Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

'whole home backup' confusion

This site may earn commission on affiliate links.
I recently had a system installed (16kW, 2PW+, backup gateway 2) and all of my paperwork references 'whole home backup'. Basically, they pulled all of my loads from my main panel into a subpanel connected to the gateway. The only breaker in my main panel now is a 100A going to the gateway. What's odd to me is that they had me upgrade my main panel and service (which required significant trenching and $$$) from 100A to 200A, but they're only feeding this subpanel (which is capable of 225A) with 100A. There are several new loads I'd like to add in the coming months (EV chargers, electric dryer, water heater), which my project advisor was aware of, but there doesn't seem to be the bandwidth to support these in the subpanel. Am I mistaken? Should I ask Tesla to update the wiring/breaker to 200A?

My understanding (which may be wrong!) is that in order to be powered by my PV system or PWs when the grid is out, loads would need to be in the subpanel as no electricity would flow to the main.

I've included the wiring diagram below. Any insight would be much appreciated.

Thanks!


n91ktoodr6i71.png
 
This is the panel I have: Square D Homeline 200 Amp 30-Space 42-Circuit Outdoor Ring-Type Surface Mount Solar-Ready Main Breaker Plug-On Neutral CSED-SC3042M200PS - The Home Depot
It is rated for 200amp main breaker with 225 bus bar. Before the install it had 125amp breakers in there feeding to my subpanel. Now it only has one 110amp breaker to the gateway.
I am 99% certain that your panel can also take the 4 pole 200A branch circuit breaker. It's hard to read the sticker on that picture, but usually those types of CSED will work. It's the very narrow ones with a center-fed bus and not enough room for the large wire that usually won't take larger than a 100 or 125A branch circuit breaker.
 
Upvote 0
In my interpretation, you need a main breaker inside the Tesla Gateway if you use subfeed lugs in the main panel.
If the service panel is emptied of all distribution breakers, and the feeder from the subfeed lugs has an ampacity properly protected by the main breaker (including the allowance of 2017 NEC 310.15(B)(7)), then there's no need for a main breaker inside the Gateway.

Cheers, Wayne
 
  • Informative
Reactions: BGbreeder
Upvote 1
I am 99% certain that your panel can also take the 4 pole 200A branch circuit breaker. It's hard to read the sticker on that picture, but usually those types of CSED will work. It's the very narrow ones with a center-fed bus and not enough room for the large wire that usually won't take larger than a 100 or 125A branch circuit breaker.
ya the sticker on the side does mention it can take in 200amp quad breakers. however the plan submitted to the city (which they wouldn't let me see until after the plans were approved) only list it at 100a. I talked to both the solar crew and the powerwall crew electricians and they all said they that the plans are correct and the gateway cannot take in more power, which just seems wrong.
 
Upvote 0
If the service panel is emptied of all distribution breakers, and the feeder from the subfeed lugs has an ampacity properly protected by the main breaker (including the allowance of 2017 NEC 310.15(B)(7)), then there's no need for a main breaker inside the Gateway.

Cheers, Wayne


Dude, can you just start your own solar and ESS company? Like why is it you know all this stuff but the company I had to go through with like dozens of experts is like 1/100th your level?

I have three 200A breakers wired in series between my home loads and the PoCo meter (with an extra disconnect in the middle just in case). I tried to actually get them to save money by removing some of these, but each time they found some weird code interpretation to add them back.

I was told the TEG2 had to be bonded neutral to ground. Which then meant a 200A Eaton CSR style thing absolutely had to be in the Gateway. I feel like both of these assertions from Sunrun are false... but then Sunrun doesn't employ enough Wayne's overseeing their design team.
 
Upvote 0
If the service panel is emptied of all distribution breakers, and the feeder from the subfeed lugs has an ampacity properly protected by the main breaker (including the allowance of 2017 NEC 310.15(B)(7)), then there's no need for a main breaker inside the Gateway.

Cheers, Wayne
In this case, would you say the 100% rule sticker needs to go on the panel? Subfeed lugs are part of the distribution bus but labeled according to 705.12.B.2.3. wouldn't really make sense. What prevents someone from adding breakers to that busbar later, which could overload that bus in the main service?

Dude, can you just start your own solar and ESS company? Like why is it you know all this stuff but the company I had to go through with like dozens of experts is like 1/100th your level?

I have three 200A breakers wired in series between my home loads and the PoCo meter (with an extra disconnect in the middle just in case). I tried to actually get them to save money by removing some of these, but each time they found some weird code interpretation to add them back.

