Tigo Optimisers on a string with 2 orientations not optimising as hoped

[EastCoastDIYer]

Yep 60mph gusts tonight forecast for here and the local highway bridge is closed - good idea re the lasso, the winds are from the South-West so it will fall away from the PV panels nicely - should pass the SWMBO test

Well, you're localish to me then I guess, if you're referring to the Orwell Bridge. Sadly I missed these posts and so missed my opportunity to pull 200KG of brick chimney off my roof in the middle of the night!

My theory on why they spilled over the two planes is that 5 is on the edge of minimum string length. As well it’s possible if 5s is on low end of MPPT TIGOs deciding to buck by 25% for no reason might make the string drop out.

Possibly. At the time they were more concerned about String A being too big, rather String B being too small. I can't recall if String A was described as being a wattage or voltage concern, but for whatever reason, having understood the numbers better myself and done the maths, and the maths being backed up by various users in this thread (thanks guys!) I've realised that String A can have 13 and be safe, and String B can have 5 and always be above VMin in normal conditions. Looking at my Tigo's logs, for whatever reason, the voltage output on B is pretty consistent apart from in very low light when there's no meaningful current anyway.

My installer has agreed to re-wire B1, B2 and B3 onto String A. They were initially hesitant to do so as they weren't comfortable with a 13 panel string A being "close" to VMax, but with a little encouragement, they're going to do it.

I'm interested in seeing if String A optimising takes any longer in the morning to start working what with panels added to it from closer to the West Facing roof being shaded for longer. But whatever the result of that is, the flip side to that is that in the summer months in the evening, the West facing panels will all be in the sun and String A will be shaded, so I don't have to worry about B1, B2 and B3 dragging the rest of B down now. At worst, Swings and roundabouts, at best an overall better performing system.

 

[SeaGal]

Well, you're localish to me then I guess,

Yes, there are a few of us on here from sunny Suffolk and Essex.

and the maths being backed up by various users in this thread (thanks guys!)

and gals

My installer has agreed to re-wire B1, B2 and B3 onto String A.

do keep us updated when that is complete - will be good to see and understand the differences.

 

[Hedges]

English may not be a Romance language that thinks inanimate objects must be assigned gender, but at least those of us from an earlier generation understood some terms were gender neutral in the plural, don't have to say "he or she" etc.

Is "guy" gender-neutral?

The plural "guys" definitely is, at least here in San Francisco — I'm often hearing all-girl companies here being greeted with 'Hi guys, how are you doing?'. How about the singular guy? Is it

english.stackexchange.com

On this forum, as in engineering, we value talent and integrity. Chromosomal configurations aren't a consideration for membership and acceptance.

 

[SeaGal]

I think it's catching on over here... the younger generation saying "you guys" and all that - it seems we gradually pick up new ideas from across the pond; like Halloween and drive-thru-fast-food outlets... Who knows, we might even start using miles again (oh wait, we do)

 

[Quattrohead]

Just wait until you get a drive-thru liquor store, spaz the shit out of my English brain ?

 

[Hedges]

Heard about that years ago.

Texas drive-thru booze: What to know and where to get it

It is very easy to crave an icy and sweet alcoholic beverage while sweating in the car due to hundred-degree July and August temperatures.

www.click2houston.com

During Covid, California wanted to keep some businesses operating, so to-go drinks from bars were allowed here too.

No distilled spirits to go in Texas, Huh? That's why people figured out fractional freezing was another method to separate chemicals. Doesn't contain the word "distilled". Of course, can't reproduce what is crafted in the Scottish highlands, lowlands, etc.

 

[millsan1]

FWIW, I am happy with my TIGO optimizers, but not happy with support. I have 2 that are performing poorly, or not at all, and tech support keeps asking me about my VPN, which I do not have.

I have 2 extra optimizers and 2 extra panels, and will figure it out myself in the Spring.

 

[zanydroid]

tech support keeps asking me about my VPN, which I do not have.

They’re just doing slow psychological torture on you.

 

[EastCoastDIYer]

By the way, do rooftop PV panels not cause fires in the UK?

I was going to say "No" as normally any instance of something which could be considered "Green" not performing exactly as it's supposed to would have the Daily Mail coming out with the pitchforks saying they're going to kill us all/take our jobs. However, apparently it DOES cause fires here. The Fire Prevention Association reported that from Jan 2023 to July 2023 there were 66 fires related to solar panels, whatever "related" means. In 2019 it was 63 for the whole year....

 

[EastCoastDIYer]

do keep us updated when that is complete - will be good to see and understand the differences.

Will do. For Science!

 

[400bird]

I was going to say "No" as normally any instance of something which could be considered "Green" not performing exactly as it's supposed to would have the Daily Mail coming out with the pitchforks saying they're going to kill us all/take our jobs. However, apparently it DOES cause fires here. The Fire Prevention Association reported that from Jan 2023 to July 2023 there were 66 fires related to solar panels, whatever "related" means. In 2019 it was 63 for the whole year....

Can you find related statistics for other categories? Electrical or gas issues?

