Neurio for new breakers

[hulvey]

I've had a strange experience with the Neurio so far, so thought I'd share here in case someone has an idea of what is going on, and maybe it will provide others with some insight on how Tesla hooks these things up. I just had a Powerwall 2.0 and Tesla solar installed and got the PTO a couple days ago. We had noticed even before that the "Grid" figure in the app almost always said 0.1kW no matter what we were doing in the house, and I had called them thinking this didn't seem right, so they said they'd send someone out to fix an incorrect part. That never happened, though, and after PTO, the Grid number was still 0.1kW. We did notice some spikes now and then. After throwing some breakers, we realized that the 0.1kW seemed to be due to our hot tub which was not backed up, and the spikes appear to be the heater coming on occasionally to maintain the hot tub temp. We also have our AC and double-convection-oven on the "not backed up" panel, and found that those, too, would cause the Grid number to go up, but nothing else in the house impacted the grid number. It seems that most of the time, the "Home" number shown in the app is just the Grid minus solar minus battery, but since the grid number isn't right most of the time, the Home number isn't right, either. So, I figured they must have installed a current sensor on the wrong wire or something. They said they'd send someone out next week, but being impatient, I thought I'd take a look to see if I could figure out how to fix it myself. I found the Neurio sensor box, and the CT modules. Searching on "Neurio" and "Tesla" is how I then found this forum thread. One set of the Neurio CT sensors seems to be on the solar panel breaker, which I'd expect. There are 4 other CT modules, 2 sets of 2 which are each twisted together. It looks like they put one around all the high-current breaker wires which all seem to exist the top of the panel with the exception of the hot tub. That sensor is then twisted it together with another sensor on just the hot tub wire. None of the sensors appears to be measuring the grid or the connection to the back-up gateway. I was thinking that maybe they had just missed putting the gateway connection through the CT along with the spa wire, since the breakers are right next to each other. However, then it would seem simpler to just put a single CT on the main grid connection coming into the panel rather than trying to catch all the breakers going out of the panel.

So, does anyone know ... should the Grid CT modules be measuring actual Grid? Or just the "grid" feeding the back-up gateway? I assume it should be the actual grid, and thereby include the non-backed-up loads as well as the backup gateway, as that would seem to explain why they twisted another pair of CT sensors in there which sense all the non-backed-up loads other than the hot tub.

I'm also curious why they wouldn't just pull the PV solar number from the inverter, and the grid number from our SoCal Edison smart meter. I used to have a dongle in my Lowe's Iris system that would give me a real-time reading from the meter, but Lowe's had to change from a 1st-gen hub to a 2nd-gen hub, and they never added back support for that. Anyway, it seems that the Neurio is superfluous in all of this.

 

[mongo]

I've had a strange experience with the Neurio so far, so thought I'd share here in case someone has an idea of what is going on, and maybe it will provide others with some insight on how Tesla hooks these things up. I just had a Powerwall 2.0 and Tesla solar installed and got the PTO a couple days ago. We had noticed even before that the "Grid" figure in the app almost always said 0.1kW no matter what we were doing in the house, and I had called them thinking this didn't seem right, so they said they'd send someone out to fix an incorrect part. That never happened, though, and after PTO, the Grid number was still 0.1kW. We did notice some spikes now and then. After throwing some breakers, we realized that the 0.1kW seemed to be due to our hot tub which was not backed up, and the spikes appear to be the heater coming on occasionally to maintain the hot tub temp. We also have our AC and double-convection-oven on the "not backed up" panel, and found that those, too, would cause the Grid number to go up, but nothing else in the house impacted the grid number. It seems that most of the time, the "Home" number shown in the app is just the Grid minus solar minus battery, but since the grid number isn't right most of the time, the Home number isn't right, either. So, I figured they must have installed a current sensor on the wrong wire or something. They said they'd send someone out next week, but being impatient, I thought I'd take a look to see if I could figure out how to fix it myself. I found the Neurio sensor box, and the CT modules. Searching on "Neurio" and "Tesla" is how I then found this forum thread. One set of the Neurio CT sensors seems to be on the solar panel breaker, which I'd expect. There are 4 other CT modules, 2 sets of 2 which are each twisted together. It looks like they put one around all the high-current breaker wires which all seem to exist the top of the panel with the exception of the hot tub. That sensor is then twisted it together with another sensor on just the hot tub wire. None of the sensors appears to be measuring the grid or the connection to the back-up gateway. I was thinking that maybe they had just missed putting the gateway connection through the CT along with the spa wire, since the breakers are right next to each other. However, then it would seem simpler to just put a single CT on the main grid connection coming into the panel rather than trying to catch all the breakers going out of the panel.

