Due to several requests I have created a 5min version of this video that allows you to share with people who don't want a 30min video: kzfaq.info/get/bejne/hNV5nbd8l5qddGQ.html

5:51 "one of the curses of being an engineer [...] is you know too much" So much this - When watching a show, or an art installation you automatically start questioning every little detail and always wonder "how did they do it" instead of just enjoying the show's "magic" in awe like everyone else.

I took my antenna down when I had panels installed because my guy said it would affect my power production. I was sceptical, but I trusted him because he knows his stuff. This video confirmed his advice.

Did you update your SMA inverter with the latest firmware ? They introduced their "magic" shadefix algorithm (basically a quicker MPPT swipe) which greatly improves changing the MPP in case of shade. I'd love to see your update about it !

I see your problem...you need to install "wiggle wiggle wiggle yeah" inhibitors :D

I experienced this as a kid playing with some solar panels. Called it mouse effect after my plush mouse that if I put it on the panel, while only shading one of the cells in the panel, the output of the whole panel was nearly gone.

Light shading and compound shading is tough on older string inverters. Basically your bypass diodes aren't fully conducting, and the shaded cells in the series are limiting the current of the whole array. Modern string inverters can look for alternate maximum power points. I have 3 MPPT channels on my newer Sunny boy, with shading from a water heater stack and plumbing vent that affect two of the channels after a certain time in the day. But I only loose the power of shaded panels (or portions of them) when the stacks throw shade - not the large reduction you see. This tech seminar describes how newer SMA inverters work (most vendors employ similar techniques). kzfaq.info/get/bejne/idNggciBt726aY0.html

Soo... Exactly what I suspected then. When we installed a 3kW system many years ago on my dad's house, we put it into operation with the scaffolding still in place. There was just a thin shadow from the railing across half of the panels (essentially as thick as a water pipe). After we took the railing down, the output increased a solid 1kW just because that shadow was gone. We were quite surprised the effect was THAT significant as well. But at least we figured it out immediately.

Newer panels no longer have diodes, instead using a small IC that has an integrated power MOSFET, so that when current is detected flowing through the body diode the tiny bit of circuitry controlling the gate turns the MOSFET fully on, and shorts out the body diode and thus reduces power flow in the device. No power dissipation, and the energy to turn the MOSFET on is stored in a small ceramic capacitor using the normal off state voltage. Turns off the switch every few seconds, to both recharge the capacitor, and also check if the panel shading is transient and thus panel is back at full output.

Unfortunately right at the start of this video I knew your problem. I was a radio tech from the mid 80 and worked for the police in Queensland. At the time the cells we were running at remote repeater sites were as low as 45W per unit and we used to have large strings of panels to keep sites alive (given there critical use). One particular site we had particular problems with during the Winter and it kept dropping out. We ended up at the site one day and were checking the current output from each cell trying to put a finger on the problem (previously we'd replaced all the cells and the Solrex regulator with no result) and discovered half the panels were low output. We then noticed the guyed wire from the 25 ft wind up Hills master crossed only the panels showing low output. Like a light bulb we'd found the problem after nearly a year of attempts to fix this repeater site. After a further trip we ended up building a new stand for the panels and moved the array away from the tower and the problem was solved. I do follow you on Twitter, I must have missed this series of tweets. Great video, wish we had KZfaq in the late 80 :)

Really glad you went on to explain what the bypass diodes were for, thanks!

Rising voltage in evening is industrial and commercial customers reducing load, along with the home consumers reducing load from electric cooking, so the overall load in the area is dropping, and the input voltage from the power plant is essentially constant, so all those small reactances and resistive line drops add up.

I never would have figured partial cell shadows would be that drastic, that's pretty crazy.

Wow, this looks to have a much bigger impact than it looks to the human eye at first glance. Nice eco-drive watch Dave! It's a pretty good idea on the battery side of things also you never have to worry about WR after a battery change. Love it myself!

