Alternative to microinverters

[cegue]

Hi, I want to start a personal project for my small countryside house. However, the lack of a similar product in the market is making me doubt its feasibility.

The house is surrounded by trees, so shading is always present during the day. There’s no way to improve this by changing the panels’ layout. Thus, a parallel connection is necessary, and a friend suggested using microinverters. This is fine, but I already have an inverter and still want to use it. So, this idea came to mind:

Why not create an isolated DC-DC converter to be placed on the back of each solar panel? The output of each converter would be >320V DC, and then the main inverter would transform it to 230V AC.

Why doesn’t this product exist on the market? They would be cheaper and less complex than microinverters

Edit:
Similar to this product
https://www.ti.com/lit/ug/tidu404/tidu404.pdf

[Siwastaja]

Thus, a parallel connection is necessary,

Wat???

Why not create an isolated DC-DC converter to be placed on the back of each solar panel?

There already exists something much better: non-isolated DC-DC converters to be placed on the back of each panel, then those in series, together generating high enough voltage for the inverter, yet allowing separate MPPT control per each panel. These are called "optimizers" and are widely available.

Microinverters or optimizer based system are both OK for your case.

[Marco]

They're called power optimizers, but they are usually used with series connection (and I assume they just bypass under full illumination).

For parallel they would have to be mandatory, with the serial solution they can be optional.

[nctnico]

Optimisers are one way, but you can also think about arranging the panels in groups which receive shading together. That is how I organised the system on my roof.

[Siwastaja]

Also worth considering:

1) If shading affects most of the panels most of the day, consider not doing the install at all. Only in such case, microinverters/optimizers provide significant advantages over string inverter, but then again the system is more costly and still producing much less than a less obstructed system.

2) If shading, at any point in time, only affects small number of the panels (e.g., significantly less than 50% of panel count), a string inverter could do surprisingly well: partial shading is not as big as a problem as microinverter sales folks say. Panels have bypass diodes: only the production of the shaded panel drops to zero, rest produce fully, this is what microinverter or optimizer sellers want to lie to you about to make their case.

e.g. with 10 panels out of which 2 are shaded, string inverter produces 0.98*(8 * 100% + 2 * 0%)/10 = 78.4% of rated power, while microinverters produce like 0.95*(8*100% + 2*30%) = 81.7% of rated power. Not much of difference even if you had two panels out of ten shaded for the whole day! If you have partial shading for only an hour or so, you already lost your gains in lower efficiency of the micros.

But if the shading affects many panels at the same time, then make sure the rest are capable of exceeding the inverter's minimum MPPT and minimum startup voltages, otherwise exactly what microinverter folks claim will actually happen. If like 5 out of 10 are shaded more than some half and hour or so, consider micros/optimizers or not doing it at all.

If you have enough panels to satisfy two strings of a string inverter, then play the game as nctnico says. Then again if you are only barely able to exceed the minimum MPPT/startup voltages, then not dividing into two strings might be better after all.

[tszaboo]

Why doesn’t this product exist on the market? They would be cheaper and less complex than microinverters

Solaredge uses DC-DC converters, called power optimizers. They are still connected in series, so all the POs share the same current, while voltage changes with the solar input.
It works quite effectively, I have it on my roof.

[nctnico]

e.g. with 10 panels out of which 2 are shaded, string inverter produces 0.98*(8 * 100% + 2 * 0%)/10 = 78.4% of rated power, while microinverters produce like 0.95*(8*100% + 2*30%) = 81.7% of rated power. Not much of difference even if you had two panels out of ten shaded for the whole day! If you have partial shading for only an hour or so, you already lost your gains in lower efficiency of the micros.

I don't have exact numbers, but partial shading of only 1 panel influences the solar setup I have much more than the numbers you write above.

I have the same inverter brand (Growatt) as this guy and I think the relative influence is similar:


I'll try to post some numbers from my own inverter when there is some shading over the panels. But it could be the Growatt inverters are not good at dealing with shading.

Anyway, if there are trees obstructing the solar panels, the best way is to trim the trees down a bit. Using optimisers or micro-inverters adds costs which can make a solar install un-economic quickly especially since many energy companies seems to be making it less economic to supply solar energy to the grid nowadays.

[Siwastaja]

I don't have exact numbers, but partial shading of only 1 panel influences the solar setup I have much more than the numbers you write above.

The key question is, does the string's (minus 1 panel) total MPPT voltage fall below inverter's minimum MPPT input voltage? This is what happens in a marginal design with minimum number panels per string. If not, then another explanation could be very poor MPPT tracking algorithm which gets stuck in local minimum, but I would be surprised to hear this still being a problem (unless your Growatt inverter is already pretty old. I don't remember all the model numbers and when they were introduced.)

