Tesla Solar, Powerwall, Gateway System Problem

[skullalpha]

Posting this problem that I am facing after installation for others to know and follow:
Having issues with Tesla Gateways resetting and power interuption.

Current system:
2 gateways, 4 powerwall, 24kw solar array
2 powerwalls are for Main House
2 powerwalls are for Guest Suite

Problem:
When the main house’s powerwalls and guest suite’s powerwalls are being charged by solar panels while connected to the grid; the 2 gateways glitches (house power goes off and then on) and reset themselves many times over the span of charging cycle. Glitch is noticeable because a momentary power outage occurs.

A) No it’s not power grid surges
B) Pattern is established and is reproduceable
C) Only happens when all the following conditions are true:

1. Main House’s Powerwalls are under 100%
2. Guest Suite’s Powerwalls are under 100%
3. Main House’s Solar Panels are charging Powerwalls
4. Guest Suite’s Solar Panel are charging Powerwalls
5. Both system is On-Grid

Workaround Util they fix this:
I put the guest suite's system on Off-Grid at night. Let the that part of the home run offgrid until the powerwalls are at 100%.

 

[Matt-FL]

This happens to both gateways simultaneously? If so, sounds like something might not be wired correctly.

 

[Vines]

Are the CT's for the 2 separate systems separated, and also do not overlap each other electrically?

I'd wonder if the Powerwalls are both trying to offset the same loads?

 

[Vines]

Also, what does your service panel configuration look like? Is it a 400A or 200A service?

 

[skullalpha]

This happens to both gateways simultaneously? If so, sounds like something might not be wired correctly.

When power glitches, both gateways are affected and gets into a reset cycle.

 

[skullalpha]

Also, what does your service panel configuration look like? Is it a 400A or 200A service?

Main House 1 Gateway, 2 Powerwall, 2 Inverter, 1 200amp Panel,
Guest Suite 1 Gateway, 2 Powerwall, 1 Inverter, 1 200amp Panel

All loads are separated, meaning there is no overlap. At least that was the way it was before Tesla install.

 

[skullalpha]

Are the CT's for the 2 separate systems separated, and also do not overlap each other electrically?

I'd wonder if the Powerwalls are both trying to offset the same loads?

I would assume the current transformers are separate and should not overlap. Odd thing is the design of the 2 systems should be completely segregated. 1/3 of my solar panels are dedicated to guest suite system and the other 2/3 of the panels are hooked into the main house.

 

[BGbreeder]

That scenario would drive me bananas. Have you had a DVM or oscilloscope on the various lugs to check for voltage surges or other transients?

Without data, and as a complete armchair quarterbacking, my guess would be either you have a ground / neutral issue either internal to your panels/gateways or your transformer, or alternatively possibly a loose connection, again probably at the gateway.

There was a small issue during our install of a Tesla wiring lug that was tight, but not centered correctly on the bus bar. (Hard to describe, but basically the lug had two tabs projecting into the gateway to bracket the bus bar. I think as an anti-rotation element. It turns out that it is possible to tighten the lug askew on the bus bar. Don't ask.)

I would leave open the possibility that there is an odd harmonic between your two gateways, but that would not be my first guess as there are more than a few people here with dual gateways.

All the best,

BG

 

[GregBallantyne]

So if these two systems are physically separate from a hardware perspective, that leaves influences between the two that do not require a physical connection such as thru wifi or network connecton, Bluetooth, or EMFs. I'm a bit curious if there is spatial separation..... but it sounds like the panels are mounted as one array, just not via wiring. Or an external influence that is affecting both systems, but the workaround probably eliminates that possibility....

 

[BGbreeder]

So if these two systems are physically separate from a hardware perspective, that leaves influences between the two that do not require a physical connection such as thru wifi or network connecton, Bluetooth, or EMFs. I'm a bit curious if there is spatial separation..... but it sounds like the panels are mounted as one array, just not via wiring. Or an external influence that is affecting both systems, but the workaround probably eliminates that possibility....

To be clear the OP outlined that only the control circuits are not overlapping. They are, however, not physically separated: both are connected to each other and the grid electrically, which leaves open a number of avenues for feedback between the systems.

 

[GregBallantyne]

To be clear the OP outlined that only the control circuits are not overlapping. They are, however, not physically separated: both are connected to each other and the grid electrically, which leaves open a number of avenues for feedback between the systems.

Ahh, and a bit more likely than a wifi or similar issue.....

 

[skullalpha]

What @BGbreeder said above makes a lot of sense. The system is conjoined at the grid meter.

46 Solar Panels to 7 Strings to Inverter 1 and 2 to charge Powerwalls 1 and 2
14 Solar Panels to 2 Strings to Inverter 3 to charge Powerwalls 3 and 4

So what i don't get is that the system only glitches when solar is feeding both sets of powerwalls, while feeding both houses with power while being on the grid.

The same scenario above with one or both systems off grid does not result in any power glitches.
Even when the gateways are off grid; isn't the system still conjoined at the grid meter?

 

[BGbreeder]

@skullalpha I could spin many tales of possibles, but I would start with loose connections to power / neutral / ground somewhere, but I certainly leave open the possibility that there is an unusual impedance on your grid feeder that is ok with either one of your systems, but not with both. (Basically there might be a harmonic that gets overdriven with both gateways, like a microphone feedback, and that would easily take your two systems plus what is out on "your" transformer, or the transformer itself to make it happen.)

Maybe one of your Powerwall has a funky DAC in the AC forming circuit that is OK with its buddy, but not everything else, maybe it is made worse by an unstable neutral...The thing about digitally generated AC is that it can push pretty strongly on the rest of the system, especially with quad powerwalls. Tesla requires that the line impedance gets monitored before installation and imposes limits on the power quality as the number of powerwalls increases.

If the power / neutral / ground connections. and ground impedances all check out, I would be tempted to do some power quality monitoring at your meter and work backwards toward your gateways, solar, and major appliances. Data logging power quality monitors have gotten a lot cheaper over the years. To me the great thing about your issue is that it is so reproducible (hats off to you for nailing that down). (On the subject of major appliances, have you tried taking your HVAC / EVs / pool pump(?) off line to see if the issue changes?)

Good luck!

BG

 

[skullalpha]

Update: @BGbreeder, looks like your guesses might be right. Finally got Tesla guys out here and they suspect the impedance threshold on their equipment maybe too low. They will not be able to adjust until they have special equipment to measure it for proof. Hopefully a solution will be had by end of this weekend. I will post more as this whole thing plays out.