Powerwalls installed with undersized wiring.

[SSonnentag]

OK, it's a bit of a clickbait title, but it's true. about 6 weeks ago our 3 Powerwalls became unresponsive and started flashing their green LEDs in a cyclic manner. I went through diagnostics and such with technicians over the phone, but it was ultimately decided that an on-site visit had to be made. The first appointment was 5 weeks out, which was yesterday. After the technicians poked around for a couple of hours, they have determined that the communication wiring (22 gauge) is too small to get a satisfactory voltage through from the units to the communication module. Apparently a minimum of 10.5 V is required for reliable communication, and we're only getting 10 V. So now we're waiting for another 2-3 weeks (estimated) before the team can come back out with larger wiring for a redo. The tech claims this has been seen on multiple installations and that, to his knowledge, the larger wiring has been working.

I also mentioned the ramping up to 66 Hz that we see, and which causes our induction stove to go berserk. He seemed to think the Powerwalls should only vary between 59 and 61 Hz, and promised to look into the situation once the rewiring job has been complete. We shall see. I will be on site with my frequency meter to verify the results.

 

[wwhitney]

For reference, how long is the communications wire in your installation?

Cheers, Wayne

 

[astrorob]

thanks for that info, maybe i can mention this to my installers.

as for the frequency, i was under the impression that older inverters might need the higher frequency to turn off, whereas the newer 'intelligent' inverters will cut out at lower out-of-spec frequencies. is that not the case? IOW the 66Hz thing might be intentional, or is at least a conservative setting.

 

[SSonnentag]

For reference, how long is the communications wire in your installation?

Cheers, Wayne

I'd estimate it to be around 25 feet (8 m).

 

[SSonnentag]

thanks for that info, maybe i can mention this to my installers.

as for the frequency, i was under the impression that older inverters might need the higher frequency to turn off, whereas the newer 'intelligent' inverters will cut out at lower out-of-spec frequencies. is that not the case? IOW the 66Hz thing might be intentional, or is at least a conservative setting.

Possibly. Our SMA Sunny Boy inverters are around 10 years old (2009). I don't know what their synch cutoff frequency is, but hopefully it's far lower than 66 Hz. The specs I have only list the inverter output frequency range (59.3 Hz - 60.5 Hz).

Edit: Now that I think about it, since the output frequency range stops at 60.5 Hz, it is impossible for the inverters to sync up and produce power above 60.5 Hz, which indicates to me that 61 Hz from the Powerwalls should cause the inverters to disconnect.

 

[astrorob]

i have a ca. 2017 sunny boy SMA5000; i sure hope that it also will cut out at ~61Hz as well. these are all things i'm going to have to get right at installation time, cause from the sound of it if there is a problem they won't be back for months and months.

 

[jboy210]

Forget the silly questions, but how is this communication wiring run? Between Powerwall and inverter, to some other device and to Powerwall, or ??? Also, what is the purpose of the communication wiring?

 

[woferry]

The Powerwall2 seems to be just a battery pack and inverter (makes sense when you put multiple PW's together, only duplicate what needs to be duplicated). All of the "brains" to determine how much these inverter(s) should be consuming or producing lives in what's called the Gateway (or Tesla Energy Gateway, TEG), which is a separate box. So the Powerwalls have a 4-wire comms cable that connects all of them to the Gateway (I think someone determined that it's a CAN bus, plus presumably a ground and DC power wire). The Gateway gets the state of charge from the battery packs and controls the PW inverters over this cable. The cable is physically run through the same conduit as the AC wires, using appropriately-certified wiring for that application (you can't just use any wire when it's run in the same conduit as power lines, whether PV or AC, per codes).

The Gateway does not connect to the solar inverters at all (other than via the AC power circuits), but it does monitor the solar inverter output using a pair of Current Transducers (CTs). A second pair of CTs monitors the current going to or coming from the grid. It seems the Gateway derives the "Home" power from whatever is left over considering the grid current, solar current, and what it's instructed the PW inverters to output/input. And depending on the mode the system is configured to, the Gateway seems to largely be playing a game of making the PW output either zero-out the grid consumption, or to cause the grid production to match the solar production (i.e. PW powers the house, all of the solar goes to the grid).