Combining MicroInverters and String Inverters

[Siwastaja]

open modbus protocol on the inverter to read out data

To be fair, this seems to be standard feature on every random Chinese inverter these days. Sometimes with a bit of trouble finding the correct document with Google - e.g. Sofarsolar has at least three different interface specifications with no documentation to which model numbers they apply -, but you can always do it.

[nctnico]

I spend more money than necessary in other areas of life and derive pleasure from the extra expense so if people want to pay too much for DC optimisers or microinverters and feel good about it then so be it.

Of course but people should have access to the information and know what they are doing. Pay premium for premium features that actually do what is promised. We (as a player on the smart energy business) have seen aggressive microinverter sales companies who sell FUD (false claims about others) and false promises of their own systems.

I would like to see microinverter systems evolve and gain real traction on the market through real advantages but currently it seems they are falling behind in the competition, i.e. quality is going down (e.g., they have no resources to update their product lines for new higher-power panels -> hence they clip power, totally unacceptable!), while competion is going forward (string inverters having more and more MPPT inputs, smaller start/minimum voltages, battery option for very little extra cost, better efficiency...

I hate to see it when the underdogs don't get their shit together but instead start to play dirty. FUD and lying in marketing are reasons alone I will not recommend microinverters to anyone. If a fresh new player who does something new on the field pops up I'm super happy to change my mind.

I don't see microinverters ever becoming a viable option from a financial standpoint. Having a single unit (=string inverter) doing the job of 20 seperate units is always going to be cheaper due to manufacturing, maintenance and handling costs alone.

From what I've read from Enphase, getting a system running requires a lot of fiddling with software, registering online, etc, etc. They should drop all that crap and make their system work out of the box. Plug & play. That alone would save a lot on installation and maintenance costs. Nobody cares about monitoring.

[Siwastaja]

I don't see microinverters ever becoming a viable option from a financial standpoint.

I believe it could be possible but it requires a different mindset. I.e., they should be solving problems, not adding them. They concentrate on a few made-up/exaggerated problems and forget all the rest.

But there is room for improvements. E.g. DC cabling between panel and inverter. DC connector mess with MC4 and compatibles. Microinverters do not even try to do aynthing about this. You still wire the panels to the inverters. Shorter wires, yes. Still wires. Still connectors. Still authorities saying you must match the connector brands. How the fuck you are supposed to do that, buy Enphase branded MC4 connectors, strip off the existing connectors from panel stubs and re-crimp the Enphase MC4 connectors? At least on a string inverter you have much fewer connectors to crimp.

No, microinverters would need to integrate into the panel itself, to create an AC panel. Then it needs a well designed weather sealed AC plug system. Just like MC4, but single plug for L,N,PE. Then add a reliable and problem-free power line communication or wireless system. Engineer it all top-notch reliability, but also engineer it for low cost (increasing NRE to decrease unit cost), and then apply economy of scales to bring the final product price down. Which is chicken-and-egg, you need money upfront for all that work.

And do all of the above really well, there is not much room for error because competition from normal string inverters is harsh, and optimizers cater pretty well for those who want to improve trackability of panels + partial shaded production.

[nctnico]

That sounds like integrated solar panels to me 
It could be worthwhile but the major problem is that it will take time for the technology to mature and builders & architects to embrace this technology. IMHO even the retrofitted solar panels aren't mature technology at this point. The whole issue with non-compatible MC4 connectors blew my mind. I also think the panels are too big too handle safely and make a system harder to install than it should be. Needing to bolt the panels down is also not convenient. Especially when you need to take a single panel out. Some kind of click / lock system would be much better for maintenance but it would have to be an industry standard.

[tszaboo]

That sounds like integrated solar panels to me 
It could be worthwhile but the major problem is that it will take time for the technology to mature and builders & architects to embrace this technology. IMHO even the retrofitted solar panels aren't mature technology at this point. The whole issue with non-compatible MC4 connectors blew my mind. I also think the panels are too big too handle safely and make a system harder to install than it should be. Needing to bolt the panels down is also not convenient. Especially when you need to take a single panel out. Some kind of click / lock system would be much better for maintenance but it would have to be an industry standard.