Now that was really interesting! Trap for (not so) young players! I'd like to see an experiment, now that you have the micro-inverters I'd like to see how much dust affects power output. You could do this by washing one of the panels say every other day and see if it outputs any more than the rest. Also temperature difference (this would be trickier). But I guess this wouldn't be an issue, if not manufacturers would sell a watercooled version lol

Excellent discussion of solar panel architecture. Now I understand why one of my panels outputs exactly 2/3 the voltage of all the others, but the same current. A single bad cell or open internal connection will cause this even though all my panels are in full sun with no trees (or coax cables) nearby. Congrats on solving your problem. Every kWh is important.

Just put the damn pannel on the road. It seem very efficient 😅🤣

This is an excellent video! Very informative and I love how you cover the problem, troubleshooting, fix and explanation.

Effectively when a bypass diode kicks in, you get a second, but lower maximum in the MPPT curve. Not so sophisticated MPPT algorithms will get stuck on this lower maximum and do not search for the higher one. Would be interesting if a reboot of the MPPT controller helps, so it may search the whole area again an sees the higher maximum?

Hi Dave, Jason from Prohelion here. As an ex solar car racing peep, this makes complete sense. While driving down the Stuart highway at roughly 90 km/h, even the thinnest tree would cause the MPPT to drop its power output by 60-80% for upwards of 3 seconds. Thankfully there is not many trees in the outback. #DontShadeTheArray

Great video, Dave. Recommend making this a short. It's what all the cool kids are doing!

I was confused about the direction of those diodes until you explained it. Learned something new today from you.

Hi Dave, absolutely agree with your assertion that we engineers can jump to diagnostic conclusions that satisfy our need for complex, rather than simple, problems. In my early career I spent the better part of a morning attempting to urgently diagnose a "motor won't run" problem with a 3 HP three-phase synchronous motor. It had a complicated relay-based starter & direction-changer control box {this was 1972 - all brute force stuff}. Studied the schematic until my eyes were bleeding, tried many fixes, but no joy. Then a buddy of mine in the same plant dropped by for a chat, and he asked if I'd checked the fuses. Bingo - one of the three 100 ampere fuses for the incoming line was open. Replaced it, and all was well again. I never forgot the lesson...

Your panels have 60 cells separated into 3 power generating columns protected with bypass diodes. It only takes a slither of shade to drastically reduce the current output of a column. If you know a shadow will traverse the entire panel you can consider reorienting to landscape and only impact one column of 20 cells.

Wow, if that little shadow was able to wreck so much havoc, I wonder how effective a reflector and lens that tracks the sun and reflects onto the panels would be. Well it would be the solar surface area of the reflector divided by the area the lens projects to, not good, and a shadow might still heat your little cells.

As soon as I saw your time-lapse, I knew exactly the problem. It's the shadow from the antenna. My system, as I was told during installation, a shadow would basically cause the individual panel to stop producing power. Which reminds me, I need to go up and clean my panels. Dust, dirt and grease (from cooktop exhaust) needs to be cleaned off. Which needs to be done before the sun hits the panels. My issue is trying to find a safety observer willing to get up at 0dark30 😅

Now all you need is a rotator motor to turn the house to track the sun.

I would love to see an experiment to see how the new array would react to a small diffuse shadow

working with and teaching solar cells I did not believe that these narrow shadows could drop your output that much with proper bypass diodes installed - but anyway you should see this from the dc current on the solar panels which you could measure with a current clamp. The shadow of the mast could be perhaps 10% of the area on a cell which would reduce the current in the affected cell and substring in the maximum power point also by 10%, reducing the power by 10%. The bypass diode will only carry current when the substring voltage drops so much that the bypass diode becomes forward biased because the other cells in the string reverse bias the shaded cell. In your case the shaded cell(s) were just reducing the full current by 10-20%...

Super interesting! Thanks for sharing!

What got me thinking, a solar installation with micro inverters could actually work as a huge format camera .It wouldnt be focused at a specific point, but could stil show movements of clouds.

