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

Disappointed by safety systems impact on production

This site may earn commission on affiliate links.
We have a new Solar Roof (12.7 kW on 5 strands) and one of the things that's been very surprising (and disappointing) is that the "slope" of the production curve is very steep in comparison to older installations. There's virtually no output during early morning/late evening and on heavily cloudy days, in comparison to older panels.

I've been monitoring the output from our Delta inverters: using BLE and the M Professional app, I can watch the output of the strands throughout the day. One strand will be a small amount (a few hundred watts) and the other 4 will be completely at 0. It seems that a particular strand needs to get to the critical level of brightness before any production on that strand happens. This is in contrast to an older systems or even a modern standalone solar generator where there's current flowing in even near-dark conditions (though quite weak).

For example, two days ago, it was a stormy day and we produced only 7 kWh whereas an older system would have eked out around ~15 kWh on a day like that.

I asked Tesla about this and they said that the NEC 2017-required MCI (mid-current interrupter?) rapid-shutdown devices (RSD) are the culprit: they do not allow low production to flow. I found these Delta RSS (PDF) referenced in a public Tesla submittal. If those are the ones that they use today, then they have a minimal operation current of 4W (to power itself). Based on Tesla training materials I've found in Google searches, it looks like they do about 10 tiles per 1 MCI. Doesn't seem like that would account for the behavior that I'm seeing unless I made a mistake with the math.

Can anyone confirm what I'm seeing? Why are they designed like that, if so?
 
If your system has about 500 solar tiles and each set of 10 has a minimum of 4W the minimum for your system to start would be around 40W per string or 200W total, is that tha amount you are missing? Hopefully it wont be an issue in a month or two when days are longer\ more sunny. I am pretty close to you and my 5.2kw south facing traditional panels produced 5.7kwh two days ago so your numbers dont seem that far off if your array isnt all south facing
 
Can anyone confirm what I'm seeing?

I haven't been patient enough to watch my individual real time string production through M Professional for long enough to notice particular trends. Although this may be too high-level to satisfy your concerns ...

Tesla's customer care rep explained that they use the following estimation rubric:
For winter (Nov-Jan): 1 x system size
For spring and fall (Feb-Apr, Aug-Oct): 3 x system size
For summer (May-Jul): 5 x system size

So for my 9.94 kW system, Tesla expects me to average at least 9.94 kWh per day in November, December, and January.

My solar roof is only slightly north of yours on the other coast, so I'd guess this estimation would work for you too. If so, 7kWh on a cloudy day in winter doesn't seem too bad as long as you're averaging closer to 13 kWh per day for the month.
 
I haven't been patient enough to watch my individual real time string production through M Professional for long enough to notice particular trends. Although this may be too high-level to satisfy your concerns ...



My solar roof is only slightly north of yours on the other coast, so I'd guess this estimation would work for you too. If so, 7kWh on a cloudy day in winter doesn't seem too bad as long as you're averaging closer to 13 kWh per day for the month.
Those numbers seem to be what I am seeing. BUT, no way one can say my system is a certain size and expect anything. Depends on what direction panels are pointed, angle, shading, etc!
 
BUT, no way one can say my system is a certain size and expect anything. Depends on what direction panels are pointed, angle, shading, etc!

Maybe I'm misunderstanding something, but I thought the "size" of the system reflected the peak output of the entire system taking into account panel direction, etc. Therefore, you could certainly base expectations on the system size - in fact, I would think that is whole point (happy to be corrected if this is wrong ...).

Put a little differently ... I assumed that system size was the peak output of a system as configured. I think you are suggesting that system size is simply the sum of the peak outputs of all PV panels/tiles in the system assuming each one is optimally positioned.
 
Last edited:
Maybe I'm misunderstanding something, but I thought the "size" of the system reflected the peak output of the entire system taking into account panel direction, etc. Therefore, you could certainly base expectations on the system size - in fact, I would think that is whole point (happy to be corrected if this is wrong ...).