So, does anyone know ... should the Grid CT modules be measuring actual Grid? Or just the "grid" feeding the back-up gateway? I assume it should be the actual grid, and thereby include the non-backed-up loads as well as the backup gateway, as that would seem to explain why they twisted another pair of CT sensors in there which sense all the non-backed-up loads other than the hot tub.

I'm also curious why they wouldn't just pull the PV solar number from the inverter, and the grid number from our SoCal Edison smart meter. I used to have a dongle in my Lowe's Iris system that would give me a real-time reading from the meter, but Lowe's had to change from a 1st-gen hub to a 2nd-gen hub, and they never added back support for that. Anyway, it seems that the Neurio is superfluous in all of this.

Can you posts some pictures of your setup?
The gateway (normally) has CTs also around its bus bars.
As shown with the Lowes issue, tying to interface to other HW can be a pain, the Neurio allows independent sensing.

 

[miimura]

So, does anyone know ... should the Grid CT modules be measuring actual Grid? Or just the "grid" feeding the back-up gateway? I assume it should be the actual grid, and thereby include the non-backed-up loads as well as the backup gateway, as that would seem to explain why they twisted another pair of CT sensors in there which sense all the non-backed-up loads other than the hot tub.

I'm also curious why they wouldn't just pull the PV solar number from the inverter, and the grid number from our SoCal Edison smart meter. I used to have a dongle in my Lowe's Iris system that would give me a real-time reading from the meter, but Lowe's had to change from a 1st-gen hub to a 2nd-gen hub, and they never added back support for that. Anyway, it seems that the Neurio is superfluous in all of this.

In my install, the Grid CT's were pre-installed by Tesla over part of the transfer switch inside the Gateway enclosure. It only measures the grid power going through the gateway. I have other loads that are outside the Gateway and they are not measured in any way by the Powerwall system. The Home figure displayed by the system is indeed calculated from the Grid, Solar, and Powerwall values. In order for us to give you any more clarity on what might be going on in your system, you should attach pictures of how your system is wired.

The Neurio is not superfluous at all. It is integral to the whole operation of the system. The Gateway continuously commands the Powerwall charge and discharge based on the values it gets from the Neurio. Certainly, there are other ways to design the system and I do wish it would consider the data transmitted by my SmartMeter, but today it does not.

 

[hulvey]

Here are some pics. I don't see CT sensors on the gateway busbars. They are at the top and the bottom of the main panel (the narrow one with meter). As I mentioned, those are 2 pairs, each twisted together, with top sensors apparently for the AC and convection oven in the house, and lower one going to our hot tub. The connection to the gateway seems to be via that same exit hole at the bottom, so I'm guessing they just forgot to loop in the big wire going over to the gateway.

 

[miimura]

It appears that your installer extended the wires for the CTs with wire nuts and additional wire. That may be related to your problems. All of my CTs are installed without modification to the factory original CT wires and connectors. It doesn't matter where you put the CT along the same wire, so I don't understand why some of the CTs were moved outside the Gateway enclosure.

 

[Shygar]

The Tesla guys that came today told me that some of the CT clips had stickers put on the wrong side. The CT's have a sticker that says point towards the source to show you what way it should be facing. They found out the hard way that some of them are on the wrong side.