If you had a professional installer, two guys can place an entire installation in like 3 hours.
Its really not big of a deal.

[nctnico]

That sounds like integrated solar panels to me 
It could be worthwhile but the major problem is that it will take time for the technology to mature and builders & architects to embrace this technology. IMHO even the retrofitted solar panels aren't mature technology at this point. The whole issue with non-compatible MC4 connectors blew my mind. I also think the panels are too big too handle safely and make a system harder to install than it should be. Needing to bolt the panels down is also not convenient. Especially when you need to take a single panel out. Some kind of click / lock system would be much better for maintenance but it would have to be an industry standard.

If you had a professional installer, two guys can place an entire installation in like 3 hours.
Its really not big of a deal.

Have you ever seen those 'professional' installers use any kind of safety gear? Most (if not all) don't because handling a big panel while being strapped in a safety harness doesn't go together so well. And wait until the wind picks up... I'm very sure professional installers would like smaller panels instead of larger. Keep in mind roof tiles are typically about 35x35cm; that is not some random coincidence. If it where more practical to make these larger, they would have been larger.

[Siwastaja]

If you had a professional installer, two guys can place an entire installation in like 3 hours.
Its really not big of a deal.

This is true, that's why I'm saying current situation is good enough so that competing against it will be tough.

Current systems are not ideal from reasons nctnico points out but popularity and competition has wed out inefficient install companies and forced those who want to remain in the market to invent efficient processes and deal with the details. Professionals can do two domestic install sites per day, so 10 install a week. Maybe installs could be made 5% faster and 50% simpler to learn but simplicity does not matter to professionals who do this and only this every day, and 5% faster is still the same 10 installs a week.

[nctnico]

If you had a professional installer, two guys can place an entire installation in like 3 hours.
Its really not big of a deal.

This is true, that's why I'm saying current situation is good enough so that competing against it will be tough.

Current systems are not ideal from reasons nctnico points out but popularity and competition has wed out inefficient install companies and forced those who want to remain in the market to invent efficient processes and deal with the details. Professionals can do two domestic install sites per day, so 10 install a week. Maybe installs could be made 5% faster and 50% simpler to learn but simplicity does not matter to professionals who do this and only this every day, and 5% faster is still the same 10 installs a week.

No. You are overlooking that the majority of the installers who do 10 installs a week do not adhere to work safety regulations at all. If you want to see acrobats without a safety net, go watch solar panel installers. Evel Knievel wouldn't do such crap if he where still alive. Needless to say this is not viable a long term  situation either through fines or Darwin awards. The proper installers either use safety harnesses or use scaffolding (which is more convenient). But they won't be doing 2 installs are day because dealing with the scaffolding alone takes a couple of hours by itself. IOW: if you want to make a more competitive solar panel solution, a good long term strategy is to create a system which is also safer to install.

[Siwastaja]

No. You are overlooking that the majority of the installers who do 10 installs a week do not adhere to work safety regulations at all. If you want to see acrobats without a safety net, go watch solar panel installers. Evel Knievel wouldn't do such crap if he where still alive. Needless to say this is not viable a long term  situation either through fines or Darwin awards. The proper installers either use safety harnesses or use scaffolding (which is more convenient). But they won't be doing 2 installs are day because dealing with the scaffolding alone takes a couple of hours by itself. IOW: if you want to make a more competitive solar panel solution, a good long term strategy is to create a system which is also safer to install.

I would want to agree with you, but it seems the current system is working, from the fact this has been going on for years and years. So it's probably more sustainable than you think. Just like it's a compromise on every other respect, it's also a compromise on worker safety. Clearly laws and regulations are ignored. Yet surprisingly few people actually die doing that. Go figure.

[nctnico]

True, but it leaves room for a manufacturer to create a demand for safer to install solar panels by swaying public opinion. In the NL the government is actively going after solar panel installers who work unsafe BTW as there are several lethal accidents each year and many more accidents with serious injuries. IMHO just leaving the situation as-is because only a few people die (or get seriously injured) is NOT acceptable. Especially if it is just to make a few bucks extra.