[nctnico]

It is not a sunny day but covering a few cells makes the output power for that string go down to 355W from 480W. I also see the DC voltage going up for that string. In this particular situation it means a reduction of around 25%. The inverter is less than 2 years old.

[tszaboo]

Here is my production data for today. The panels below get a bit of shading, reducing their output. 11x360W system.
The optimizers don't make the system more expensive, so IMHO it's worth it. You don't need DC isolator with these, and the inverter is cheaper than comparable inverters. When I bought the system, the optimizers were ~10% of the total system cost, including installation.
Though it's never going to beat the price of a system with cheap chinese inverter and bottom of the line panels.
Apparenly you can get panels now for 60 EUR, so get more panels.
https://www.solar-outlet.nl/jinko-440wp-all-black-tiger-neo-n-type.html

[Siwastaja]

The optimizers don't make the system more expensive, so IMHO it's worth it. You don't need DC isolator with these, and the inverter is cheaper than comparable inverters.

Note this is only true in some particular region / market. Here no one buys optimizers, therefore it is hard to find installer for them, and they would likely increase the cost of the system by at least 30% and never pay back for themselves. Obviously, if you can get the deal you describe, it would be fool's errand not to get them.

[cegue]

Thanks everyone for the clarifications.
Yes, power optimizers look like a solution, and probably it would be the best one in my case.

But in general, would not it better to have isolated DCDC converters? Parallel connection, with every PV panel having a converter, stepping up the voltage to >320V and then a central inverter.

I can only see benefits over series connection with optimizers or microinverters...
- PV panels are isolated from HV, hence less risk of shock if there are cracks
- Reduced wire gauge from PV panels to the central inverter
- The inverter does not require step up stage, hence less complex
- If one PV panel breaks or is shutting down for maintenance, the rest can continue working

Why this solution has not been adopted in the industry??

[tszaboo]

Because for parrallel connection you need twice the amount of wiring than in series. And standard wire thickness is wasted there.
Panels breaking is a non-issue. You have maybe a week downtime in 25 years.
The inverter doesn't require a step up stage with solaredge either.

I don't think that power optimizers are going to be economical for the future. 2 MPPT input microinverters, like the Hoymilles  HMS-2000-4T make them obsolete. It accepts 4 panels. It has 2x MPPT, so shading losses are a good compromise, and it's very very cheap.
The whole solar industry is just trying to get a good compromise. There will be always better systems, but if they aren't economic ...

[Siwastaja]

I don't think that power optimizers are going to be economical for the future. 2 MPPT input microinverters, like the Hoymilles  HMS-2000-4T make them obsolete. It accepts 4 panels. It has 2x MPPT, so shading losses are a good compromise, and it's very very cheap.
The whole solar industry is just trying to get a good compromise. There will be always better systems, but if they aren't economic ...

Yeah, plus also classic "string inverters" having larger number of MPPT inputs and lower minimum MPPT voltages than before. You can get a small one with 2-3 MPPT inputs and 150V min voltage, that is pretty close to what microinverters have to offer already, so the gap is smaller than before (e.g. 1 MPPT input and 200V min voltage).

E.g. my older inverter really requires 12 panels to work well. Nowadays a 8-9-panel system is possible with nearly every string inverter and some are happy with 6.

[nctnico]

Modern solar panels also have higher output power (due to bigger sizes and improved efficiency) so you need less panels for more power. In my install I have 2 strings with 6 panels each which add up to 4500Wp. The 4.2kW inverter I have has an MPPT range starting from 80V. So it would be happy with 3 to 4 panels. From what I've seen from micro-inverters is that many can't handle the maximum power of the solar panels unless you buy the really expensive ones. In my case that matters because my roof has a near perfect angle and orientation.

[Siwastaja]

From what I've seen from micro-inverters is that many can't handle the maximum power of the solar panels unless you buy the really expensive ones. In my case that matters because my roof has a near perfect angle and orientation.

This is an interesting note, because microinverters are marketed with the idea they will give the highest possible output over the wider operational area than string inverters (e.g., insolation differences). Yet if you get a 400W microinverter for a 400W panel, you will be flattening out the production during better-than-rated conditions and these conditions do exist in real world maybe up to +25% of rated, basically the case of partially cloudy day, combined full direct sunlight + indirect light from light clouds, maybe boosted by cold temperature.

It's not a huge loss for annual total production but microinverters are marketed by the very idea they extract the every last drop of what is available. Which is also why the 2-3 percentage point worse efficiency is kind of a problem IMHO.

[max_torque]

Given the huge fall in costs of solar panels recently (such that for my system, the mounting rails and bracketry cost MORE than the panels themselves!!) the first priority is simply to fit every single square mm of space with panels, ie buy more panels not more optimisers!