But yea your setup doesn't look right. Neurio makes this so simple but Tesla complicates it. According to Neurio, you put two CT clips on your main lines coming into the main panel, and you put two CT clips on your solar breaker, that's it. They complicated my setup but mine was fairly easy to see since it was just 3 breakers with one being a single subpanel. I'm not sure what's going on with yours but you should call the Tesla customer service number and make sure they have a ticket created for you to come out.

 

[mongo]

Here are some pics. I don't see CT sensors on the gateway busbars. They are at the top and the bottom of the main panel (the narrow one with meter). As I mentioned, those are 2 pairs, each twisted together, with top sensors apparently for the AC and convection oven in the house, and lower one going to our hot tub. The connection to the gateway seems to be via that same exit hole at the bottom, so I'm guessing they just forgot to loop in the big wire going over to the gateway.
View attachment 284984View attachment 284985View attachment 284986

Thanks for the pics, they help.
Your setup is not bad, the main issue is that the meter and main breaker panel are one unit, and it does not look like there was an easy way to get the CTs on the main feeds.
So what you have is two sets of CTs that should be clamped over all the feeds from the breakers (one set came from the gateway). I'm guessing the amount of wires was too large for just one, so they used two sets. The 3rd set is measuring solar in the new box, yes?

There could be a couple sources of the issue, the CT could be backwards (possibly labeled backwards per @Shygar ). Or they joined them incorrectly (which can produce the same effect). CTs can be connected in series to provide a combined load value. If done, it should be white to black, white to black to keep label polarity the same (connecting one backwards is the same as installing it backwards, difference instead of sum). I see a couple wire nuts with pairs of white wires in the old box, that that might need checking.

So 2 sets for all loads (feed to gateway and all other un-backed up breakers) feeding one channel.
1 set for solar feeding the other channel

Be careful out there.

 

[mongo]

Here are some pics. I don't see CT sensors on the gateway busbars. They are at the top and the bottom of the main panel (the narrow one with meter). As I mentioned, those are 2 pairs, each twisted together, with top sensors apparently for the AC and convection oven in the house, and lower one going to our hot tub. The connection to the gateway seems to be via that same exit hole at the bottom, so I'm guessing they just forgot to loop in the big wire going over to the gateway.
View attachment 284984View attachment 284985View attachment 284986

Follow up: it is hard to tell from these shots, but they may have connected the CT's in parallel which would not be correct.
I think I see three wires in the nuts, which is the wrong way.
Wrong way: (two sets of this)
White wire from new panel (extension) <-> white wire from CT1 & White wire from CT2
Black wire from CT1 & Black wire from CT2 <-> Black wire back to new section (extension)

Right way is two sets of this:
White wire from new panel (extension) <-> white wire from CT1
Black wire from CT1 <-> White Wire from CT2
Black wire from CT2 <-> Black wire back to new section (extension)
For 6 wire nuts, I only see 4 which makes me think they put the CTs in parallel which is not going to work right.

 

[hulvey]

> they may have connected the CT's in parallel which would not be correct.

Do you have a pointer to any documentation to support a series rather than parallel configuration of the CTs to sum them? I dug around a bit on the Neurio site, but couldn't find anything.

I actually went out this afternoon and put the back-up sub-panel feed through the same CTs that my hot tub was going through, so assuming that the parallel thing is correct, the sum of all of these non-backed-up loads plus the sub-panel should match the grid input as there's nothing else coming out of the main panel. The "Grid" reading in the app now seems to match what I see on our meter outside, so I think it's working now.

The Tesla folks are supposed to come back next Tuesday to check it out, so they can double-check what I did, and probably neaten it up. I'm still not sure why they wouldn't just use one pair of CTs directly on the grid input rather than summing together all the output lines.

 

[mongo]

> they may have connected the CT's in parallel which would not be correct.

Do you have a pointer to any documentation to support a series rather than parallel configuration of the CTs to sum them? I dug around a bit on the Neurio site, but couldn't find anything.

I actually went out this afternoon and put the back-up sub-panel feed through the same CTs that my hot tub was going through, so assuming that the parallel thing is correct, the sum of all of these non-backed-up loads plus the sub-panel should match the grid input as there's nothing else coming out of the main panel. The "Grid" reading in the app now seems to match what I see on our meter outside, so I think it's working now.

