Undersized inverter with my new Tesla Solar panels installation

[ai4px]

~7600 is kind of a magic number in the residential PV world. Like most people you probably have a 200A main panel. Most PV systems are connected in this panel. The max back-fed breaker in a 200A panel is 40A to prevent overloading the bus. The max sustained current for 40A is 32A or ~7600w.

I'm not sure this about max back feed. In our case we have a generator transfer switch and ended up double lugging the electrical meter, so none of the 16.5kw of solar goes through the house 200amp panel. Oh the other hand, the two solar inverters both back feed through a 100 amp sub panel right off the meterbase and one of them is 12kw, and it most certainly back feeds greater than 40amps.

The only oddity I ever heard of for a backfed breaker was that it had to be screwed in.

 

[wwhitney]

The 120% rule is one of several NEC compliance option for panelboards connected both to the utility and to solar. It is typically used for panels with many load breakers in them.

Another option is that the sum of all breakers in the panel, excluding the breaker on the utility feed, add up to 100% of the bus rating or less. So a panel feed via a "line-side splice" on the service conductors, and with a 100A bus, can be back fed with up to 100A of solar, if there are no other breakers in the panel.

As to the requirement that backfed breakers have a holddown, that does not apply to the breakers that are fed only by a grid tie inverter.

Cheers, Wayne

 

[nwdiver]

I'm not sure this about max back feed. In our case we have a generator transfer switch and ended up double lugging the electrical meter, so none of the 16.5kw of solar goes through the house 200amp panel. Oh the other hand, the two solar inverters both back feed through a 100 amp sub panel right off the meterbase and one of them is 12kw, and it most certainly back feeds greater than 40amps.

The only oddity I ever heard of for a backfed breaker was that it had to be screwed in.

Sub panels are different. The concern is that Current in via the main + Current in via solar could exceed that ampacity of the panel bus. I designed a ~23kW system on a home with 400A service. Our solution was to tap directly off the meter base to a sub panel exclusively for solar (no load breakers permitted). To be super safe we used a 250A panel and downgraded the main breaker to 150A. It's not a very well known rule. I doubt most inspectors or electricians that don't deal with PV often would be aware of it.

 

[thx1139]

We installed ours in December and it went live in January. I was noticing this as well and brought up to our installer. They explained that it was more effective to not get a larger inverter for that little bit extra. What I have noticed is our peak period is less in summer then winter. Overall our summer days producing far more energy just dont reach the peak for as long a period. So I think they were correct. Our installer wasnt Tesla Energy. They didnt open in Illinois until this spring.

 

[wwhitney]

Sub panels are different.

I'm going to have to disagree with that. It's not being a subpanel that makes a difference--it's the supply-side connection (solar is connected before the service disconnecting means) versus the load-side connection (solar is connected after the service disconnecting means). With a load-side connection, every panel that the solar generation goes through has to qualify under one of the options in 2017 NEC 705.12(B)(2)(3) (e.g. option (b), the 120% rule).

With a supply-side connection, 705.12(B) doesn't apply. In the supply-side 23kW example you gave, it would have been fine to use a 125a panel with a 125A main breaker.

Cheers, Wayne

 

[nwdiver]

I'm going to have to disagree with that.

Meant sub panel as in supply only sub panel. Not sure if the same 120% rule would apply or not if there's no load breakers. We complied just in case.

 

[wycolo]

Tilt-mount.

More gizmos is always better.
--

 

[spoonwzd]

My experience with inverter clipping is a similar story.

I had two 3.54kW systems installed on my East/West facing roofs back in February. The installer recommended and installed a SolarEdge 3.68kW inverter, as this matched the standard grid export limits for the UK.

It was pretty obvious even in February that my system was capable of producing much more than the inverter could handle and with my Powerwall 2 and planned EV purchases, I was capable of consuming more if it was being generated.

My installer was pretty amazing and within a month I had a second SolarEdge 3.68kW inverter with with my East/West strings dedicated to each. As I had a SolarEdge modbus export meter fitted with my original install they were able to daisy chain the inverters together but still limit the total grid export to 3.68kW in the event that I was not consuming enough PV.

I have seen production peaks over 6.8kW on a patchy cloudy day between the clouds and have regularly generated over 40kWh on a hot sunny day. My Powerwall 2 set to 5kW (21A) and now 2 EVs ensure I almost never export!

 

[nwdiver]

I had two 3.54kW systems installed on my East/West facing roofs back in February. The installer recommended and installed a SolarEdge 3.68kW inverter, as this matched the standard grid export limits for the UK.

That's not very typical. I've never heard of an array with an oversize ratio of 1.9. SolarEdge recommends no more than 1.55. It would be interesting to see how much the second inverter increased production.

I do have an increased appreciation for how much location effects the 'ideal' oversize ratio. Someplace that has really hot summers should probably be oversized by ~1.3-1.5 while a more temperate location should probably be something closer to ~1.2.

 

[wwhitney]

I had two 3.54kW systems installed on my East/West facing roofs back in February.

So I take that to mean one 3.54 kW system facing East, and one 3.54 kW system facing West (180 degrees apart in azimuth), is that correct? What roof slope (elevation angle)?