I knew it was a"moving vertical shadow" the moment I first saw the graphs! (and I don't even know HOW I knew that) glad that you got it sorted :)

Dave, noone thought you didn't check your roof for shadows. And just because of that, the simpliest explanation turned out to be right explanation.

Dave you might want to check if your old SB3000TL-21 already has the "shade fix" feature. In older firmware versions it might be called "opti trac global peak". This helps a lot with problems like this.

Thanks for the current flow clarification!

I would have never thought that such diffuse shadow would cause an effect that drastic. You learn something new every day.

Wow, all that difference from a thin shadow? It's a good job nobody ever thought of putting solar panels on roads, isn't it. ;)

I have two 100 watt solar panels driving two 12 V ventilation fans in the attic of a separate garage. One panel for each fan. When even the slightest bit of shade hits one of the panels, the fans drop off like crazy. I could even it out with some batteries, but it is nice just to have the panels directly connected to the fans!

On some of the Sunny Boy inverters, you have multiply inputs so you can divide your grid into smaller units. In that way shade will only drop output on that smaller grid. The drawbacks are higher cabling costs, and an isolators is needed for each circuit.

Woweeee. I always knew a shadow can create an issue. But that small shadow?! Wow great job discovering this David.

I've designed a couple of mppt chargers and I've noticed how a "small" shadow can make a huge difference to the panels output. Not so bad if a bird poops :0) on one cell , but if a few series cells are obscured , even slightly , it's very noticable.

This looked like a click bait, but isn't one. Really useful video

Very fascinating! It just goes to show how important it is to keep your panels clean from dust and bird droppings. Not only because it significantly reduces their power output but it can also reduce their life. Thanks so much for sharing. ;)

I loved the commentary on "knowing too much." I work in product development and spent an entire team meeting discussing bizarre test results. Everyone had weird ideas to explain what we were seeing. I eventually just retested the prototypes myself (a technician did it the first time) and found that the data sheet must have just rotated all results by one ID (results for A were recorded on the sheet for B, and so on). I was probably 2 years into my role, and it's funny that you put it like you did... That was the exact thought I had when I solved the riddle. "Dang it. I know too much." When you don't have the knowledge to invent these minuscule probability explanations, you check the simple ones first. That was a great lesson to learn.

Amazing find, after watching this, i must go on the roof, and check my array. thank you.

FWIW there's also the SolarEdge system. Modified string inverter with an MPPT box on the back of every panel. Fixes this issue and gives per-panel monitoring.

Now imagine that (in)efficiency on solar roadways...

Yeah, I believe that! I have microinverters with split-cell panels, and early morning sun catches a few thin tree branches, and I can see the low output move across the panels as the sun rises. One of the reasons I went for a microinverter system in the first place.

WOW.... this video is amazing, crazy useful knowledge to keep in mind. I knew shadows caused power dips but never imagined something so small causing that huge of a power loss.

Great video! Would love to see you do an in-depth video on MPP. Another interesting video idea would be teardown and explanation of DC optimizers, going over both DC to DC conversion optimization and the communication method that they use.

You still watch broadcast TV? Wow, I ditched that over a decade ago and never regretted it for a moment.

Great explanation of bypass diodes on a solar panel - thanks.

I had this exact issue with overhead powerlines on the first small-scale ground install I did for my off-grid building install. Good catch!

Thanks, I'm amazed at how much that little shadow affected the output

OMG I'm so happy you mentioned why the diodes were in reverse!! It was driving me crazy. I love this man.

It would be interesting to cover up that entire panel with something opaque, and see if the power increased, compared to with the fuzzy coax shadow. The 3 diodes should conduct the full current of the remaining panels, and lose only 8.5% of the total power.

The voltage drops in the evening when demand peaks from everyone having dinner. It rises again as the demand trails off to a minimum at around 4am.