I was told the TEG2 had to be bonded neutral to ground. Which then meant a 200A Eaton CSR style thing absolutely had to be in the Gateway. I feel like both of these assertions from Sunrun are false... but then Sunrun doesn't employ enough Wayne's overseeing their design team.
The TEG should not have the Neutral to Ground Bond (Removal per page 21 of the manual), unless it is the service equipment like if you started with a meter socket only in a typical east coast installation, instead of a combo panel/CSED like a typical west coast installation.

Neutral to ground should only be bonded once per structure, or you risk a ground loop. The code says that bond happens in whatever enclosure is the first one with an overcurrent device, after the meter. So in your case, the main CSED has the Main Bonding Jumper, and the main breaker. There is no sense putting an overcurrent device at both ends of a single wire, as long as its not tapped and therefore has a fork in it. There is also no sense putting the main breaker in the TEG, if the MSP has a branch circuit breaker in it feeding the TEG.
 
  • Like
Reactions: BGbreeder
Upvote 0
In this case, would you say the 100% rule sticker needs to go on the panel? Subfeed lugs are part of the distribution bus but labeled according to 705.12.B.2.3. wouldn't really make sense. What prevents someone from adding breakers to that busbar later, which could overload that bus in the main service?
The 2017 NEC, the basis of the 2019 CEC, basically overlooked the issue of subfeed lugs in 705.12(B). There was some language added in the 2020 NEC, but I find that new language pretty unclear. We'll see what the 2023 NEC brings, the first draft language is somewhat clearer. And its language would support your idea that the Gateway should have a main breaker installed, which would allow any of the 3 usual busbar protection methods.

As a practical matter, for a 200A meter/main/distribution with just 200A feed thru lugs in it, the only way to overload that 200A main busbar would be to add over 200A of load to it (or interconnect another power source). That means the 100% warning label would be appropriate, electrically (as long as the breakers serve loads only). It would only be compliant NEC-wise if you want to take advantage of the loop hole that the 2017 NEC doesn't address feed-thru lugs.

In reality what would is desired in these cases is to turn that meter/main/distribution into just a meter/disconnect. So you could slap a label on there that says "hey, no loads here." Or more emphatic would be to rivet an appropriately thick piece of galvanized sheet metal (painted grey for best effect) over the dead front to cover all the breaker spaces except for the feed thru lug location, and then put the label on that. That would also be a nice way to repair the dead front if there used to be a lot of breakers in the panel, rather than having a bunch of little plastic fillers in the dead front.

Cheers, Wayne
 
Upvote 0
In this case, would you say the 100% rule sticker needs to go on the panel? Subfeed lugs are part of the distribution bus but labeled according to 705.12.B.2.3. wouldn't really make sense. What prevents someone from adding breakers to that busbar later, which could overload that bus in the main service?


The TEG should not have the Neutral to Ground Bond (Removal per page 21 of the manual), unless it is the service equipment like if you started with a meter socket only in a typical east coast installation, instead of a combo panel/CSED like a typical west coast installation.

Neutral to ground should only be bonded once per structure, or you risk a ground loop. The code says that bond happens in whatever enclosure is the first one with an overcurrent device, after the meter. So in your case, the main CSED has the Main Bonding Jumper, and the main breaker. There is no sense putting an overcurrent device at both ends of a single wire, as long as its not tapped and therefore has a fork in it. There is also no sense putting the main breaker in the TEG, if the MSP has a branch circuit breaker in it feeding the TEG.


Lol can you please also get a job at sunrun so you call tell their designers what the heck they ought to do?

My house now has two grounding rods… the MSP is neutral bonded to the original ground, and the downstream TEG 2 is bounded to a new grounding rod they drilled in.

The inspector was like… why??? But got fixated about the disconnects and just shrugged at all the breakers and stuff.

Both grounding rods are within 36” of my gas riser so I’m expecting to die in a big 💥 if there is an earthquake and lightning strike at the same time. shrapnel from the disconnects will be flying everywhere
 
Upvote 0
In reality what would is desired in these cases is to turn that meter/main/distribution into just a meter/disconnect. So you could slap a label on there that says "hey, no loads here." Or more emphatic would be to rivet an appropriately thick piece of galvanized sheet metal (painted grey for best effect) over the dead front to cover all the breaker spaces except for the feed thru lug location, and then put the label on that. That would also be a nice way to repair the dead front if there used to be a lot of breakers in the panel, rather than having a bunch of little plastic fillers in the dead front.

