that they will only read production.

I wonder if you could move your main CTs back to where they were, and have Sense not try to calculate the ins and outs - simply show how much you’re producing and how much you’re using? It would be like having two units, reporting their data separately to the same graph instead of one unit trying to combine both of the data streams. That’s something I’d hope an engineer might be able to answer.

Well, the way the CRs are connected right now is giving me the exact same readings as before. That is to say, any loads in the shop do not register as usage, they just pull down the solar production by the amount they consume.

It seems to me the way Sense is configured, ANY load going through the Solar CTs will be “corrected” out of the usage reading from the main CTs. In a way I can see why that is, as otherwise “negative load” i.e. solar production, would have a huge impact on the usage readings from the main CTs.

I could install a sub-panel in the shop dedicated to the solar breakers, but as stated above, I have no way to extend the Solar CT leads 200’ down to the shop so that they would be the only load they would measure. Maybe a pair of “wifi” Solar CTs could be used in situations such as mine.

I told “Alex” to please pass on the word to have an engineer chime in here, but so far nothing.

I went back and looked at your diagram and I agree that there is no way this can work correctly as the solar CTs see the net of solar production minus and load that is being pulled from the 200 amp shop sub panel. I wonder if you could install the CT on the solar breaker(S) in that panel and then add 200 feet of cable to the CT output to connect it to Sense or if that extra resistance would screw stuff up.

Hey @peter - I’ve informed our team and they are taking a look. It does sound similar to the problem @Howard posted about in the Facebook group. Our team thinks they have a solution to the problem that they are working on now and should be available soon.

Thanks for providing this info and being patient with us!

1 Like

Just to comment - the issue that @Howard posted about on facebook is the same one described here: Solar customers: Recognizing the "Daytime dip" and how it might impact your Sense numbers

@HilarioAtSense, has there been some correlation drawn between supply-side vs. load-side solar, or to the original orientation of our CTs as @Howard suggested yesterday? It seems as if there are a mix of people reporting a solar dip (like Howard, myself, @Stuart, and facebook user Brian Schmidt), and others reporting a solar rise (like @ryan.lynch and facebook poster David Xenos).

I know you guys are working on it and I appreciate it more than you know - it’s honestly been my personal bugbear since purchasing so if you’re able to get it figured out and patched I’ll be over the moon.

I made a SWAG (Scientific Wild Ass Guess) that assuming this is a math issue whether you see a dip or a blip has to do if the system accepted your CTs as installed or if they had to be mathematically flipped (versus physically flipped). I bet when they get flipped by math there is some other equation in the algorithms that don’t make the flip correctly.

There’s certainly something keeping them from effectively releasing a patch. I wonder if the fix is going to be “turn one CT one way, and the other the other way.” :wink: That would probably require a reset, I’d guess.

1 Like

I thought about adding 200’ of cable to each Solar CT, but I suspect you’re right that would add quite a bit of resistance and probably some stray induction from the 500 MCM cable it would be buried with. Besides, the prospect of trenching again between the house and shop is not very appealing and it is nothing but rocks. I wasted $300 on a large trencher that could not deal with the rocks, so I had to get an excavator instead. It is going to be much much cheaper just to get a 2nd sense vs. attempting to extend the Solar CTs the 200 feet down to the Shop sub-panel.

The Sense at the house would measure usage only everything behind the main meter.

The Sense at the shop would only use a single pair of CTs connected to the Solar connector and measure nothing but solar production.

The trick would be to make them both talk to the same “Cloud Account” and give me the nice overlay on the same graph and Usage vs. Production stats, etc.

My issue is different than the “daytime dip” (Which I also suffer from), but I kinda knew going into this that would likely be a problem (having solar breaker mixed in a sub-panel with other loads, physically separated several hundred feet from where the Sense would be installed).

1 Like