Without shading your entire string has only 1 Maximum power Point With the active shading you get two MPPT points. 1 with all modules on the reduced current of the weakest cell (all other cells will rise in voltage and thus reduce theirs current) or with the MPPT Current of the other cells and the shaded parts disabled via the diodes, thus theirs voltage will be missing on the string. On modern inverters there is a global MPPT tracking where the inverter cyclicly scans for other better MPPT points. Maybe the inverter here slowly reduces the current of the entire string when the shade moves in because the shadow is to narrow or is not searching for other MPPT points because the inverter hasn't that GMPPT feature.

In other words the solar cell turns into an LED, or a Zener diode. I once made a solar LED array to prove a point to some instructors and classmates. The strings weren't long because the voltage would get high enough to turn the LEDS into rectifiers and conduct without a regular diode in series to stop it. It would take many thousands of LEDS to make a solar panel.

amazing. this video may explain why many consumers arent happy with the output of their expensive solar panel systems compared to what the suppliers promised. i have made a mental note of this for when i get solar panels

I've run across this with a small off-grid system I use when camping. One little shadow covering enough of the panel could mean the difference between a flat battery and a fully-charged one.

Very interesting , I had taken my aerial down , well actually wind had blown it nearly down , so glad I never replaced it

That was really interesting! Have you got an old panel you could dissect, at least to the level of the three strings / bypass diodes and conduct tests to show that even the diffuse shading of a cell or two can drop the output like this? The old micrometeorite panel maybe?

I am surprised they don't fab a bypass diode in the individual cells. They're already fabbed silicon. Also, is there any advantage to ganging cells in parallel as opposed to series? You're trading voltage for current, but if you can use microinverters per panel, I'd tend to think resistance and bus gauge issues would be minimal. It definitely would require a completely different inverter design, though.

Honestly I haven't seen this on twitter, and first thing I thought when you explained the issue, was about shadows, then you mentioned the mains voltage and I was like "ok this could also be the issue", and actually I was sold on the mains voltage idea when you showed the time laps, there was like 0 shadow. I honestly couldn't believe that tiny narrow shadow killed the entire array, think having a micro-inverters system would have helped maybe saving some power-percentage meaning you would have lost less as you say. Insane anyway I still can't believe that tiny shadow dropped 20% power.

Fantastic video series over, what, 2 years minimum! Thanks for the detail!

Wow - glad you found this. I am just setting up a 12V system and I know that one panel is going to get some shade. I had been musing about serial vs parallel so I am going to wire the 4 panls as 2 x 2 parallel

Very interesting. I remember getting a report at high school, it may have been my leaving testimonial, saying that "Anthony will often overlook a simple explanation in order to seek out a deeper reason for a problem", so you're not alone in that regard!

Yep. Waiting for my new house to be ready so I'm not installing on this house, but I set up 7X 100W panels I got for cheap at an auction to experiment with panels, batteries, charge controllers, etc. I found that one little leaf dropped on a series string had a dramatic impact on output, shockingly so. I knew what your problem was as soon as I saw the video title. I was hesitant to go to the expense of microinverters for the new house system until I saw the result of even minor shading. It makes me think that regularly washing dust from panels is probably a good idea.

All you had to do was download an update for your Sunny Boy that provides Shade Fix. Of course, you can also replace the robust string technology with module electronics. It is always a good idea to place complex electronic components on the roof where they are exposed to the maximum possible temperature fluctuations.

Thank You. I would not have believed that small shadow would do that.

Nice observation! 253V is where it's mandatory for a inverter to shut off (or limit) here. So mains voltage theory wasn't that wild.

I have a client that had similar issues and what I suggested and he did was add 3 more panels to boost the output and that did work. In his case the total daily output was 40% below what was expected it has been over 5 years and the system is still working at full capacity.

Exactly, a great example of why panel level microinverter or panel level mppt optimization (Tigo etc.) is so important and advantageous to strings. Some shade indeed makes a big impact.