[Siwastaja]

Well, that's also one reason why I believe alternatives will have a chance. But they need to hit the nail in many respects simultaneously. Not introduce new problems; solve enough of the old. And bootstrap the market from scratch with massive investments so they can be affordable from day one to buy. Not an easy task.

Microinverters are crap because they introduce more new problems than they solve:
* More DC-to-DC connector interfaces
* Having to do both AC and DC connection
* Having to mount and wire up 20 inverters instead of 1
* Track record of reliability issues due to placement of inverters
* Worse production under some conditions, due to
  * Worse efficiency
  * Clipping of peak production power because of slow response to the market increasing panel power
* Higher idle consumption
* Possibly a dangerous product which does not have the mandated grid isolation relays due to cost optimization done in desperation by dying manufacturer

What problems do they solve?
* Increase in production under some other conditions
* Per-panel data for troubleshooting
* Wiring on the roof is AC instead of DC

[fastbike]

If the online claims about Solis handling MPPT "shadow scan" much better than others, that places Solis into pretty good position because it's not expensive. My biggest complaint about Solis is that external MODBUS RS485 port shares the same interface with the datalogger so you can only have either one, not both. For some companies this is a showstopper for our control box, for others they are fine accessing data through our service only instead of Chinese cloud.

Can you help with some advice on the RS485 port on the Solis. We have a S6-EH3P10K-H-AU inverter. I'm trying to use a Waveshare RS485/Ethernet bridge but cannot get it to respond to requests from Modbus Poll - I just get timeouts even though I can see the ACT light blinking.
I've followed the instructions at https://github.com/alienatedsec/solis-ha-modbus-cloud

[brucehoult]

[Have you ever seen those 'professional' installers use any kind of safety gear? Most (if not all) don't because handling a big panel while being strapped in a safety harness doesn't go together so well. And wait until the wind picks up... I'm very sure professional installers would like smaller panels instead of larger. Keep in mind roof tiles are typically about 35x35cm; that is not some random coincidence. If it where more practical to make these larger, they would have been larger.

I got ridiculed online when I was talking about the prices of "200W"(1 m^2) panels."You need to look at 400W panels" I was told.

Actually, the best prices in NZ now are for "650W" (3 m^2) panels.

I ended up buying 6x 440W panels JAM54D40-440/GB for US$60 + tax each. The shop was trying to push 645W LONGi LR7-72HVH-645M panels but I consider them to be past the size and weight that is comfortable for one person to load / unload / carry. I can deal with the 22kg 2m^2 ones.

But for sure 1m^2 would be more convenient if they were price competitive -- they might have a chance of fitting in my station wagon, which even the 2m^2 ones don't.

The 2m^2 panels I bought (and I think most of them) would only have about a 1/4" clearance (1/8" each side) between the wheel wells of a Ford Ranger, and only just fit in the longer bed of a single-cab, and not at all in a double-cab version (the most popular).

The 645W panels are the same width (1134mm) but 2382 long. That's almost 7'10" for Americans.

But anyway, I'm just using my panels on the lawn. They're never going on the roof (at least not at this rented house). They are organised in two strings of three panels, feeding into two 1200W MPPT controllers at around 100V 12A.

At this time of year they should be at a 35º angle, but actually with 10% over-paneling (2640W into a 2400W system) just about dead flat is fine. In December a 12º slope will be ideal, but 0º (or 6º with 2 concrete blocks on one side and one on the other) will be basically perfect.

Now that I've got to the point of spring that I don't need heating all night I'm effectively about 90% off-grid, as in the past few weeks I've been paying for on average no more than about NZ$0.50/day for electricity on top of the fixed NZ$1+GST ($1.15) daily charge. That's down from $5.44/day in October 2024, $8.86/day in October 2023, $6.74/day in October 2022. It's $1.66/day for the 9 days so far this month and it's actually been pretty rainy and/or overcast most of the time with just 800W-1200W produced most of the time.

That NZ$1.66/day is US$0.95, €0.82, £0.72.