The second point is imo, if you array is roof mounted and therefore diffcult to access, putting active electronics in this environment is a false gain. Any failure here is going to hint you hard finanically speaking because of the effort required to replace parts up their (esp true these days in countries with strong HSE rules for "working at height")

[Someone]

Chasing the last few percent is usually uneconomic:
https://www.nrgsolar.com.au/blog/enphase-fronius-testing-station
which takes you to:

[nctnico]

From what I've seen from micro-inverters is that many can't handle the maximum power of the solar panels unless you buy the really expensive ones. In my case that matters because my roof has a near perfect angle and orientation.

This is an interesting note, because microinverters are marketed with the idea they will give the highest possible output over the wider operational area than string inverters (e.g., insolation differences). Yet if you get a 400W microinverter for a 400W panel, you will be flattening out the production during better-than-rated conditions and these conditions do exist in real world maybe up to +25% of rated, basically the case of partially cloudy day, combined full direct sunlight + indirect light from light clouds, maybe boosted by cold temperature.

It's not a huge loss for annual total production but microinverters are marketed by the very idea they extract the every last drop of what is available. Which is also why the 2-3 percentage point worse efficiency is kind of a problem IMHO.

In my opinion micro-inverters make no sense at all. In every scenario they are more expensive compared to a string inverter. And when you oversize the string inverter (as mentioned in the video Someone linked to), you can get the little extra on a cold (say 10 degrees C) sunny spring / fall day. Going one or two steps up for a string inverter costs like 50 to 100 euro extra for the entire install.

For micro-inverters the prices easely double. For example: the cheapest Enphase micro-inverter I can find at the shop I bought my solar gear from costs about 63 euro and can handle 295W. So that would add up to 756 euro / 3540W for my installation. The 380W micro-inverter costs a whopping 125 euro which would add up to 1500 euro / 4560W. And for Enphase you'd need to buy their control box and relay (which I've seen burned out; from looking at the board, the voltage rating is a tad low on the protection components).

In comparison: a 5kW, single phase inverter costs 539 euro (including an option to connect batteries to it). The 5kW inverter would really take every bit of power from my solar panels as they can produce around 5kW on a sunny spring / fall day. The panels rated for 375W deliver around 417W in that situation (11% more than rated). Then again, it does cost extra and the question is whether it is worth it. After all, every bit of money you spend on the solar installation eats into the ROI.

[Kjelt]

In my opinion micro-inverters make no sense at all.

Until a single solar panel malfunctions, you get 0 output and you have to go up to the roof to replace it or let a company replace it and you are directly break even on the costs of the company replacing a single panel. 

It makes perfect sense, there is not a single point of failure anymore and you can wait till a couple panels have broken down.

[Siwastaja]

Until a single solar panel malfunctions, you get 0 output and you have to go up to the roof to replace it or let a company replace it and you are directly break even on the costs of the company replacing a single panel. 

It makes perfect sense, there is not a single point of failure anymore and you can wait till a couple panels have broken down.

So why are not cars equipped with ten parallel tires on each axle so that you don't need to stop and change tires if one blows? I will answer that myself: because that would reduce efficiency, add cost, and tire failures are rare enough that replacements are a better option than redundancy. Or why only server PCs have dual power supplies, but ones used at offices (still doing work that matters) not?

Engineering needs numbers and analysis, and that is why larger strings win in all professional installations, and microinverters are only sold to households where they can be explained by oversimplified "common sense", and often outright lies.

And don't get me wrong, there is still potential in microinverters; distributed over centralized is often a good choice, but it's all about spot-on engineering. The situation changes as soon as an "AC panel" is available i.e., a pre-wired panel + microinverter, with easy to install AC wire harness systems (plug and play like MC4 is supposed to be), which costs at most the same as equivalent string inverter system. But so far the microinverter manufacturers seem to struggle even supplying legal, safe and up-to-date with power ratings products. It's not a nice market to be in, some fundamental design paradigm shift is needed to make it work: for example the "AC panel" plus "safe AC plug" concepts, avoiding the electrician on the roof.

For micro-inverters the prices easely double. For example: the cheapest Enphase micro-inverter I can find at the shop I bought my solar gear from costs about 63 euro and can handle 295W. So that would add up to 756 euro / 3540W for my installation. The 380W micro-inverter costs a whopping 125 euro which would add up to 1500 euro / 4560W. And for Enphase you'd need to buy their control box and relay (which I've seen burned out; from looking at the board, the voltage rating is a tad low on the protection components).

It seems microinverter manufacturers have not kept up with the trend of increasing panel size, efficiency and thus higher output power. 63 euros sounds appealing, until you realize 295W maximum power is suitable for a 270W or so rated power panel, which is ten years old technology. You might be able to find a good deal of some new old stock on such ancient panels, but... Probably better just to buy modern stuff.