Wow that was very interesting who would have thought that a slight shadow would have had such a dramatic effect on output power. You've right, sometimes you can look to deeply into a problem and not see the wood for the trees. I liked the way you analysed the problem and eventually came to a logical conclusion. Well as an excellent engineer we shouldn't expect anything less, should we!!!!!!!!

Wow...I totally wrote off that shadow when you first showed it. Crazy that much loss from cascade of issues.

Deep insight with the investigation and expanation as always. Thank you, Dave.

Newer SMA inverters have ShadeFix, which is supposed to solve most of the issues with light shading.

Thanks for explaining the (seemingly) reverse-biased diodes in the panel diagram. Definitely not used to thinking in terms of current sources. 👍

The funny thing is that I SAW that pole shadow in the time laps video. I though that this might be the problem and sure enough, to me surprise, it was the only problem. I guess that the benefit of watching the EEVBLOG video on a 40 inch screen. Even the small details are clearly visible.

Nice work Dave, I discovered this with a smaller caravan unit . Shading is a real issue for matrix panels. There is probably great room for improvement .

That's quite surprising - people always said they were fussy, but this video proves it conclusively. I took a 455W panel up to our block some months ago, mid winter, and in FULL shadow (no direct light, limited other light) I was getting best case 45W from it :-D At home the same panel was doing 200-250W with full winter sun on it. Ended up charging my battery with an idling car multiple times per day, ha ha :-D

I'm surprised you only have just discovered now that when your neighbors have solar feeding the grid it can over volt the street and reduce the amount of you're feed in power. It screws you're pay back models as you may not get the full feed in power as expected, and its impossible to know if your neighbors will get solar too in the future as its bad for you. Also the dramatic power loss on partly shaded series string panels is a issue too..

Mains voltage going up/down would be correlated to peak and off-peak consumption of energy... lots of people using the grid around 5-7pm ;)

I like the idea and simplicity of microinverter designs. But being required to keep an internet (or any other type of) data connection from the microinverters to Enphase or other company's servers is non-negotiable deal breaker for me.

Round 1 to the Aussie Sundial! I had tested shading in my solar experiments starting about 2013. I was trying different combinations of panels, bypassing diodes and various configurations trying to see how to maximize yield in less than optimal conditions. My system is DC as I charge batteries to provide power isolated from the mains. My present test system has 6 x 350W panels in pairs on 3 masts that I rotate throughout the day (I'm retired and haven't put up linear actuators). The single best way of connecting them is each pair in series and then the 3 pairs feeding a combiner box where they each have a Schottky blocking diode feeding the cable to the MPPT charge controller. In this way, any pair of panels that is shaded is simply isolated until such time as it becomes illuminated. I have some more industrial Schottky diodes picked up at some flea market that work quite well for this purpose. I tolerate the minimal series voltage drop going through the Schottky diode as simply being the price I pay for convenience. The MPPT controllers that I've tested each have quirks about when they recompute and reset themselves to new input voltage and current conditions. One hopes that the controller will discover new operating points within a few minutes of things changing.

You can just bolt solar optimisers to those panels in the string. Essentially turns a string inverter setup in to a micro inverter setup without significant additional cost. I got them for my panels that get shade.

I know it's just semantics but impedances for AC on DC circuit's we call it resistance.

You can adjust the voltage range in SunnyBoy Inverter but you’ll need a dealer code. In the US this is for installations that are either first or last in line from the power plant.

Did you hear about the private electricity company(s) that want to reduce the payments or completely stop them to people with solar? They said they wanted to reduce the payments or completely stop them made in order to offset what those without pay. I believe it is wrong simply because they are extensively profit driven and are focussing on those they are paying out money to with solar as it is eating into company profits. The way they worded and described it was in terms of the money people were getting by putting electricity back into the grid, they are getting way too much money, this is causing a divide between those with and those without, and is the reason the prices were increased towards those without solar in the first place. They already price hiked the electricity a number of times since the grids were privatised which affected millions of households with/without solar and deliberately neglected to say this. Therefore, the amount received by those with solar is going to be significantly reduced anyway and by blaming those with solar as the reason of the increased prices shows the extent these grubs will go to, to get their own way.

