Tesla Gateway registering solar at night 😊 from array 800 feet away

[JJ-Moonstone]

Tesla Gateway registering solar at night from array 800 feet away.
We are troubleshooting our system and would welcome input. Our setup has a 1.5 kWh pole-mount with micro inverters and a 11 kWh roof-mount SunPower with micro inverters, separated by 800 foot distance through the woods to the Tesla Gateway and Powerwalls. Due to the setup and with various loads on the property, the Gateway is getting inaccurate readings and thinks that we are producing solar at night! Our Tesla app shows this negative load because the CT’s are placed at the Tesla Gateway 800 feet away from solar panels: Ideally the CT’s should be placed on the lines next to the solar array. We didn’t think this negative load had consequences until recent extended power outages. The Powerwall registered it was nearly empty and repeatedly turned off and had to be reset when it was actually 96% full.

Our installer wants us to dig an 800 foot trench through trees and rock driveways to run an Ethernet line with extenders for communication between the solar at the house and the Tesla Gateway. We are trying to come up with a solution other than digging a trench!

We are looking into the option of adding a Neurio-meter for communication.
Both locations have their own Starlink for Wi-Fi to the cloud. Might a Neurio Meter be able to solve the communication problem?

Any other ideas or solutions? Thank you.

 

[aesculus]

It has nothing to do with the 800 feet but how your CTs are configured. Are the circuits that have the solar output only solar or do you have loads on them too (ie the gateway is measuring a remote sub panel with both loads and solar inverter breakers). If just solar then a simple CT switch should fix it. However ...

BTW I have a Tesla solar roof on a circuit that has both the inverter on it and lightly used loads. I had the same problem that you do. Somehow Tesla was able to put an additional CT on the load side for the same circuit and that corrected my problem. I am not sure what else they might have done on the server side to accommodate this or perhaps the GW just knows to throw away the negative loads it gets when its producing solar.

We tried various scenarios of loads and solar production and it all indicated correctly. Both the Tesla employee that was on site doing the work and myself were scratching our heads trying to understand what was happening, but the engineering team at Tesla had some magic going on.

NOTE: I had a Tesla PM as my liaison and did not go through product support so I am pretty sure we were deep into the technical side. He stated engineering was who was in the back end.

 

[dareed1]

Perhaps I don't understand your system, but the idea that the CTs should be located at the inverters does not make sense from an electrical or instrumentation point of view. CT stands for Current Transducer, and it measures the electrical current. The current is the same at the GW as it is 800 feet away. It's Kirchhoff's Law.

I'm also puzzled by the idea that measured solar currents can affect the PW reporting of charge state. They measure charge state (I think) by looking at the battery output voltage, and not by summing up currents in and out.

 

[JJ-Moonstone]

Perhaps I don't understand your system, but the idea that the CTs should be located at the inverters does not make sense from an electrical or instrumentation point of view. CT stands for Current Transducer, and it measures the electrical current. The current is the same at the GW as it is 800 feet away. It's Kirchhoff's Law.

I'm also puzzled by the idea that measured solar currents can affect the PW reporting of charge state. They measure charge state (I think) by looking at the battery output voltage, and not by summing up currents in and out.

Thanks Dareed1,
There’s lots to learn here for a layperson like myself. I understand that the reporting issues at our site are related to multiple loads as well as the two separate solar arrays and all spread over a couple of acres. I agree what you say about Kirchhoff’s Law. Tesla has made some remote adjustments and our system is at least running smoothly again. Thanks!

 

[JJ-Moonstone]

It has nothing to do with the 800 feet but how your CTs are configured. Are the circuits that have the solar output only solar or do you have loads on them too (ie the gateway is measuring a remote sub panel with both loads and solar inverter breakers). If just solar then a simple CT switch should fix it. However ...

BTW I have a Tesla solar roof on a circuit that has both the inverter on it and lightly used loads. I had the same problem that you do. Somehow Tesla was able to put an additional CT on the load side for the same circuit and that corrected my problem. I am not sure what else they might have done on the server side to accommodate this or perhaps the GW just knows to throw away the negative loads it gets when its producing solar.

We tried various scenarios of loads and solar production and it all indicated correctly. Both the Tesla employee that was on site doing the work and myself were scratching our heads trying to understand what was happening, but the engineering team at Tesla had some magic going on.

NOTE: I had a Tesla PM as my liaison and did not go through product support so I am pretty sure we were deep into the technical side. He stated engineering was who was in the back end.