Put a little differently ... I assumed that system size was the peak output of a system as configured. I think you are suggesting that system size is simply the sum of the peak outputs of all PV panels/tiles in the system assuming each one is optimally positioned.
The latter is correct - it is the total size of all the panels or tiles. Every person who has a "medium" system from Tesla (currently 24 340W panels) will be listed at 8.16kW. It does not matter whether those panels point north, south, east, west, or a mix. Tesla also provides an annual output number, which is what you would tend to look at for performance. However, it does not tell you about how it will behave seasonally. At the extreme, the seasonal difference could be significant between two systems, both of which are projected to produce 10 MWh annually, but one pointed due south and one due north. (The north one would, of course, need more panels and thus have a larger size in order to get to 10 MWh.)

The rated wattage of the panel is what it is expected to produce under a standard set of nearly-ideal conditions. It actually can (though rarely) produce more in just the right circumstances. However, at most times it is expected that the output would be less than the wattage.
 
No, the size is purely an ideal operating condition value for any given panel/tile, times the number of them used. It has nothing to do with the installation details or other conditions that can affect production. Like my Solar Roof totals to 13.32kWp, 561 tiles at 23.75W each. But those tiles are in various orientations, so some will never produce nearly as much as others, the realistic peak production has hardly ever reached the 10.4kW max output of the two inverters (never on a clear day, but has hit the inverter limits briefly a bunch of times on partly cloudy days due to the cloud edge effect). That same 13.32kWp worth of tiles on a different roof could produce more or less energy (different orientations/pitches, different latitudes w/different sun angles, hotter/colder climates, shading, average cloud cover, etc). So the estimate of total annual production can vary wildly for a system with the same number of panels/tiles.
 
  • Like
Reactions: charlesj
Maybe I'm misunderstanding something, but I thought the "size" of the system reflected the peak output of the entire system taking into account panel direction, etc. Therefore, you could certainly base expectations on the system size - in fact, I would think that is whole point (happy to be corrected if this is wrong ...).

Put a little differently ... I assumed that system size was the peak output of a system as configured. I think you are suggesting that system size is simply the sum of the peak outputs of all PV panels/tiles in the system assuming each one is optimally positioned.
Guess it depends. I have a 14.5KW system. I will never get this. First, panels are never 100%, and over time they get less. Second, I have said, sun angle, sun time, etc. And then depends on the inverter. I have a 11.4K inverter, so I cannot go above this. I have seen once in a while a little bit of clipping, but 99% of the time I am no where near even that.

So there are so many variables, you may or may not get something, but ....
 
Got it - thanks to all for setting me straight :). I wonder if there is something else about Tesla's process (like limiting sub-optimal positioning) that allows them to assume that systems of a certain size will have certain minimum production (making a 1x, 3x, 5x rule possible).

As an aside, it still surprises me that summer production is expected to be 5 times winter production.
 
Got it - thanks to all for setting me straight :). I wonder if there is something else about Tesla's process (like limiting sub-optimal positioning) that allows them to assume that systems of a certain size will have certain minimum production (making a 1x, 3x, 5x rule possible).

As an aside, it still surprises me that summer production is expected to be 5 times winter production.
This again depends on the setup. If a person is adjustable panels, facing south, that is totally different than again lets say me.
1/3 of my panels face west so that is a negative in winter. Plus, with low sun angle, they get tons of shade of the structure of my garage.
So with this, I can peak in summer at like 96kwhs, but in winter, well, depends on if it is raining, then I get 5 kwh in a day
 
That's still not a rule, just a rule of thumb. If your solar system was at the equator it wouldn't be true at all. But in some parts of the US it is probably a reasonable ballpark estimation. From pure sun angle my system certainly doesn't vary 5x from winter to summer. A good winter solstice day is around 20kWh, while the best summer day is around 70kWh. But when you factor in how much cloudier it is in the winter vs. summer here, that hurts the winter average more than the summer figure, it's probably still not quite a 5x delta, but probably somewhere around 4-4.5x, I'd have to go back and crunch the numbers to get an exact figure.
 
Back to the original question, it only makes sense that boxes powered from the solar system will consume some power on their own, and require some minimum amount of solar production before they can do their job (which in the case of a RSD is passing power through to the inverter). The same is true of the optimizers used per-panel on Solar Edge systems, and one of the down-sides to having so much distributed electronics.