Cheers, Wayne

@Vines can I get a galvanized piece of sheet metal with a laser etched “these are not the droids or breakers you’re looking for since there’s nothing here except a neutral bonded to ground?”
 
  • Like
Reactions: X-pilot
Upvote 0
In reality what would is desired in these cases is to turn that meter/main/distribution into just a meter/disconnect. So you could slap a label on there that says "hey, no loads here." Or more emphatic would be to rivet an appropriately thick piece of galvanized sheet metal (painted grey for best effect) over the dead front to cover all the breaker spaces except for the feed thru lug location, and then put the label on that. That would also be a nice way to repair the dead front if there used to be a lot of breakers in the panel, rather than having a bunch of little plastic fillers in the dead front.
Interesting. My home's installation will have the same scenario where all the branch circuits relocated from the main panel so it only contains a backfeed breaker to the Gateway. I was having an offline conversation with @Vines to ask if using a piece of sheet metal like you suggest would be allowed, but he said the little plastic fillers were the sure method because that's what is specified in the code.
 
Upvote 0
Interesting. My home's installation will have the same scenario where all the branch circuits relocated from the main panel so it only contains a backfeed breaker to the Gateway. I was having an offline conversation with @Vines to ask if using a piece of sheet metal like you suggest would be allowed, but he said the little plastic fillers were the sure method because that's what is specified in the code.
I am interested in your opinion of this approach as well.
 
Upvote 0
I am interested in your opinion of this approach as well.
I assume this question was not directed to me, as you were quoting someone else.

But I'm curious to ask if as a practical matter when emptying a service panel to turn it into a service disconnect only (at least functionally) with a feeder to a Backup Gateway, whether it is easier to source 200A plug-on feed thru lugs rather than an actual 200A breaker (presumably both are 4 position devices). As if not, 705.12(B)(2)(3)(c) is clearly complied with when using a 200A breaker, so that would be the regulatorily easier solution, even though there's no need for that wire to bus connection to have OCPD that duplicates the service panel's main OCPD.

Cheers, Wayne
 
Upvote 0
So, with a whole home backup routed through a West coast type main service panel with a 200A service disconnect and a single 200A breaker, which one is the recommended device to shut off power; the service disconnect to the nearly empty main panel, or the sole breaker in the main panel, the 200A breaker to the gateway? Surely there is an NEC logic to this...

I can't believe the cost and wall space being used for these Powerwall installs because a direct meter feed to a gateway is not permitted west of where, the Mississippi? Don't mind me...

All the best,

BG
 
Upvote 0
I assume this question was not directed to me, as you were quoting someone else.

But I'm curious to ask if as a practical matter when emptying a service panel to turn it into a service disconnect only (at least functionally) with a feeder to a Backup Gateway, whether it is easier to source 200A plug-on feed thru lugs rather than an actual 200A breaker (presumably both are 4 position devices). As if not, 705.12(B)(2)(3)(c) is clearly complied with when using a 200A breaker, so that would be the regulatorily easier solution, even though there's no need for that wire to bus connection to have OCPD that duplicates the service panel's main OCPD.

Cheers, Wayne
Sorry Wayne, my quote skills missed the mark. I am interested to know what you think about modifying a deadfront with sheet metal to cover a bunch of holes where breakers used to be. The plastic covers arent ideal though listed, but what is your experience with inspectors, do they have an issue with this?

To answer your question we choose the 4 pole breakers, as they seem to be readily available. When we do use a lug connection it is usually to a panel that came with lug connections from the factory.

I totally agree that having a pair of 200A breakers which basically do the same thing is pretty redundant, but it keeps us 100% within the code.
 
Upvote 0
Sorry Wayne, my quote skills missed the mark. I am interested to know what you think about modifying a deadfront with sheet metal to cover a bunch of holes where breakers used to be. The plastic covers arent ideal though listed, but what is your experience with inspectors, do they have an issue with this?
As far as experience with inspectors, I'm the wrong person to ask; I've only dealt with a handful of inspectors for one AHJ (the building department in the city I live in).

But Article 312 covers Cabinets (enclosures with a door, like for a panelboard), and there is no requirement that metal cabinets be listed. You can buy a panelboard (the guts of a load center) and put it in your own metal enclosure as long as the spacing and construction requirements of Article 312 are met. So if you use a suitably corrosion resistant, suitably thick (at least 53 mils) piece of sheet metal, there's no question that the cabinet deadfront modification is compliance with the NEC.