Once we get into actually good weather that should go to $1.20-$1.30/day (it is already, some days)

On the sunny days my 6kWh battery is full again by midday (after getting me from ~6 PM to ~9 AM), and there is a ton of free electricity in the afternoon for aircon, cooking, clothes and dishes washing, etc.

[Siwastaja]

If the online claims about Solis handling MPPT "shadow scan" much better than others, that places Solis into pretty good position because it's not expensive. My biggest complaint about Solis is that external MODBUS RS485 port shares the same interface with the datalogger so you can only have either one, not both. For some companies this is a showstopper for our control box, for others they are fine accessing data through our service only instead of Chinese cloud.

Can you help with some advice on the RS485 port on the Solis. We have a S6-EH3P10K-H-AU inverter. I'm trying to use a Waveshare RS485/Ethernet bridge but cannot get it to respond to requests from Modbus Poll - I just get timeouts even though I can see the ACT light blinking.
I've followed the instructions at https://github.com/alienatedsec/solis-ha-modbus-cloud

Note that the RS485 pinout is different for every Solis S6 model series, go figure. Try RS485 A = RJ45 pin 5 (blue-white if you use standard Ethernet cable) and RS485 B = pin 4 (blue). Also, always try swapping A and B. Check slave address - 1 by default on Solis I guess, but worth verifying from the inverter settings.

These Homeassistant based things get so complicated with dozens of failure points - a massive number of hardware boxes then massive number of software layers - so they are nearly impossible to debug. It's a hobby in itself.

[Siwastaja]

I got ridiculed online when I was talking about the prices of "200W"(1 m^2) panels."You need to look at 400W panels" I was told.
...
I ended up buying 6x 440W panels JAM54D40-440/GB for US$60 + tax each.

So it seems you got the message, and bought the right things 

So we were right. Sorry if it felt like "ridiculing".

[tszaboo]

I got ridiculed online when I was talking about the prices of "200W"(1 m^2) panels."You need to look at 400W panels" I was told.

I think you went into a discussion opinionated, with bad information, and you just got some helpful info on how to fix your misconceptions. Are you salty, that you got help?

[nctnico]

[Have you ever seen those 'professional' installers use any kind of safety gear? Most (if not all) don't because handling a big panel while being strapped in a safety harness doesn't go together so well. And wait until the wind picks up... I'm very sure professional installers would like smaller panels instead of larger. Keep in mind roof tiles are typically about 35x35cm; that is not some random coincidence. If it where more practical to make these larger, they would have been larger.

I got ridiculed online when I was talking about the prices of "200W"(1 m^2) panels."You need to look at 400W panels" I was told.

Actually, the best prices in NZ now are for "650W" (3 m^2) panels.

I ended up buying 6x 440W panels JAM54D40-440/GB for US$60 + tax each. The shop was trying to push 645W LONGi LR7-72HVH-645M panels but I consider them to be past the size and weight that is comfortable for one person to load / unload / carry. I can deal with the 22kg 2m^2 ones.

But for sure 1m^2 would be more convenient if they were price competitive -- they might have a chance of fitting in my station wagon, which even the 2m^2 ones don't.

That is exactly my point. IMHO while these solar panels offer best price / performance, they aren't primarily made to be put on roofs but are made for use in large scale solar farms where they use lifts of even robots to install panels. A bigger panel simply means less installation costs per surface area. During the long planning phase (years!) of my own solar install I noticed that panels got bigger and bigger making the power output look higher but the increase in power output per surface area didn't increase that much. I have 375Wp panels (19kg each) on my roof but per surface area they have the same power output compared to the panels you got.

A panel of over 2.3 meters in length and nearly 30kg in weight, would be very hard to handle if you have to mount it on a roof. And their size makes them less flexible to get to an optimum layout on a roof in where it comes to chimneys, dorms and roof windows.

Anyway, at least you have the space to put the solar panels in the garden (and didn't have to haul them onto the roof >6m high like I did). Although, in your case I would try to get proper mounts for flat roofs which are weighed down. The way the panels sit now, they are likely to be blown away by the first storm.

PS: are those banana trees in the background? If yes, do they produce edible bananas?