Good breakdown and interesting to see in real life how a small shadow is affecting the setup. For the breakdown of the panel, (from 18:30), there is also a "half-half" in that the diodes are in he middle and there are 6 strings (3 on top and 3 bottom) allowing it to produce at lest 1/2 power when top or bottom is obstructed.

Increasingly, panels are now constructed with their own internal bypass diodes (e.g. that can bypass even subsections of the panels). So these newer panels would take care of them problem as well.

If your mains voltage is out of spec, the inverter will usually shut down completely, or at least throw an error. Had an issue with a customer whos inverter would just shut down production at certain times of the day. Measured mains voltage, and one of the phases was really low, like 200V maybe. Went to check the main fusebox for the incoming supply, and one of them was absolutely scorching hot, and crackeled when given a little twist. The fusebox was from like the 1940s, and pretty worn out this point. So it turns out that everytime there was too high of a load on that phase the voltage would drop just enough to trip the inverter, and shut down until there was almost no load on that phase and it would work for a while.

I have a single panel on a battery reserve system, and my bypass diodes actually went dead short. I had to remove them in order to restore system operation. It was actually kinda funny how I discovered it. I went out one day and noticed that, while I had panel voltage, it wasn't nearly enough. So I took the panel down and checked the connection box. One of the diodes had come loose! So I reconnected it. Went back down, and ... wtf? zero volts? Checked the panel again, and the fresh flux fumes POURED out of the connection box, and both diodes were rather spicy. Since I'm only running one panel, I took the diodes out completely, and the system has run great ever since.

Start of the video my first reaction were the sight of a pole on the roof. I don't know the direction of the sun, but considering how the event clocked in at about the same time day after day, then it is most likely a shadow. Thermal shutdown would be more sporadic and dependent on ambient temps. Mains voltage is however logical given how many solar panels there is locally. So overvoltage protection kicking in wouldn't be strange. But the graph instantly killed that idea. Since no reason for it to drop so sharply a 20 to 10 in the morning, when it delivers just fine at 11. Though, long series strings can be sensitive to small things like birds flying past. I am though curious to the micro inverter solution. I have seen some other solutions using DC-DC converters instead and having a main high voltage DC bus going over to the inverter.

Would be interesting to see a test where you create a thin shadow earlier in the day for experimental confirmation.

I was talking about almost exactly this with my boss who worked both as a solar install contractor and as a grid engineer. The voltage problem is a really common issue actually. They have to oversize new transformers because of solar output at this point because some houses will push out 250-ish where the inverters cut off, that's 230 and 10% error, 253, which your inverter exactly hovered over. So that could still be a problem even if it wasn't the root cause.

Hi and thank's for the great Video and sharing your experince here. In theory, as probaply most, I have heared about this issue with (small) shadows but not seen it myselve yet. It's seriously a thing to avoid when installing modules in a string. In an other video I saw that the loss also depends on the orientation of the modul(s), depending on how they are internally wired together it might be a bit better to install them in landscape orientation. Best is of course no shadow... ;-)

The whole Shading effect on String Systems vs Micro Inverters was completly busted just a couple of months ago by a Bunch of Aussie Solar Installers using proper scientific methodology. There is no difference between Enphase Micro Inverters being shaded and String Inverters. Dave hit the nail on the head when he said it was his Inverter. I suspect he has an old SMA unit that was made before SMA added "Shade Fix" their patented shading fix algorithm that fixes their own mistake. BTW every modern Inverter already has algorithms that prevent this issue. Here is a link to the Video and there are others who in the last five weeks have repeated his tests. 100% proof that Shading in Strings is no different than Micro Inverters. kzfaq.info/get/bejne/i7dpht2Z0ZfHlYU.html

Interesting logical derivation and explanation from a practical example!