Tesla Solar Panels & Inverter (no powerwall) with Neurio and two electrical panels

[shred86]

Yeah, checked that multiple times as well. I can't physically see the service lines when they exit the main electrical panel and go into the meter, but I just compared it to the electrical diagram that's posted on the inside door (which is what I used in the original post image). I've checked the CT clamps, voltage reference wires and phases multiple times on multiple days just to make sure I'm not going crazy, haha.

Nope, no Neurio in the inverter, just in the main electrical panel. The Tesla inverter just reports solar production. I had PTO for about a month before I got the Neurio installed so I only got solar production in the app and it was always correct as far as I can tell. The Tesla inverter page also shows the "Instant Flow" which is always correct. I'm guessing this is why the solar production graphs are correct, because it's just using the reported solar production.

Interestingly, I have seem some Neurio setups on Tesla solar installs where they used a third CT clamp to measure solar production which is connected to the CT4 port. This seems redundant though since the inverter reports production and the Neurio simply reports its information to the inverter which is then pushed to the Tesla servers. I have a feeling this is just how they use to install Neurio on non-Tesla inverters that maybe didn't report solar production?

In the last photo below, I included a picture of the Tesla solar inverter main page. You can see "Instant Flow" is the solar production the inverter itself is reporting. "Site Meter" is the sum of CT1 and CT2 from Neurio. "Tesla Solar Inverter Meter" says Not configured because I don't have the CT clamp going to CT4 on Neurio. Other installs that do have that will show the solar production reported here as well.

 

[shred86]

I should note, I realize I have the CT 2 clamp on the left (or bottom) main service lines, but it matches with the red voltage reference wire circuit breaker. I don't know which side is technically Phase A versus Phase B, but all of the wiring in my panels has the red wires going to the left bus/phase and black wires going to the right bus/phase. Even in the 100A panel, it has a black service line with a red line going to the left bus bar. My understanding is black is usually Phase A (L1) and red is usually Phase B (L2), but I'm not sure if my panel just has Phase B on the left and Phase A on the right... or maybe I just have it all backwards (I'm not an electrician).

Either way, it shouldn't matter as long as CT1 is on the same phase as the black voltage reference wire and CT2 is on the same phase as the red voltage reference wire. I've physically checked the circuit breakers to see which bus bar it's attached to so I'm confident the pairings are correct.

 

[shred86]

Lastly, for trend data I compared day-by-day from 1 Jul to 22 Jul what the Tesla app was reporting for grid import/export compared to what my energy company report showed. Over those 22 days, the Tesla app is showing I imported 1545 kWh from the grid and exported 1093.4 kWh to the grid. The energy company report shows they delivered 1007.13 kWh and received 757.84 kWh. So basically, the Tesla app is showing that I'm both importing and exporting significantly more than I actually am (assuming the energy company report is correct).

 

[jgleigh]

Phase A (black) vs Phase B (red) is arbitrary and just general convention. As long as the CTs are aligned with the correct phases it doesn't matter.

It looks correct, but something is adding home usage to your solar production instead of subtracting it.

If you want to move the CTs around without the splitter, you'll need Phase A on CT1 and CT4 and Phase B on CT2 and CT3. And all of them set as Site.

 

[jgleigh]

And check out this option of moving CT2 to CT4 and flipping the CTs around.

Tesla Neurio Device Accuracy

Home and Grid data is back for me. We’ll see how it works going forward. Same here, so far it’s working. I really won’t know for sure until tomorrow late afternoon or so.

teslamotorsclub.com

Tesla Neurio Device Accuracy

Did you get it going? Mine has worked now for over 24 hrs so I’m thinking if it sticks maybe better instructions with pics. Everyone setup could be a little different however. Good thing m3solar figure this out if it works for everyone. I have a feeling more are going to need a solution with...

teslamotorsclub.com

 

[shred86]

I reset the inverter and Neurio and basically watched the values on the Tesla inverter page and Tesla app for an hour while turning different devices on and off. Summarized and detailed observations below.

Summarized observations:
- Tesla app appears to update using Neurio provided values roughly every several seconds to several minutes. It doesn't seem to be consistent. This is really important because it's tough to just look at the Tesla app to troubleshoot anything.
- Tesla app will sometimes show a value for a few seconds that doesn't make sense (different from what Neurio is passing) but reverts back to "normal" values (matches Neurio).
- Neurio appears to be reporting an abnormally high Site Meter values, certainly when exporting. This only appears to occur when home consumption is low (no ACs running, no car charging, etc.), although it could just be masking an underlying problem with Neurio. However, seeing how the Tesla app seems to be reporting higher grid import and export as compared to my utility meter readings, I think it applies to both import and export. It's just more significant at higher kW values (almost like a multiplier being applied).
- When the Neurio Site Meter export value exceeds the inverter reported solar production (Instant Flow), this is when I see the 0 kW home consumption and solar production in the Tesla app main screen (which is sometimes well above the inverter rating) that doesn't match the actual inverter reported solar production (Instant Flow); it just matches what the Site Meter export value is. However, the solar production graphs are always correct so I'm assuming these are generated using the inverter solar production data directly.
- If I'm importing from the grid (high energy demand from the home), there doesn't appear to be an issue. At least everything looks believable but the numbers are probably still slightly off based on my comment above in the third bullet.
- If I'm exporting very little to the grid (still fairly high energy demand from the home), there also doesn't appear to be any issues. Same caveat as the above statement.
- If I'm exporting a lot to the grid (low energy demand from the home), this is when I see the issues with what Neurio is reporting thus causing Tesla app issues.