The question that arises is whether you can therefore claim compliance with 705.12(B) because the equipment no longer "is capable of supplying multiple branch circuits or feeders, or both." I would say yes. ButI also think that even if you don't modify the deadfront, a main service panel with only a set of feed thru lugs complies with 2017 NEC 705.12(B)(2)(3)(c) both in letter and in spirit. The fact that 705.12(B)(2)(3) doesn't address feed thru lugs does leave you open to an AHJ saying that other rules should apply, but only a pig-headed AHJ would try to fail the installation. Unfortunately, I understand that those exist. : - )

Cheers, Wayne

P.S. A similar discussion came up at NEC 2017 705.12(B)(2)(3)(c)
 
  • Helpful
Reactions: BGbreeder
Upvote 0
Well I got an email saying they are doing a main breaker switch and they set up an appointment for next week. I thought it was good news, but I called my advisor and he said they are switching out the breaker feeding the gateway from 110amps to 100amps as per the plans. He said he's going to ask the design team to see if it can be switched out to 200amps feeding the gateway or at the very least 125amps feeding the breaker.

Tesla never told me what loads where backed up until a couple days before the install when they sent over the approved permit plans. I just assumed that since I had a 200amp main service and that the gateway can support 200amp, everything would be able to be backed up while still pulling 200amps from the grid.
 
  • Funny
Reactions: Vines
Upvote 0
Well I got an email saying they are doing a main breaker switch and they set up an appointment for next week. I thought it was good news, but I called my advisor and he said they are switching out the breaker feeding the gateway from 110amps to 100amps as per the plans. He said he's going to ask the design team to see if it can be switched out to 200amps feeding the gateway or at the very least 125amps feeding the breaker.

Tesla never told me what loads where backed up until a couple days before the install when they sent over the approved permit plans. I just assumed that since I had a 200amp main service and that the gateway can support 200amp, everything would be able to be backed up while still pulling 200amps from the grid.

The cost of isolating the real designers from the customers is issues like this. Surely its the cheapest way to do things but does it serve the end goal?
 
  • Like
Reactions: hydro
Upvote 0
As far as experience with inspectors, I'm the wrong person to ask; I've only dealt with a handful of inspectors for one AHJ (the building department in the city I live in).

But Article 312 covers Cabinets (enclosures with a door, like for a panelboard), and there is no requirement that metal cabinets be listed. You can buy a panelboard (the guts of a load center) and put it in your own metal enclosure as long as the spacing and construction requirements of Article 312 are met. So if you use a suitably corrosion resistant, suitably thick (at least 53 mils) piece of sheet metal, there's no question that the cabinet deadfront modification is compliance with the NEC.

The question that arises is whether you can therefore claim compliance with 705.12(B) because the equipment no longer "is capable of supplying multiple branch circuits or feeders, or both." I would say yes. ButI also think that even if you don't modify the deadfront, a main service panel with only a set of feed thru lugs complies with 2017 NEC 705.12(B)(2)(3)(c) both in letter and in spirit. The fact that 705.12(B)(2)(3) doesn't address feed thru lugs does leave you open to an AHJ saying that other rules should apply, but only a pig-headed AHJ would try to fail the installation. Unfortunately, I understand that those exist. : - )

Cheers, Wayne

P.S. A similar discussion came up at NEC 2017 705.12(B)(2)(3)(c)
Wayne, thanks for your input regardless, it is always valuable. You are absolutely right that it's often just a few inspectors pushing back can be a real headache.

When existing feed through lugs exist that we are tying into, I generally just tell our guys to use the main breaker. I also see the merit in your argument that it is no longer capable of accepting breakers, but probably not something I'd try.

Also for really damaged dead fronts, it's good to know there are other options besides the simple plastic fillers.
 
Upvote 0
Similar story here. I just had a 12kW/3PW install done, and Tesla did something similar to me. House originally had a 200A service with a 125A breaker running to a subpanel where most of the loads are. The plans I signed off on had AWG 2/0 running between the original panel and the gateway, and the gateway and the (whole house) load box.

When Tesla was done installing, they'd run AWG 3 everywhere (instead of AWG 2/0), and added 100A breakers on both sides of the gateway, limiting the entire house to 100A. AWG 3 is thinner than the wires running to the subpanel! Have about 5 meters of it between utility & gateway, and 5m between gateway and load panel.

I asked Tesla about it and they insisted AWG 3 is fine for 100A. Doesn't sound right to me (if you're using 2/0AWG for 200A, shouldn't you use 2AWG for 100A?), but lights aren't dimming when 12kW loads turn on and off so it's not comically bad.

Hope things turn out well for you!
 
Upvote 0