I don't know how the newer Solar Roof systems are being done, I've seen pictures showing the newer (and smaller) RSDs but didn't realize there might be many more of them throughout the system. In my v2 system there are a total of 6 strings of varying size, that get combined into the 4 inverter channels (2 inverters each with 2 inputs). My system has 5 Delta RSDs, because 2 of the combined strings exit the roof close enough that they could be combined before the RSD, and the other two combined strings are far enough apart that they must each have their own RSD (so they get combined after the RSDs). My RSDs are rated at <4W consumption each, so they should be drawing less than 20W from the total system power, a small amount compared to the multiple kW of normal output. But if the newer v3 installs are using many more RSDs and they still draw about 4W each then it could be a bigger deal.

The inverter itself will also need to see a certain amount of power before it will turn on, it's not like as soon as there's 1W of power hitting the panels that it'll turn that 1W into AC power. At least for my system, I think that inverter number is quite a bit larger than the RSDs, probably more like 50-100W per-inverter. So yes, if conditions are really bad it might be that the little bit of sun power hitting the panels doesn't get converted to any electricity. Ultimately the overall system has an efficiency curve, and it would say that below a certain amount of solar power the efficiency is 0. But by the time there's more than a few hundred W hitting the panels, those efficiencies will be much be much, much higher, and the tiny bit being drawn by the RSDs, cabling losses, etc. won't amount to much.
 
Last edited:
That's still not a rule, just a rule of thumb. If your solar system was at the equator it wouldn't be true at all. But in some parts of the US it is probably a reasonable ballpark estimation. From pure sun angle my system certainly doesn't vary 5x from winter to summer. A good winter solstice day is around 20kWh, while the best summer day is around 70kWh. But when you factor in how much cloudier it is in the winter vs. summer here, that hurts the winter average more than the summer figure, it's probably still not quite a 5x delta, but probably somewhere around 4-4.5x, I'd have to go back and crunch the numbers to get an exact figure.
Do you like I can get 5KWH on my setup in the winter with rain clouds, 96 kwh in the summer with perfect conditions. This just just about 20x difference. So yep, lots of rule of thumbs, but everyones mileage will be different.
 
Got it - thanks to all for setting me straight :). I wonder if there is something else about Tesla's process (like limiting sub-optimal positioning) that allows them to assume that systems of a certain size will have certain minimum production (making a 1x, 3x, 5x rule possible).

As an aside, it still surprises me that summer production is expected to be 5 times winter production.
I am not certain Tesla's "rules" (I saw one page on their site that listed up to 7x difference) are always applicable to a specific system - I know sometimes when you call in they seem to indicate they have monthly estimates, but even then I'm not certain where those come from. The most reliable number is the one you get with the contract, but it is only an annual number. PVWatts should give you a pretty good idea of how to allocate the Tesla number by month.

It is worth noting that the more optimized the system, the less it will vary over they year. As an example, our south-facing roof, which is close to optimal, will (per PVWatts) produce 1.46 the amount of solar in June as December. However, our north-facing shingles will produce 8.96 the amount, and that is due mostly to production loss in winter (it actually is estimated to produce 94% of the south-facing amount in June, so almost as good.) While there is probably a good margin of error in these numbers (particularly the 8.96 number) it illustrates how widely the summer/winter ratio can vary with panel location.

I think the Tesla ratios are intended to kind of be an average, and are also used by front-line support to try and assuage customers' concerns without having to actually look into whether the concerns are valid.
 
I don't know how the newer Solar Roof systems are being done, I've seen pictures showing the newer (and smaller) RSDs but didn't realize there might be many more of them throughout the system. In my v2 system there are a total of 6 strings of varying size, that get combined into the 4 inverter channels (2 inverters each with 2 inputs). My system has 5 Delta RSDs, because 2 of the combined strings exit the roof close enough that they could be combined before the RSD, and two of the strings are far enough apart that they must each have their own RSD (so they get combined after the RSDs). My RSDs are rated at <4W consumption, so they should be drawing less than 20W from the total system power, a small amount compared to the multiple kW of normal output. But if the newer v3 installs are using many more RSDs and they still draw about 4W each then it could be a bigger deal.

My v3 install is, as I understand it, similar to yours as far as the RSDs. We initially have four strings of 35 shingles each. Each set of two strings are then combined and run through an RSD in our attic, so we have two RSDs - one for each face of our roof. The two combined strings of 70 shingles each are then fed to our inverter, using two of the four channels available. The RSDs and inverter are also Delta products.