More detailed observations:
- Prior to the inverter being fully initialized (no solar production), the "Site Meter" was indicating about ~0.5 to 0.6 kW import which makes sense (no ACs running, car not charging, etc.).
- After the inverter was initialized and solar production was occurring, the Site Meter indicated I was now exporting but a value higher than what the inverter was actually producing/reporting. The inverter (Instant Flow) was showing 2.9 kW (cloudy at the time) and the Neurio (Site Meter) was showing 3.8 kW export.
- I could see the Tesla app was updating maybe every 10-20 seconds from what the inverter page was showing. I also noticed sometimes the Tesla app would randomly show values that didn't make sense (e.g. 0 kW home consumption) for several seconds but then correct itself. I never saw these values on the inverter page which appears to be nearly real time.
- One of the ACs turned on and I could see the Site Meter export kW drop on the inverter page. My second AC turned on during the same time and I could see the export drop to now become import which all makes sense. The Tesla app appeared to be following along with a 10-20 second delay.
- At some point several minutes later, the Site Meter jumped up to ~10.6 kW export which is above my inverter limit (7.7 kW is the max I see). Not entirely sure when this happened, but it is when the ACs were off, so low home consumption. The Instant Flow (solar production) on the inverter page was showing 7.7 kW during this time (clipped at max production). The Tesla app was showing 10.6 kW solar generation, 10.6 kW grid export and 0 kW home consumption. This is the strange behavior I typically notice.
- During this time, I started charging my car and the ~10.6 kW export becomes 0.8 kW import. Makes sense, large energy draw to charge my car. The Tesla app reflected this correctly now showing 7.7 kW solar generation, 0.8 kW import and 8.5 kW home consumption. Stopping the charging, it jumps back up to ~10.6 kW export. Tesla app is back to where it was previously (10.6 kW solar generation, 10.6 kW grid export and 0 kW home consumption). Of note, my car charger circuit breaker is located in the 200A panel which is where the inverter circuit breaker is.
- To test increased electrical load in my 100A panel, I turned on a bunch of devices in my den (lights, fans, TVs, receiver, etc.) and I could see the ~10.6 kW export drop to around ~10.5 kW export which makes sense. This seems to indicate my theory that the Y-splitter cables were the issue might be incorrect since it's accurately showing more energy draw when devices in the 100A panel are used.
- Even when solar production drops (e.g. clouds), I still notice the issue with 0 kW home consumption. For example, right now I am seeing 3.8 kW solar generation, 3.8 kW grid export and 0 kW home consumption. The inverter Instant Flow is only 2.8 kW. The only thing that's consistent is it's always when there's low electrical demand from the home (no ACs running, no car charging, etc.) and solar production is occurring (never see this issue at night).


I'm still waiting to hear back from Tesla. I don't want to mess with my CT clamps or setup right now in case they're actually trying to troubleshoot it. There's definitely something weird going on with the Neurio itself. I don't think this is just an app front or backend issue. The Neurio is clearly reporting incorrect measurements. It also doesn't appear to be just a matter of values being doubled (at least in my case).


Screenshots of inverter page compared to the Tesla app. Note the actual solar production being reported by the inverter (Instant Flow) compared to what the Neurio is incorrectly reporting as export (Site Meter). The Tesla app in this case will show the Site Meter export value as solar production on the main screen. The Site Meter value jumps around a bit and the Tesla app doesn't update real time hence the slight difference in values.

Same thing except when energy is being imported from the grid, everything appears "correct".

 

[shred86]

Just to update anyone that comes across this thread, I'm trying the solution @jgleigh posted above that's outlined in this thread. I moved the CT clamps (again, I have Y-splitter cables that goes to two sets of CT clamps) plugged into CT2 on the Neurio over to CT4 and flipped the orientation of those CT clamps. Leave the Metering page on the Tesla inverter settings as is, which I realize is very confusing because it looks like it's misconfigured now (see screenshot below). You would think this setup only measures one Phase which CT1 is connected to, but as far as I can tell, it appears both Phase A and B import/export is being reported through what's labeled as "CT1" on the Tesla inverter page, and you can basically just ignore what CT4 is showing. I tested this out by turning on devices that use the phase that the CT4 CT clamps are attached to, and I can see it being reflected in the totals reported by CT1. This seems to be further supported by what this user did when he performed this fix and only having the CT clamp for CT1 attached resulted in a very high and inaccurate reading until he attached the CT clamp for CT4 in the reverse orientation. Assuming this fix is accurate, the Neurio has to be reading energy flow from both CTs, so the only logical explanation to me would be that it's taking the sum of those are reporting it all as CT1.