I have seen production peaks over 6.8kW on a patchy cloudy day between the clouds

Seems to me that could only happen if the elevation of the panels is very low (almost flat roof), so that when the sun's azimuth is near due south, both sets of panels are significantly insolated.

Cheers, Wayne

 

[mongo]

So I take that to mean one 3.54 kW system facing East, and one 3.54 kW system facing West (180 degrees apart in azimuth), is that correct? What roof slope (elevation angle)?


Seems to me that could only happen if the elevation of the panels is very low (almost flat roof), so that when the sun's azimuth is near due south, both sets of panels are significantly insolated.

Cheers, Wayne

I have seen production peaks over 6.8kW on a patchy cloudy day between the clouds

Nonblocking clouds (edge of cloud effect) can boost the light/ energy levels at the panels. It's something installers have to take into account when matching panel OCV to the inverter.

 

[nwdiver]

Nonblocking clouds (edge of cloud effect) can boost the light/ energy levels at the panels. It's something installers have to take into account when matching panel OCV to the inverter.

'Cloud edge effect' is negligible for OCV. It's temperature that's the limiting factor. A really cold overcast day will see much higher voltages than a partly cloudy day. Cloud edge effect DOES need to be considered for ISC. I've seen a 20A fuse blow from 2 8A strings because of cloud edge effect (Which wasn't even required...)

Current is mostly sunlight dependent.
Voltage is mostly temperature dependent.

 

[spoonwzd]

That's not very typical. I've never heard of an array with an oversize ratio of 1.9. SolarEdge recommends no more than 1.55. It would be interesting to see how much the second inverter increased production.

I realise my system isn't typical, but my installer basically did what I asked. I thought there would be a fairly obvious switch from the East to West roof in terms of production, but in practice ambient and scattered light can still produce a decent amount of energy when in shade. This is particularly obvious on a day where there is full cloud coverage but the cloud is not heavy - both East and West systems produce quite evenly across the day.

Here's a good production day when I had just a single inverter. I hit the inverter limit at 11:15 and production didn't drop under the limit until 15:15 - and this was back in April when there was less daylight.

Compare that to a good generation day in the height of summer with the two inverter setup. There was not a cloud in the sky, but it was very humid and what I can only assume to be high-level / stratospheric haze. My ststem only peaked at 5kW generation but still generated nearly 46kWh over the day.

And then you have days like today where you have patches of heavy cloud that part to reveal huge bursts of sunshine - even a 15 minute average shows the peak at 6.5kW, but in reality there was a sustained 5+ minutes of 6.8kW PV generation. Again, I assume that this is because the panels are generally cooler and the lack of haze when the clouds part.

I would be interested to know the actual reason for these differences in production between a seemingly 'clear' day and one that has patches of sun.

 

[spoonwzd]

So I take that to mean one 3.54 kW system facing East, and one 3.54 kW system facing West (180 degrees apart in azimuth), is that correct? What roof slope (elevation angle)?


Seems to me that could only happen if the elevation of the panels is very low (almost flat roof), so that when the sun's azimuth is near due south, both sets of panels are significantly insolated.

Cheers, Wayne

Yes that's right - I have one 3.54kW system on my East roof and another identical system on the West. My roofs are at a 32° pitch and are facing 121° and 301°. They're not flat.

 

[nwdiver]

I would be interested to know the actual reason for these differences in production between a seemingly 'clear' day and one that has patches of sun.

As was referenced earlier this is due to 'cloud edge effect'. If there are clouds not blocking the sun you're getting direct sunlight AND reflected sunlight. These can combine to produce higher power levels than would otherwise be possible on a clear day.

How much additional energy (kWh/day) are you getting after adding the second inverter?

 

[spoonwzd]

'Cloud edge effect' is negligible for OCV. It's temperature that's the limiting factor. A really cold overcast day will see much higher voltages than a partly cloudy day. Cloud edge effect DOES need to be considered for ISC. I've seen a 20A fuse blow from 2 8A strings because of cloud edge effect (Which wasn't even required...)

Current is mostly sunlight dependent.
Voltage is mostly temperature dependent.

View attachment 318758

Interesting - so it's probably temperature that's limiting production to 5kW when it's 28°C ambient (hot for the UK!) and the sun's been baking the panels all morning. I knew temperature affected things, but had no idea it was to the tune of 25%+

Thanks for the info guys

 

[spoonwzd]

How much additional energy (kWh/day) are you getting after adding the second inverter?

It will be hard to know this for sure until a similar time of the year in 2019, but I would estimate this to be somewhere between 40-60%.

 

[nwdiver]

Interesting - so it's probably temperature that's limiting production to 5kW when it's 28°C ambient (hot for the UK!) and the sun's been baking the panels all morning. I knew temperature affected things, but had no idea it was to the tune of 25%+

Thanks for the info guys

Yep;

 

[spoonwzd]

Yeah I guess the panels are much hotter than 44° in the sunshine. A 25% power loss probably puts them in the 90-100°C range!

 

[wwhitney]

Yes that's right - I have one 3.54kW system on my East roof and another identical system on the West. My roofs are at a 32° pitch and are facing 121° and 301°. They're not flat.

Thanks for the info and the data, it's enlightening.

Cheers, Wayne