EEVblog 1426 - This Problem can DROP Solar Output by 20%

[EEVblog]

Dave tracks down why his old 3kW system that was moved was now showing a 20% drop in output power at the same time every day.
Is it inverter temperature, inverter mains voltage, solar panel cracking, dodgy connection resistance, or shading?

[SilverSolder]

That is a very interesting effect, amazing that such a small change in input is "amplified" in this way...   It would be good to understand what those max power tracking algorithms actually do.

[PartialDischarge]

Nice.
By leaving the mast down not only you win the extra solar power but also the time not wasted watching the useless content in the TV.

[Dread]

As Dave stated the problem he was having was almost totally due to the fact that he is using an older SMA inverter that does not have the SMA patented fix for their dodgy software, it's called "Shade Fix" and yes they actually Patented it!  Almost all the newer inverters do not have this issue and it's only a few of the older models that suffered from it.

THE STRING VS MICRO INVERTER SHADING EFFECT LIE HAS BEEN BUSTED!

Just about six weeks ago a couple of Australian Installers released controlled scientific test to once and for all prove that there is ZERO benefit to Micro Inverters when it comes to Shading.  Here is one of the main Videos but you can find the others easily.



Next up is the big Clipping Lie.   If you have batteries or sell back to the Grid then clipping is costing you a lot more than they want you to believe.

[Dread]

More Tests by a different installer in a Sunnier Part Of Australia.

[wasyoungonce]

Excellent video Dave.  I was going to put a small mast antenna on our roof...that would shade the panels (Lilydale), 5.8Kw string system.   Suffice to say I won't now

cheers

[tszaboo]

So probably it finds another local max instead of the global max for the power.
It is still surprising to see that such small shading can effect the output.

[Kleinstein]

The bypass diodes seem to be dividing the panel in vertical stripes of 1/3 a panel. Onece shaded the section would either reduce the total current for the whole string, or the whole section is bypassed by the diode. So 10% of shading for a single cell would give 10% less total current and power, or if the bypass diode is used (the MPP software decides which voltage / current to use) the 1/3 panel gives 0 voltage (actually some -0.6 V).

The way the shadow is, just the small mast can get some -10% to several of the vertical stripes. Maybe they should offer panels with a smaller subdivision for use with more shaded areas. With shade often coming from the bottom (not here), I would expect it to be better to have more a horizontal stripes, so the shaddow would like by more effecting only few subdividsion.

Chances are better MPP software could improve things a little.
Here the shaddow of the mast is close to the point of doing the most trouble with little shade, effecting 2-3 cell each with some -10% in the current.

[mrnewton]

Could one of you be so kind to explain to me how i can find the twitter discussion on this topic? I've never used twitter before and i figure there must be a better way than to scroll through Dave's entire twitter messages. (using the browser search within page for "solar" and other key words did not yield any results).

[tszaboo]

The way the shadow is, just the small mast can get some -10% to several of the vertical stripes. Maybe they should offer panels with a smaller subdivision for use with more shaded areas. With shade often coming from the bottom (not here), I would expect it to be better to have more a horizontal stripes, so the shaddow would like by more effecting only few subdividsion.

They do, it is called "half cut cells". The solar cells are cut in half, you have 144 cells in a panel instead of 72 (eg) two parallel string of 72, six bypass diode. Shading only kills half the panel, and it is more efficient by itself, because lower resistance.

[floobydust]

Dave, I suggest using a thermal imaging camera to regularily inspect and troubleshoot the system.
https://photovoltaikbuero.de/en/pv-know-how-blog-en/checking-bypass-diodes-part-2/ has some interesting thermograms, even for finding hot spots with poor connections in the array.

[EEVblog]

Dave, I suggest using a thermal imaging camera to regularily inspect and troubleshoot the system.

I did, it didn't show anything. Didn't include that in the video.

[EEVblog]

Could one of you be so kind to explain to me how i can find the twitter discussion on this topic? I've never used twitter before and i figure there must be a better way than to scroll through Dave's entire twitter messages. (using the browser search within page for "solar" and other key words did not yield any results).

Multiple threads unfortnately. And Twitter search sucks.
https://twitter.com/eevblog/status/1439871878798082048

[floobydust]

Dave, I suggest using a thermal imaging camera to regularily inspect and troubleshoot the system.

I did, it didn't show anything. Didn't include that in the video.

That's odd but maybe expected as the shaded cell doesn't make more much heat? I note this guy does short-term torture of entire arrays to get them to heat up. Example, his blocking-diode test involves injecting reverse voltage up to 400VDC at 2.5A to reverse crowbar an array. Youch.

[Kleinstein]

I don't think the bybass diode are bad here. This would more like bring the whole panels down in power, as the parts shorted by the bypass diodes would not contribute. Just measuring the cell voltage would be enough to see a missing section (1/3 panel).

The point here is that it is enough to have 1 cell per vertical row shaded to bring down the whole row. The shadow is just in the direction to shade 1 or 2 cells per row, so close to worst case orientation. The other point than seems to be a not so good MPP tracking algorithm in the old inverter.

Heating up solar cells with external current brings back memory to university. Was kind of funny to learn about cell quality from how fast they heat up on the front and back side. How much power you need depends on the sensitivity of the thermometer.