[nctnico]

In my opinion micro-inverters make no sense at all.

Until a single solar panel malfunctions, you get 0 output and you have to go up to the roof to replace it or let a company replace it and you are directly break even on the costs of the company replacing a single panel. 

It doesn't have to be. I have made my installation so that I can service all electrical connections to and between the solar panels from the windows in the roof. If 1 panel fails, I can disconnect it and bridge it if I want to. But that is a perk of how I did my installation.

More realistically, a micro-inverter has a lot more components in it (and subjected to a very harsh environment) compared to a solar panel which is mainly a passive element. There is probably a 10 to 1 chance a micro inverter breaks before a solar panel does. You also need to consider what could cause breakage. In case of a solar panel breakage, this is likely a weather related event which is covered under insurance so replacing a solar panel wouldn't be super costly. I'd be more worried about sourcing a panel of the same size as panels seem to get bigger and bigger.

And don't get me wrong, there is still potential in microinverters; distributed over centralized is often a good choice, but it's all about spot-on engineering. The situation changes as soon as an "AC panel" is available i.e., a pre-wired panel + microinverter, with easy to install AC wire harness systems (plug and play like MC4 is supposed to be), which costs at most the same as equivalent string inverter system. But so far the microinverter manufacturers seem to struggle even supplying legal, safe and up-to-date with power ratings products. It's not a nice market to be in, some fundamental design paradigm shift is needed to make it work: for example the "AC panel" plus "safe AC plug" concepts, avoiding the electrician on the roof.

This would only work if solar panels are part of the roof and money is saved on roofing material. IOW: A dual use solution. So far the track record of such systems isn't great. In case of fire, the fire brigade can't get under the roof to put the fire out. There have been some serious incidents in the NL with such systems causing way more damage than necessary. Another problem is long term availability of replacement roofing material. Roof tile size & shapes (clay or concrete based) are pretty much standarised and you can still buy the correct tiles for homes built 60 years (or longer) ago.

[mikeselectricstuff]

In my opinion micro-inverters make no sense at all.

Until a single solar panel malfunctions, you get 0 output and you have to go up to the roof to replace it or let a company replace it and you are directly break even on the costs of the company replacing a single panel. 

It makes perfect sense, there is not a single point of failure anymore and you can wait till a couple panels have broken down.

But more hardware ( microinverters) means more potential failure points.
How often do panels go bad ?

[mikeselectricstuff]

Why not create an isolated DC-DC converter to be placed on the back of each solar panel? The output of each converter would be >320V DC, and then the main inverter would transform it to 230V AC.

Why doesn’t this product exist on the market? They would be cheaper and less complex than microinverters

Edit:
Similar to this product
https://www.ti.com/lit/ug/tidu404/tidu404.pdf

FYI there are a number of cheap inverter "back-end" PCBs on Aliexpress that can do the DC-AC conversion if you want to persue this :
https://www.aliexpress.com/item/4001030294883.html?spm=a2g0o.order_list.order_list_main.225.1ece1802XhFtR8
https://www.aliexpress.com/item/1005005448265485.html?spm=a2g0o.order_list.order_list_main.230.1ece1802XhFtR8
https://www.aliexpress.com/item/1005001937739215.html?spm=a2g0o.order_list.order_list_main.236.1ece1802XhFtR8

Most seem to be based on this fairly well-dcoumented controller
https://www.aliexpress.com/item/1005006160607383.html
https://server4.eca.ir/eshop/000/EGS002/EGS002_inverter_digitale_a_onda_sinusoidale_pura_1.pdf

[Siwastaja]

More realistically, a micro-inverter has a lot more components in it (and subjected to a very harsh environment) compared to a solar panel which is mainly a passive element. There is probably a 10 to 1 chance a micro inverter breaks before a solar panel does. You also need to consider what could cause breakage. In case of a solar panel, this is likely a weather related event which is covered under insurance.

Yeah, and this is not just like idle talk. Surprisingly good reliability of PV panels is well known, and so is the fact that microinverters do fail and are harder to replace than the string inverter. Probably someone like Someone comes up with a good reference again as I fail to deliver.

The argument is that production does not go to zero when either a panel or inverter fails. But what now, you are being sold a system which has higher likelihood of failure (just with smaller consequence), are you supposed to then just live with 90% of the production you used to have for the rest of the system lifetime? No way, of course you will replace the broken panel or broken inverter. So why push it further in the future, obviously you want to fix it ASAP to get back to full production. But then again grid tied PV generation is not mission critical that say 2 weeks of downtime with MTBF of say 50 years would be any kind of problem. You have such "downtime" regularly anyway due to dark, cloudy weather. So what you lose in a string inverter is minimal anyway.