66 fires in all of the UK for an entire year sounds like a similar volume to the number of toothbrush related hospitalizations.

 

[SeaGal]

66 fires in all of the UK for an entire year sounds like a similar volume to the number of toothbrush related hospitalizations.

Agree is very small number & it's not something I've ever seen mentioned on the news.

However, the vast majority of installation in the UK are small (< 16A / 3680W) grid-tied ones driven by the (previously) generous feed-in-tariff's and the simplicity for connecting <16A systems (in that it is just inform after install, rather than needing prior approval).

So I'm guessing here but, because of that, most installations will be 1 or 2 single serial strings - hence with no paralleling of panels maybe the fire risk is mitigated?

 

[Hedges]

3 or more parallel panels is where per-string OCP is required. Only an issue if a string shorts (maybe not so likely). Failed bypass diodes would also be a short, and can happen due to shadows while other panels are getting full sun. A few could fail, but not all because no other panels producing enough to burn them up. With only some shorted panels in series, backfeeding would not be all that much current.

I think single strings are just as likely to cause fires as multiple paralleled strings. It is poor series connections, overheating and arcing (IMHO). Poor quality connectors is one cause, and interspecies mating another. With MLPE, you get two of those for every PV panel.

(Said as an armchair failure analyst. I've only done it professionally perhaps once, not for this particular application, but it was for a fire caused by electronics.)

 

[Sigunas]

My Tigo TS4-O optimisers on one of my strings is not behaving as I would expect and would like a second opinion on what Tigo support have left me with please.
Taking B1 as an example, it's neighbour A5 at 11:20 is producing 303 watts from 35v x 8.66amps. B1 is producing 30 watts from 39v x 0.77amps. Every other of the 8 string B panels is producing at least 0.71amps at 34v or higher. B1 continues to underperform compared to A5 until about 12:00 when over the next 30 minutes it's amps ramps up until it's matching A5. At 12:00 there's still one panel, B5, with 0.78 amps due to some shading and being west facing, but the rest of the west facing have more.

View attachment 189846

And current
View attachment 189848
Thank you, good people!

Optimizers work as DC-DC down (buck) converters.
Lets pretend you have all 8 panels in one direction, but one of them is shaded and it produces only 2A@33V, when all the others could produce 9.4A@33V. In this case optimizer would convert shaded panel's 2A@33V down to something like 9.4A@7V and you would get 7* (9.4A*33V) + 9.4A*7V ~= 2235W@238V. All is good.
In your case 3 panels could produce 9.4A@33V, but the rest 5 panels could only 1.5A@33V each. So in ideal world 5 optimizers should down convert shaded panels to 9.4A@5.2V, i.e. 3* (9.4A*33V) + 5*(9.4A*5.2V) ~= 1175W@125V. All is good? No, because 125V are probably too low to drive yours inverters MPPT input. So your inverter and optimizers "agree" (maybe incorrectly) that the best they can do is something like 8* 1,8A*33V ~= 475W@264V.

 

[EastCoastDIYer]

Optimizers work as DC-DC down (buck) converters.
Lets pretend you have all 8 panels in one direction, but one of them is shaded and it produces only 2A@33V, when all the others could produce 9.4A@33V. In this case optimizer would convert shaded panel's 2A@33V down to something like 9.4A@7V and you would get 7* (9.4A*33V) + 9.4A*7V ~= 2235W@238V. All is good.
In your case 3 panels could produce 9.4A@33V, but the rest 5 panels could only 1.5A@33V each. So in ideal world 5 optimizers should down convert shaded panels to 9.4A@5.2V, i.e. 3* (9.4A*33V) + 5*(9.4A*5.2V) ~= 1175W@125V. All is good? No, because 125V are probably too low to drive yours inverters MPPT input. So your inverter and optimizers "agree" (maybe incorrectly) that the best they can do is something like 8* 1,8A*33V ~= 475W@264V.

Thank you Sigunas, that makes a lot of sense. I hadn't thought about how the voltage drop to get the amps up could take it below voltage threshold at times.
So, for my Solis inverter S5-EH1P6K-L with 2 strings, the start-up voltage is 120v, the max input is 600v. The MPPT voltage range is 90-520v, so from my understanding, once 120v is reached, even a little shading taking voltage below 120 will be fine as long as it's above the 90v for the MPPT.

Something else I don't understand is, is the start-up voltage for the inverter as a whole (as in, as long as one of the strings is meeting this start-up voltage) or per string MPPT? If it's at inverter level, then String A which is running happily and making a lot more than 120v will be waking up and keeping the inverter running. Then the whole of String B could run off of the voltage of just 3 fully lit panels as that would be more than the 90v needed for the MPPT, right?

 

[zanydroid]

Thank you Sigunas, that makes a lot of sense. I hadn't thought about how the voltage drop to get the amps up could take it below voltage threshold at times.
So, for my Solis inverter S5-EH1P6K-L with 2 strings, the start-up voltage is 120v, the max input is 600v. The MPPT voltage range is 90-520v, so from my understanding, once 120v is reached, even a little shading taking voltage below 120 will be fine as long as it's above the 90v for the MPPT.

