Think I have an issue with my Tesla Energy panels

[Az_Rael]

Sigh, I am back, again with what I think is an issue with my panels that Tesla Energy is not acknowledging. (My previous experience is here: How to tell if my Tesla Energy solar system isn’t producing up to spec?)

So, I want to poll the community here to see if there are any suggestions for how to tell what might be going wrong with my panels. So I can maybe convince the Level 1 support guy to elevate me to Level 2.

Background on system: I have 2 arrays, with 2 Delta string inverters. Larger east array is on the 6.6 inverter and smaller South array is on the 3.8 inverter. Last year we had problems with the East array and Tesla fixed the wiring in August 2018.

Tesla is arguing that when they fixed our system in August 2018, we had too many strings hooked up to the small (south) inverter. They changed the wiring so that 3 strings are on the large (east) inverter and 1 string is on the south inverter. They claim that is why our south array is not performing as well this year vs last. However, based on the performance against PV Watts estimates, I think the South array is actually under-performing altogether starting in July 2019.

Since I am still 1% above the total performance estimates with Tesla, they say the system is "good, greens across the board". Last time I had to wait until the performance dropped about 20% over the total before they would do anything, and I am afraid I will have to play that waiting game this time. Unless someone has any suggestions for troubleshooting the system myself.


Here are the % against PV watts estimates for my two arrays. You can see the large divergence that happens starting July 2019. East array is over performing vs PV Watts by 14% and South array is under-performing vs PV Watts by 15%.

Here are the actual monthly totals by inverter:

I did take some FLIR images of the South array - the lower panels do look hot, but I can't say for sure that isn't just a weird FLIR reflection. It does look like I have some small hot spots on the upper two panels - not sure if those could reduce output significantly or not. Maybe I should strap the camera to my drone or something, LOL. I don't like getting up on my roof since it is tile.

Wish I had the solar edge inverter. Might be easier to see what the heck is going on.

Maybe I shouldn't care so much because my system in general meets all our energy needs, but I can't help it when I see trends in the data like that.

 

[ctslc]

It's worth noting that a FLIR - or similar - thermal infrared camera is not reliable in determining the actual temperature of any surface that is reflective - particularly at IR - so determining the temperature of the panels with it is problematic. When I measure the temperature of glass - or shiny metal - I have to attach something to the panel, such as a piece of tape - that has a more favorable reflectance. The angle, dirt/dust on the panels can affect their reflectance and this could explain why the panels "look" different.

From what you mention, it's hard to know the reason why you might have seen some drop - and there isn't quite enough information to be sure, but here are some possibilities:

One thing to note is that any PV panel will lose around 0.5%/C of output which means that on a hot summer day, one may easily see 10% of drop-off on a hot day compared to a cooler day - and I would expect the south array - and the roof - to be warmer during the peak of production - and summer heat. Comparing this years and last years production is a bit tricky as one must also take into account temperature into the equation - as well as things like weather (sun "quality") and the cleanliness of the panels themselves (e.g. dust - or lack of it.)

As for having "too many strings" hooked up, if you experienced "clipping" (e.g. hitting the maximum power of the inverter) - or if, under some conditions, the open-circuit voltage exceeded the inverter's ratings - I can see changing it. I would hope that the south inverter has several DC inputs - each with MPPT - and that each MPPT input was from a series of panels without mixed aspects (e.g. all panels on each MPPT input had the same orientation) - but if one MPPT input has a mix of both south and east panels, you will probably not get as much power as it would not be possible to fully optimize the string - but Tesla would surely have recognized this.

Finally, as for "clipping", even moderate amounts of clipping result in less loss than one might expect - particularly when one integrates the power over the entire day, and over the entire year: Clipping is likely to result only for an hour or two at most - and only during the summer-ish months, particularly when it is cooler and panel efficiency is higher - and compared to the longer (and lower) power production of the rest of the day - and year - the actual loss can be surprisingly low.

 

[Az_Rael]

As for having "too many strings" hooked up, if you experienced "clipping" (e.g. hitting the maximum power of the inverter) - or if, under some conditions, the open-circuit voltage exceeded the inverter's ratings - I can see changing it. I would hope that the south inverter has several DC inputs - each with MPPT - and that each MPPT input was from a series of panels without mixed aspects (e.g. all panels on each MPPT input had the same orientation) - but if one MPPT input has a mix of both south and east panels, you will probably not get as much power as it would not be possible to fully optimize the string - but Tesla would surely have recognized this.

Finally, as for "clipping", even moderate amounts of clipping result in less loss than one might expect - particularly when one integrates the power over the entire day, and over the entire year: Clipping is likely to result only for an hour or two at most - and only during the summer-ish months, particularly when it is cooler and panel efficiency is higher - and compared to the longer (and lower) power production of the rest of the day - and year - the actual loss can be surprisingly low.

Thanks for responding! Haven't seen any noticeable clipping of either inverter, so if it is happening, it happens for less than 1 hour.


If what the Tesla customer support person told me was true, they had previously hooked up panels from different orientations to a single string inverter, so who knows what they actually did. I had a lot of problems with my installation, including the wrong panels being installed, so the panels had to be removed from my roof and replaced. I am pretty sorry I used Tesla Energy.

One thing to note is that any PV panel will lose around 0.5%/C of output which means that on a hot summer day, one may easily see 10% of drop-off on a hot day compared to a cooler day - and I would expect the south array - and the roof - to be warmer during the peak of production - and summer heat. Comparing this years and last years production is a bit tricky as one must also take into account temperature into the equation - as well as things like weather (sun "quality") and the cleanliness of the panels themselves (e.g. dust - or lack of it.)

Well, I was hoping that since I had two arrays to compare, I could isolate the weather as the factor. Right now I am seeing the South array had higher numbers last year, but the East array had lower numbers compared to this year. The trends should be the same (both higher or both lower) if weather or sun was the factor I would think. (Can’t imagine I got a ton of cloud shading consistently only on one roof - that would be a heck of a micro climate).

t's worth noting that a FLIR - or similar - thermal infrared camera is not reliable in determining the actual temperature of any surface that is reflective - particularly at IR - so determining the temperature of the panels with it is problematic. When I measure the temperature of glass - or shiny metal - I have to attach something to the panel, such as a piece of tape - that has a more favorable reflectance. The angle, dirt/dust on the panels can affect their reflectance and this could explain why the panels "look" different.

Totally understand the FLIR comment. Was able to previously troubleshoot my previous problem with my East array using the FLIR, so I have found it useful in the past. Turns out these 6 panels were the ones with the problem back then:

Eh, at this point we may just have to wait it out for our production to drop below a level where Tesla will act.