The Tesla folks are supposed to come back next Tuesday to check it out, so they can double-check what I did, and probably neaten it up. I'm still not sure why they wouldn't just use one pair of CTs directly on the grid input rather than summing together all the output lines.

Hi,
Spent 10 minutes digging for a good CT tutorial, but only found industrial themed ones. In a nut shell, a CT is a transformer with a large winding count on the secondary (output) and single turn in the primary (sensed wire). This produces a voltage proportional to the primary current when a burden resistor is places across the secondary. On these CT that resistor is built in (safer), so you end up with a voltage output proportional to current. If you connect CT in series, these voltages sum (assuming correct polarity) and you get a final voltage signal equivilent to the sum of the primary currents. If you connect them in parallel, you get half the resistance (half the voltage output for same total current) and are also paralleling two inductors which will have different effects depending on the connections.

I could be wrong about how they are connected, like I said the pic is hard to determine.

If the new panel feed was not running through a CT, that was deffinately a/the problem for house load sensing.

Based on your meter/ panel setup, there is no way to get the CTs around the preinstalled feeds to the 175A main breaker. The feeds are shielded to make the panel safe to work on when mains are off and may be busbars themselves.

 

[hulvey]

I think it depends on whether the Neurio is measuring a current input (in which case parallel makes sense) or voltage (in which case series makes sense). I'm actually an MSEE, though I'm more of a microelectronics guy and not a "macroelectronics" guy. I've seen some current sensors that are called Rogowski coils, but my understanding is that those do not have metal cores. They are a coil of wire, but one end of the coil goes back through the center of the coil so that the 2 leads are on the same end of the coil, and the coil itself is wrapped around the conductor to be sensed.
Optimizing Performance from Rogowski Coil Current Transformers

Rogowski coils do output a voltage proportional to the current passing through the sensed conductor, so I'd guess those would need to be connected in series in order to sum 2 sensors. However, I know the Neurio sensors did have a metal core since I took them apart and shoved the sub-panel feeds in there with the hot-tub feeds that were already there. So, I believe the Neurio sensors are the metal-core transformer type of CTs.

I found at least a couple of sites where they seem to show CTs being connected in parallel:
Paralleling Current Transformers | Continental Control Systems
Placing Multiple Current Transformers in a Single Channel

One thing I'm not sure about is whether my hot-tub and the back-up sub-panel would be on the same phase. It seems to be giving me reasonable numbers now, though, so I think it's likely correct. The Tesla folks are still coming out next week, so I figure they can double-check my hacks.

 

[mongo]

I think it depends on whether the Neurio is measuring a current input (in which case parallel makes sense) or voltage (in which case series makes sense). I'm actually an MSEE, though I'm more of a microelectronics guy and not a "macroelectronics" guy. I've seen some current sensors that are called Rogowski coils, but my understanding is that those do not have metal cores. They are a coil of wire, but one end of the coil goes back through the center of the coil so that the 2 leads are on the same end of the coil, and the coil itself is wrapped around the conductor to be sensed.
Optimizing Performance from Rogowski Coil Current Transformers

Rogowski coils do output a voltage proportional to the current passing through the sensed conductor, so I'd guess those would need to be connected in series in order to sum 2 sensors. However, I know the Neurio sensors did have a metal core since I took them apart and shoved the sub-panel feeds in there with the hot-tub feeds that were already there. So, I believe the Neurio sensors are the metal-core transformer type of CTs.

I found at least a couple of sites where they seem to show CTs being connected in parallel:
Paralleling Current Transformers | Continental Control Systems
Placing Multiple Current Transformers in a Single Channel

One thing I'm not sure about is whether my hot-tub and the back-up sub-panel would be on the same phase. It seems to be giving me reasonable numbers now, though, so I think it's likely correct. The Tesla folks are still coming out next week, so I figure they can double-check my hacks.

I believe the difference is whether the CT has an internal burden resistor. If it does then it has a voltage output and gets connected In series. If not, it has a current output and gets connected In parallel. I assumed these were voltage output due to being safer in the event if the output being open.