I'm planning on comparing the Tesla app with my energy company energy report in a few days to see if this fix really works for me. Other users have confirmed it though by looking at their electrical meter that shows import/export in kWh and comparing it with another energy monitoring device (e.g. Sense), and they seem to match.

 

[jgleigh]

Do you have a current meter? If you do, you could compare to what the CTs are reporting.

 

[shred86]

I actually do have a digital clamp meter but it only measures amps or volts, which I realize I can multiple to get watts but I was under the impression the readings might not be accurate based on this information.

 

[Mase408]

The quality of these Neuria CT's are very poor. I captured this same issue during my install. The tech's had to swap a few different modules before they could get the right reading.

 

[jgleigh]

I actually do have a digital clamp meter but it only measures amps or volts, which I realize I can multiple to get watts but I was under the impression the readings might not be accurate based on this information.

Pretty sure that's all the Neurio is doing too. It measures current at the clamp and then voltage based on the breaker it's attached to. Just suggested cause you can take a few measurements at the different CT clamps and see if the total power draw aligns with what the inverter page is displaying.

 

[shred86]

So I went ahead and compared it with my digital clamp meter... measured around 123V for each service line so I just used that to multiply with my Amps measurements. I pulled the two CT clamps off going to the 100A panel so I was just looking at the clamps going to the 200A panel.

Bottom line, it's definitely not matching with this "fix". At least from what I'm seeing, it's pretty clear this "fix" is also just measuring the phase on CT1. When I pulled off the CT clamp going to CT2, it did not change the readings on CT1 at all.

I reverted back to how Neurio is supposed to be installed and what I found was CT2 always seemed to match fairly close. However, CT1 seems to be the issue. When exporting to the grid, it looks like CT1 is basically double the value of the measurements I'm getting with the digital clamp meter. When I was exporting much less to the grid (higher demand from the home), the numbers were closer aligned and at one point, they seemed to match. I also noticed this behavior while the inverter was booting up and solar generation wasn't occurring. High grid import numbers were inflated.

Example, Tesla inverter page showed CT1 exporting 5,200W. I measured around 21.1 Amps, multiplied by 123 = ~2,600W.

I'll still compare my energy company report when I get the data with the proposed "fix". I ran it like that all of yesterday so I'll just compare that day. However, based on what I saw today, it seemed pretty clear to me this wasn't fixing my issue.

 

[shred86]

I compared the Tesla app numbers from 28 Jul (when I had CT2 moved to CT4 and in reverse orientation) with my energy company and while they're slightly closer aligned, they're still off quite a bit. I think this makes sense if I'm only measuring CT1 with this setup and it's reading high.

Tesla App / Energy Company	Tesla App	Energy Company
From Grid / Delivered	44.8 kWh	40.416 kWh
To Grid / Received	38.8 kWh	33.597 kWh

Still haven't heard from Tesla support...

 

[SMao]

I have simple setup, one 7.6kw inverter (no powerwall), 8.8kw solar panel, and only one 200A panel, and one Neurio meter with two lines (CT1 and CT2).

Still same problem, for example, during 1PM peak production, solar production 7.4KW, however, Neurio meter reported 12.6KW export, while the actual export is 12.6/2=6.3KW

The issue also happens during night time (no solar production), when the home consumption is 0.5KW, Neurio reported twice the actual value, which is 1KW.

Root cause should be at Neurio meter or inverter, and Neurio always reports twice the actual value

Some updates, today, Tesla technician came, and fix the issue. He first updated the Neurio firmware, which does not solve the problem. Then, he found that both CTs are bad, reporting 2x the actual value, he did find 2 good ones from his vehicle after testing 5+, and per him, more than 2/3 of the CTs are bad, that explains why so many people are affected recently

 

[shred86]

Some updates, today, Tesla technician came, and fix the issue. He first updated the Neurio firmware, which does not solve the problem. Then, he found that both CTs are bad, reporting 2x the actual value, he did find 2 good ones from his vehicle after testing 5+, and per him, more than 2/3 of the CTs are bad, that explains why so many people are affected recently

Thanks for the update! I have four CTs and tested them individually and they all read high as well, although it’s not always 2x. Sometimes it’s fairly accurate and other times it’s basically 2x. Fingers crossed this is will be the solution to the problem many of us are seeing.