Hi Aesculus. Yes, the circuits with the solar output have loads on them too which causes the complication. Initially we tried switching and then replacing the CT’s, but to no avail. It sounds like your Tesla PM came up with a good solution - it makes sense to measure the load side for the same circuit, but in our case all of that is 800 feet away too.
Currently today we are considering digging that 800 foot trench…. the installer would be using a low-tech hardwired solution that should work. He said the solar CT requires two wires, and he uses 16awg wire when it’s all fairly close. With the 800’ he would use (2) 12awg wires pulled through a communication conduit to make the solar CT connection at the house.
At first we were fine with having this inconsistent data showing a negative load, as it didn’t seem to affect the functionality of the system on a day to day basis, but during the extended power outage at the weekend we found it did. All sorted now, Tesla techs did seem to work some behind the scenes magic too which stopped the PW from turning off. Overall of course we love being powered by the sun and less reliant on the grid!
Thanks so much for your input.

 

[aesculus]

Hi Aesculus. Yes, the circuits with the solar output have loads on them too which causes the complication. Initially we tried switching and then replacing the CT’s, but to no avail. It sounds like your Tesla PM came up with a good solution - it makes sense to measure the load side for the same circuit, but in our case all of that is 800 feet away too.
Currently today we are considering digging that 800 foot trench…. the installer would be using a low-tech hardwired solution that should work. He said the solar CT requires two wires, and he uses 16awg wire when it’s all fairly close. With the 800’ he would use (2) 12awg wires pulled through a communication conduit to make the solar CT connection at the house.
At first we were fine with having this inconsistent data showing a negative load, as it didn’t seem to affect the functionality of the system on a day to day basis, but during the extended power outage at the weekend we found it did. All sorted now, Tesla techs did seem to work some behind the scenes magic too which stopped the PW from turning off. Overall of course we love being powered by the sun and less reliant on the grid!
Thanks so much for your input.

Again. The 800 feet has nothing to do with it. You don't need any additional cabling etc. Mine is about 600 feet away FYI and I was going to try to find a way to remote monitor the inverter too, but Tesla came up with a better solution.

Just get someone in the depths of Tesla support to into engineering and explain that you are trying to monitor a remote sub panel with both loads and solar on it.

Note I have a GW 1 so not sure what you have or if you have the same options but ...

For the wires coming from the remote load/solar sub panel they have CTs on the solar inputs to the GW and put an additional set of CTs on the home/load side. Not sure if they did anything in the back end but physically this was it from my premise perspective.

And when Tesla said they were ready we tried every combination of load and solar we could and it all showed correctly.

 

[CrazyRabbit]

I don’t like the idea of a 800 foot CT extension cable, but perhaps that is the correct solution.
Perhaps using a directional wifi antenna on the neurio would help with wifi connection back to gateway. Neurio can also be connected with rs232 cable, but that probably has length limitations too.

 

[JJ-Moonstone]

Again. The 800 feet has nothing to do with it. You don't need any additional cabling etc. Mine is about 600 feet away FYI and I was going to try to find a way to remote monitor the inverter too, but Tesla came up with a better solution.

Just get someone in the depths of Tesla support to into engineering and explain that you are trying to monitor a remote sub panel with both loads and solar on it.

Note I have a GW 1 so not sure what you have or if you have the same options but ...

For the wires coming from the remote load/solar sub panel they have CTs on the solar inputs to the GW and put an additional set of CTs on the home/load side. Not sure if they did anything in the back end but physically this was it from my premise perspective.

And when Tesla said they were ready we tried every combination of load and solar we could and it all showed correctly.

Thank you again - is there any possibility your PM can let us know even the first name of the Tesla (magician) tech who helped you? We have been back and forward with them since last June and made no headway until now. That would be so helpful for us.

 

[JJ-Moonstone]

Again. The 800 feet has nothing to do with it. You don't need any additional cabling etc. Mine is about 600 feet away FYI and I was going to try to find a way to remote monitor the inverter too, but Tesla came up with a better solution.

Just get someone in the depths of Tesla support to into engineering and explain that you are trying to monitor a remote sub panel with both loads and solar on it.

Note I have a GW 1 so not sure what you have or if you have the same options but ...

For the wires coming from the remote load/solar sub panel they have CTs on the solar inputs to the GW and put an additional set of CTs on the home/load side. Not sure if they did anything in the back end but physically this was it from my premise perspective.

And when Tesla said they were ready we tried every combination of load and solar we could and it all showed correctly.

Hi. Thanks for encouragement. Here is our energy storyboard diagram to give you a visual, just to clarify the different inputs and loads on our system. I think we have Gateway 2, so maybe it is different than GW1, but we will certainly pursue your suggestion and keep calling Tesla support until we get someone who understands the situation and possible solution as they figured out for you.