If the hot tub heater/ pumps make the load reading go up,and the house loads do too, you have everything phased correctly.

 

[hulvey]

UPDATE: Tesla folks did come out and check it, and they are still in parallel. I was out of town, but my wife says they told her I had one wire backwards. I checked it again, and it looks the same as I recall, so I'm not sure what they changed. It seems to work.

 

[korinel]

I too am curious about the flexibility of the Neurio system within Powerwall. I'd love to be able to add an HPWC monitor too, ideally integrated with Powerwall's Neurio. If anyone comes up with a solution, please let me know.

 

[Shygar]

I too am curious about the flexibility of the Neurio system within Powerwall. I'd love to be able to add an HPWC monitor too, ideally integrated with Powerwall's Neurio. If anyone comes up with a solution, please let me know.

It's pretty easy, just make sure those wires run through your Neurio CT clips along with your other breakers, and it should measure it. Just keep in mind, if it measures it, it will drain the battery to feed it. So it all depends on if you want your powerwall to feed your car charging or not.

 

[miimura]

I sent Powerwall support a message about extending the monitoring outside the Gateway. My install would make running additional CTs into the main panel from the Gateway's Neurio difficult because of the 8 foot separation. Ideally, an additional Neurio unit could be added in the main panel and a RS-485 comms wire run between them. Configuration to add the values from the two Neurio units together to represent the Grid would also be necessary. I have no idea whether the scenario described here is workable.

 

[Shygar]

I sent Powerwall support a message about extending the monitoring outside the Gateway. My install would make running additional CTs into the main panel from the Gateway's Neurio difficult because of the 8 foot separation. Ideally, an additional Neurio unit could be added in the main panel and a RS-485 comms wire run between them. Configuration to add the values from the two Neurio units together to represent the Grid would also be necessary. I have no idea whether the scenario described here is workable.

You should be able to just extend the wires of the CT clips right? Or was there an issue with that?

 

[miimura]

You should be able to just extend the wires of the CT clips right? Or was there an issue with that?

I would have to add another CT pair to the same channel that is measuring the Grid currently. I would also have to extend the wire longer than what comes on the CTs. Is 8 feet too long for CT wires? I don't know. I would also have to go through the curved conduit that's already full of wires, so it will be a pain in the rear to fish the CT wires through there.

I suppose the other option is to move the existing Grid CTs to the main panel. The wires would still have to be extended, but there would only be one CT per input in that case. Does anybody know if it the Grid CT placement is important? My Gateway came from Tesla with the Grid CTs mounted onto the transfer switch.

The other consideration is that in the main panel it would be measuring current from a 200A breaker instead of the 125A breaker that currently feeds the Gateway. Do those CTs read over 100A? The Gateway is rated for 200A, so it shouldn't be a problem.

 

[Shygar]

Yea the CT's read over 100A. I have a non-Tesla Neurio in my other house and I put them on the main line coming into the panel, which is feeding my 200A main breaker and it's working fine.

On my powerwall setup, my grid CT's are spliced together with a set of CT's just for my AC and HPWC. So I have 3 sets of CT's for my one Neurio. Not sure if they need to do anything else. I would think 8 feet would be fine, but maybe you could email [email protected] and ask them the best option

 

[miimura]

Swell Energy came out to do some additional documentation on my system today. I had previously e-mailed Powerwall support about the configuration of my system (as mentioned above) and they passed the ticket to Swell Energy. While the tech was at my house he moved the Grid CTs from the grid input of the Gateway transfer switch to the main bus immediately after my main breaker. I have a 400 amp panel that can take two 200A main breakers. I only have one main breaker installed and it is connected to the breaker bus with thick round copper wire and the CT's were placed around these cables. The tech had an extension wire that already had the same connectors as the Neurio, so he just had to fish it through the conduit between the Gateway and the main panel. We did a test charging my Model 3 with varying current and watched the Powerwall page of the Tesla app update as the power flow changed. So, my Powerwall system now correctly measures the total grid draw and can offset but not back up the loads in my main panel.