 

[Donaldinc]

Technician finally came out after 2-3 weeks of level 2 support queue. Technician added a third CT clamp to port 4 to monitor the Tesla inverter and ran diagnostics across both the Neurio and inverter. All seems normal in the interface but never on the Tesla app.

The Tesla app still showed 2x value on solar production and 0 home usage. Technician was baffled until I kept pressing on him trying different ct clamps. There is a known issue with a majority of Neurio ct clamps failing or not working as intended. After replacing about a dozen Ct clamps, the measurement in the inverter finally matched the Tesla app. If you are having the same issue, I am confident that going through CT clamps until they work will eventually resolve the issue.

I tried reversing the second CT clamp as others mentioned which appears like the data is correct but unfortunately only measures half the phase on the panel. It is not the solution, force Tesla to come out and replace the Ct clamps.

 

[shred86]

Finally got a technician out after 2+ months in level 2 support queue. The issue ended up being a bad Y-splitter. Specifically, just one side of the splitter was bad where it was reading higher (almost double).

I explained all the troubleshooting I did and also the fact that other technicians ended up determining the CT clamps were bad. He just used a clamp meter to measure the amps on the service line to get the Watts (Amps * 120V) and was comparing the estimate to what the inverter page was showing for just a single CT clamp plugged into the Y-splitter with the solar system off (disabled via the app just so we could measure grid import only). I also varied the load on the house by turning the ACs off/on and starting/stopping charging on my car just to make sure we were getting matching readings with different loads. It ended up just being one specific side of one of the Y-splitters (I have two Y-splitters) was reading around 140-160W when it should have only been 60-80W. After swapping out the Y-splitter, he enabled the solar production and we totaled up the clamp meter readings with what the inverter page was showing and it pretty much matched up. I also tested the case with low load (ACs off, no car charging) with high solar production, which is where I typically saw very strange behavior of export higher than my inverter is capable of with no home load, but everything looked correct.

He also tried some new CT clamps during the troubleshooting. Oddly enough, two of the new CT clamps seemed faulty because they were showing I was exporting when it should have been import (solar production was off). Double checked "Flip" was disabled on the inverter page as well. We ended up just swapping those two back to my old CT clamps. I have a feeling there's a batch of CT clamps out there that must be wired backwards which is why there's so many posts about weird app readings and users just using the "Flip" option in the inverter page.

We also tried not using the Y-splitters but learned that if you try to configure the inverter site meter page with all four ports set to "Site", it throws an error that basically says you can't use all four ports in a "Site" config, despite the fact it seemed to still be reading/working fine.

Anyways, I'm hopeful it was just that one bad Y-splitter, but I'm going to keep an eye on things over the next couple months just to see if I'm still seeing the strange behavior in the Tesla app and also compare it with my energy company report.

In summary, if you have odd readings in the Tesla app, it can be a number of things:

• Neuro is installed incorrectly. Hopefully pretty rare with "qualified" electrical technicians, but I've seen other users post pictures of setups that make absolutely no sense (e.g. only a single CT clamp).
• CT clamps could be installed incorrectly (reversed) or you have CT clamps that read "backwards" as we discovered today. This should be fairly obvious since you'll see export when it should be import. This is something you can easily fix either by flipping the CT clamp or enabling the "Flip" option on the inverter page for those CT ports.
• Bad CT clamp(s)
• Bad Y-splitter, which are only used for setups with two service panels.

I don't believe the "fix" of moving CT2 over to the CT4 port and flipping the CT clamp is actually a solution but rather just masks the problem since CT1 is likely reading nearly double and you're just measuring a single phase. The CT reading on CT4 isn't actually calculated into your site totals (you can see this on the inverter page) and is also not configured on the inverter page. In fact, you can't configure it because it's an invalid configuration.

 

[Texasnewuser]

Phase A (black) vs Phase B (red) is arbitrary and just general convention. As long as the CTs are aligned with the correct phases it doesn't matter.

It looks correct, but something is adding home usage to your solar production instead of subtracting it.

If you want to move the CTs around without the splitter, you'll need Phase A on CT1 and CT4 and Phase B on CT2 and CT3. And all of them set as Site.

Hi three months in and utility is still showing 30 to 50KWh per day useage whilst Tesla shows zero input from utility. Cant speak to anyone any ideas?

 

[Sam88]

Hi,
Does anyone know if I can even install Neurio in this kind of panel. Where is the main feeding line.
Thanks

 

[miimura]

I can't tell from your pictures. The right place to put Grid CTs is the bus bars between the main breaker and the breaker bus. There needs to be enough clearance behind the bus bar and the opening in the center of the CT has to be wide enough for the bar to fit through. If I had to guess, I would say that the CTs that came with my Powerwall system would not work with this panel, but the ones pictured at the beginning of this thread might.