[thm_w]

As Dave stated the problem he was having was almost totally due to the fact that he is using an older SMA inverter that does not have the SMA patented fix for their dodgy software, it's called "Shade Fix" and yes they actually Patented it!  Almost all the newer inverters do not have this issue and it's only a few of the older models that suffered from it.

So are you saying this issue only affects the SMA inverter? Is this because it has a "bad" maximum power point tracking algorithm?
Or would this affect all MPPT algorithms without this specific fix in place.

Would it affect something like a LT8490 IC https://www.analog.com/en/products/lt8490.html#product-overview

[Kleinstein]

The shade would effect all inverters quite a bit.  Even though it is only a little shade, I would expect something like -10% (maybe even more) even for a good inverter. It is just the power the panels can provide is going down that much. The shade is small area, but in a way to have quite some effect. A inverter with a simple MPP algorithm like some older SMAs may make the power loss a bit larger. The linked AD/TL part seems to already search for the global optimum and thus is a good MPP algorithm.

[floobydust]

Anyone know what the "dual power peaks" is about with partial shade? From LT8490 datasheet.

[tszaboo]

Anyone know what the "dual power peaks" is about with partial shade? From LT8490 datasheet.

What I posted above.

[Kleinstein]

With partial shade and bypass didoes the power versus votlage curve can be a more complicated one, with several local maxima. So the controller has to decide which peak to use. The different sections correspond to the question of using the bybass diode and accepts a lower votlage or not use the bypassiode and accept the reduced current. Which one is better depends on how much shade, how the bypass diodes are there and how long the string is.

[floobydust]

Troubleshooting a remote site, the solar panel's output was pulsing at a few Hz and then steady good, then low.
I thought it was a connection/wiring problem until looking at a panel I saw a tiny tree branch shadow with leaves wiggling in the breeze. Then a gust of wind pushed the tree branch with wiggling leaves shadow across the cells. So it's pulsing, random changes in sunlight. Clouds are a piece of cake in comparison, slow and broad coverage. I imagine a flock of birds overhead could cause the same drama short term.

The solar charge controller seemed to get stuck hunting for MPP, which was impossible to find with input power going up and down at a few Hz. I think it needed a few seconds to complete a sweep to find MPP which it did not get. I'm not sure what a better algorithm is because you are wasting energy searching for the highest efficiency on something unstable.
At the time, IDK once the sun moved normal charging would happen, it's just that I was testing at the worst time. At least the firmware did not hang.

[floobydust]

Anyone know what the "dual power peaks" is about with partial shade? From LT8490 datasheet.

What I posted above.

I couldn't make sense of the graph, "global" isn't defined, is that the peak peak?  There's a lot more than one blocking diode involved, depending on the shade stripe orientation.
It looks like things get too complicated with shading, especially when it moves in the wind- to compute MPP?

[tszaboo]

Anyone know what the "dual power peaks" is about with partial shade? From LT8490 datasheet.

What I posted above.

I couldn't make sense of the graph, "global" isn't defined, is that the peak peak?  There's a lot more than one blocking diode involved, depending on the shade stripe orientation.
It looks like things get too complicated with shading, especially when it moves in the wind- to compute MPP?

Yes, things get complicated, because you have dozens of nonlinear elements in series-parallel. When solar is uniform, all the cells behave the same. If one gets shaded, then the I-V curve of that cell will be different. Current flows through all cells, so the end result is not trivial. You have multiple local maximums in power production, and if the inverter (or charger) locks on one of these, it might output less power than the global maximum. It needs to probe different voltage inputs to find the optimal power output.

[Geoff-AU]

As Dave stated the problem he was having was almost totally due to the fact that he is using an older SMA inverter that does not have the SMA patented fix for their dodgy software, it's called "Shade Fix" and yes they actually Patented it!  Almost all the newer inverters do not have this issue and it's only a few of the older models that suffered from it.

Yep.  It's totally unfair to blame MPPT.  Blame the specific shitty attempt that one vendor has implemented.

As you say, string vs micro-inverter myth has been BUSTED.  Well and truly debunked by NRG Solar using a Fronius string inverter.  The difference between string and micro-inverter in ANY shading or weather condition is LESS THAN 5%.

5% is not worth it.  I'd rather not have complex electronics sitting on a hot roof.  String inverters are much simpler and work every bit as well as any other solution.

[EEVblog]

As Dave stated the problem he was having was almost totally due to the fact that he is using an older SMA inverter that does not have the SMA patented fix for their dodgy software, it's called "Shade Fix" and yes they actually Patented it!  Almost all the newer inverters do not have this issue and it's only a few of the older models that suffered from it.

Yep.  It's totally unfair to blame MPPT.  Blame the specific shitty attempt that one vendor has implemented.

As you say, string vs micro-inverter myth has been BUSTED.  Well and truly debunked by NRG Solar using a Fronius string inverter.  The difference between string and micro-inverter in ANY shading or weather condition is LESS THAN 5%.

5% is not worth it.  I'd rather not have complex electronics sitting on a hot roof.  String inverters are much simpler and work every bit as well as any other solution.

The point of the video it to show what can and did happen, and for people to be aware of it.
And FYI, my model 3000TL-21 inverter does have "Shade management with OptiTrac Global Peak"
Sure, better solutions may exist. And just because one brand of string invertere does it really well does not mean that is now the standard in the industry. My bet is there are probably a lot of string inverters on the market that will have this issue to a certain extent.