Something else I don't understand is, is the start-up voltage for the inverter as a whole (as in, as long as one of the strings is meeting this start-up voltage) or per string MPPT? If it's at inverter level, then String A which is running happily and making a lot more than 120v will be waking up and keeping the inverter running. Then the whole of String B could run off of the voltage of just 3 fully lit panels as that would be more than the 90v needed for the MPPT, right?

Tigo does not worry about excessive buck voltage, not sure why. Could be oversight could be they have some dark magic like expecting the MPPT to draw less current as part of the sweep so it naturally drops buck ratio. At the expense of lost matching capability. It is likely there are situations where it interacts badly

I believe start voltage is per MPPT, and it needs Voc above start to get going/continue. While Vmpp needs to be within operating range for efficiency and to continue.

The optimizers will likely also manipulate the relationship between Vmpp and Voc to a non standard ratio

 

[SeaGal]

Something else I don't understand is, is the start-up voltage for the inverter as a whole

I believe start voltage is per MPPT, and it needs Voc above start to get going/continue

I concur. The Solis S5-EH1P6K-L will run on battery voltage alone, so that's just 45V+ for the inverter's electronics.

 

[Sigunas]

The MPPT voltage range is 90-520v, so from my understanding, once 120v is reached, even a little shading taking voltage below 120 will be fine as long as it's above the 90v for the MPPT.

You are correct. According to the specifications, 125V should be completely sufficient for the inverter’s MPPT to start and operate in MPPT mode. However, it is unclear whether at such a voltage it can actually perform a full search for the best working point and is not stuck at the local MPPT.If you have access to all inverter parameters, you can check if “Global MPPT” (or something similar) is enabled. It is recommended that this parameter be turned off with optimizers. But you can try, maybe it could do. On the other hand, if you have already agreed with the installer to switch 3 panels to another string, it may even be a better solution.

Something else I don't understand is, is the start-up voltage for the inverter as a whole (as in, as long as one of the strings is meeting this start-up voltage) or per string MPPT?

I can’t answer 100% about your inverter, but in my Huawei inverter at home and Solis 25kW inverter at work, each mppt input has to reach the operating voltage separately. That is, until a specific mppt has reached the minimum start voltage, it does not contribute to power generation.

 

[EastCoastDIYer]

Possibly a final update to this thread to relay back the practical findings: On the 1st Feb the PV installer came by and re-wired B1, B2 & B3 into String A. I waited until now to try and get a day with clear sunlight in the mornings to see what sort of difference this made.

Old setup with String B across South and West facing:
28th Jan 09:19 String A west most 8 panels started generating proper amounts of power with each panel making >140watts. East most String A panels, shaded made ~35 watts each. String B1, B2, B2 south facing, shaded made ~20 watts. The significance of this time is it shows what sort of time in the morning String A is able to optimise and generate meaningful power. Looking back at previous days, this tracks as the norm.


28th Jan 11:17. There shouldn't have been any shading on the south facing panels. You can see that B1, B2 and B3 are being limited by the rest of B being shaded as the sun was behind the roof:


New setup with String A being all 13 south facing panels including what was B1, B2 and B3. String B only being West facing 5 panels now:
12th Feb 09:15. String A west most 8 panels producing > 140watts. East most 5 panels including B1, B2 and B3 (ignore the "B", only my installer can change them to "A"s in the EI Tigo website, and they haven't yet) are producing less as shaded by roof and chimney. Optimisers doing their job it seems as B1 which is between two shadows is producing the same as the rest of A unshaded panels.:


12th Feb 11:10. It's like night and day compared to the same time on the 28th Jan. All south facing panels on the same string producing power nicely. Actual string B still in the shade hasn't really started producing proper power yet, as expected. FYI, String B reaches the per panel wattage of String A at 15:19, and its panels start producing more power than String A's as the day comes to an end, at 16:18 they produce over twice the per panel wattage, extending the power capture over the day, which is exactly what they were there for. If only I could persuade the missus to have panels on our 93 degrees east facing roof - there's room for about 13 there...


So, overall I'm glad I changed B1, B2 and B3 to be on string A. I do wonder if String A is now starting to produce proper power a little later in the morning as a higher proportion of panels are shaded. The next really sunny morning we get, I'll check what actual shading there is and how much the String A panels are producing. But right now, it's cold and cloudy again. However, looking at the total generated power numbers, even taking into account how closer we are to summer, I'm getting an increase in power generation across the day. On Sunday it was producing 6200watts at one point, which I'm pretty pleased with for a 8kW PV setup in Feb at a latitude of 52 here in East Anglia. My inverter can convert a max of 6kW into AC, and my power network has a cap of 5kW export where I live, so it was maxing that out, running the house load and (slightly) charging the battery at the same time. Excellent, more of that please.

Thanks to everyone for their support in making this decision.

 

[wattmatters]

Thanks to everyone for their support in making this decision.

Well done, good to see the hive mind get this one right.

It's a good example of how, even with optimisers, installers need to take care with string arrangements.