Thank you.

 

[JJ-Moonstone]

I don’t like the idea of a 800 foot CT extension cable, but perhaps that is the correct solution.
Perhaps using a directional wifi antenna on the neurio would help with wifi connection back to gateway. Neurio can also be connected with rs232 cable, but that probably has length limitations too.

Yes, thanks. I’ll post back the end results and final solution when resolved - ever the optimist!

 

[aesculus]

Thank you again - is there any possibility your PM can let us know even the first name of the Tesla (magician) tech who helped you? We have been back and forward with them since last June and made no headway until now. That would be so helpful for us.

Sorry but he no longer works for Tesla.

 

[aesculus]

Hi. Thanks for encouragement. Here is our energy storyboard diagram to give you a visual, just to clarify the different inputs and loads on our system. I think we have Gateway 2, so maybe it is different than GW1, but we will certainly pursue your suggestion and keep calling Tesla support until we get someone who understands the situation and possible solution as they figured out for you.

Well yours is more split than mine was but I have also seen this before here on TMC but could not help you find it.

None the less I think you will find a solution. I am kind of surprised they put the PW and GW there rather than at the house. Perhaps your 800 foot conduit was already in place and at max?

 

[BGbreeder]

I don’t like the idea of a 800 foot CT extension cable, but perhaps that is the correct solution.
Perhaps using a directional wifi antenna on the neurio would help with wifi connection back to gateway. Neurio can also be connected with rs232 cable, but that probably has length limitations too.

While RS-232 can go for 3000' or so at 2400baud, farther with inline amplifiers, are you sure Tesla neurios have that feature enabled? I ask only because when our installers were thrashing around trying to find a solution that worked for our distributed consumption and production system, RS-232 never came up.

All the best,

BG

 

[CrazyRabbit]

Yes, but you can’t do firmware updates over serial bus for whatever reason.

 

[CrazyRabbit]

OP, All that on 200 amp service or is there a non backup panel.

 

[JJ-Moonstone]

OP, All that on 200 amp service or is there a non backup panel.

There is a non backup panel next to Gateway (in my sketch = energy distribution box! ) and as we have propane for heating water/cooking, our system is just 150 amps.

 

[JJ-Moonstone]

Well yours is more split than mine was but I have also seen this before here on TMC but could not help you find it.

None the less I think you will find a solution. I am kind of surprised they put the PW and GW there rather than at the house. Perhaps your 800 foot conduit was already in place and at max?

Yeah, you are right! It was the best solution, with all given dynamics here. I will search around for similar situations on TMC. Did you read comment #6 above in this thread? It does sound like other folks have similar situations and sometimes Tesla engineers come up with unique solutions when they can. Thanks!

 

[JJ-Moonstone]

Again. The 800 feet has nothing to do with it. You don't need any additional cabling etc. Mine is about 600 feet away FYI and I was going to try to find a way to remote monitor the inverter too, but Tesla came up with a better solution.

Just get someone in the depths of Tesla support to into engineering and explain that you are trying to monitor a remote sub panel with both loads and solar on it.

Note I have a GW 1 so not sure what you have or if you have the same options but ...

For the wires coming from the remote load/solar sub panel they have CTs on the solar inputs to the GW and put an additional set of CTs on the home/load side. Not sure if they did anything in the back end but physically this was it from my premise perspective.

And when Tesla said they were ready we tried every combination of load and solar we could and it all showed correctly.

Are we correct in interpreting that you have a total of four CT’s in the Gateway? Thanks!

 

[aesculus]

Are we correct in interpreting that you have a total of four CT’s in the Gateway? Thanks!

From pictures I have four sets (1- Backup load main, 1- bundled around 2 separate solar circuits, 1 - wall adapter outside of the PW, 1- on my combo sub panel). I was looking for a picture that would give me details about how they are hooked up but its not clear. IIRC they had put some sort of splitter on that set of CTs so it could be monitored by both the solar side and the home load side. That is the mystery.

EDIT: The load panel was already being monitored as a collection which included the combo sub panel with solar. They added another set of CTs for that breaker and routed them to the solar pickup on the GW (see picture). Prior to that I would get negative loads on my house side if very little was being used and the solar was producing. That never happens now.

 

[Gwgan]

Gateway is getting inaccurate readings and thinks that we are producing solar at night!

“Producing solar at night” has been a flag for an inverted CT in many stories I’ve read and saw once myself. I recommend you insist Support review all the CT settings in the Gateway.