[EEVblog]

But anyway, I'm just using my panels on the lawn. They're never going on the roof (at least not at this rented house). They are organised in two strings of three panels, feeding into two 1200W MPPT controllers at around 100V 12A.

Looks like nothing holding them down? That's begging for wind to make them airborne.

[tszaboo]

A panel of over 2.3 meters in length and nearly 30kg in weight, would be very hard to handle if you have to mount it on a roof. And their size makes them less flexible to get to an optimum layout on a roof in where it comes to chimneys, dorms and roof windows.

Some dude grabs it, walks up a ladder, and to installs it.
I see you want to over regulate this installation market. I also don't like walking on the roof, but that's OK, because I can pay someone else to do that.

[brucehoult]

I got ridiculed online when I was talking about the prices of "200W"(1 m^2) panels."You need to look at 400W panels" I was told.

I think you went into a discussion opinionated, with bad information, and you just got some helpful info on how to fix your misconceptions. Are you salty, that you got help?

I'm never opinionated about a field I'm just starting to explore! I simply presented the best deals I'd found until that point. All the "power station" manufacturers sell and recommend 200Wish panels, and they are certainly a convenient size. There is no obvious physical reason for them to be a massively worse cost per W than 400W panels -- slightly more, sure, given unit handling and stocking costs, and that they still need connectors and use slightly more framing material per square meter.

It still took considerable searching to find a local store with 400W panels for US$60 not $100+. I don't live in Europe or North America. It was at least useful to know that it might be possible. And even then it was am industrial park place 205 km from me staffed entirely by Chinese, most of them with very little English, and offering zero support.

They even wired up the MC4 cables I ordered from them incorrectly, with one conductor having male connectors on both ends the other having female connectors on both ends. The MC4 to XT60 cables that came with my equipment had it clearly labelled which was + and which - [1] but I hadn't noted which. We ended up debating this in the warehouse for 15 minutes. They weren't even able to tell me which connector on the panels was + and which - but it just seemed wrong to wire a cable they way they did -- that would mean it's impossible to plug a panel directly into the solar controller in the field. But never having done this before I couldn't be sure of my facts and could only ask if they were sure. I got home and plugged into one of the panels and measured the voltage and sure enough it came out correct directly attached to a panel but wrong polarity via their cable (and right if I put both cables they sold me in series). I had to cut off their connectors and crimp on new ones. NBD, but you'd think a shop could get it right. To their credit, they gave me four spare MC4 connectors for free, so I didn't have to buy some to correct the problem.

In the end it all works perfectly.  Yesterday was very overcast and I only got 60% charge into the 6kWh battery by 6 PM. That was down to 5% or so in the morning by the time I cooked breakfast and the sun got high enough. But then today the battery was at 100% by 2 PM (as well as running all the loads), and in the 90 minutes since then I've cooked dinner in the air fryer and now my bed sheets are in the washer, both using electricity that would otherwise not have been taken out of the panels, with the battery staying at 100%. The air fryer uses 1700W, and the clothes washer peaks at 2kW while heating water but only uses 0.55kWh in total in a 1 hour 40º C cycle. Spin drying uses 520W getting up to top speed but only 350W once up to speed, if anyone cares. There is similarly ample afternoon spare power to run the portable air conditioner (1000W).

[1] https://www.pecron.com/products/pecron-mc4-to-xt60-cable

[brucehoult]

Anyway, at least you have the space to put the solar panels in the garden (and didn't have to haul them onto the roof >6m high like I did). Although, in your case I would try to get proper mounts for flat roofs which are weighed down. The way the panels sit now, they are likely to be blown away by the first storm.

I've had winds up to 100 km/h in the month they've been sitting there. There's not even the slightest sign of wanting to budge. It's pretty sheltered, surrounded within 5m or 10m of the house on the north side, shed (and banana) on the south, huge bamboo hedge to the west, and citrus trees and feijoa hedge 15m to the east.

If a really big storm is coming it's the work of 5 min to move the concrete blocks and put them flat on the grass. Or, I can sling some cord through the holes in the panels and around the concrete blocks ... 12x 12kg = 144kg  for three panels, plus the 66kg of the panels themselves. At a 6º angle, with the shed in that direction? Not going to move.

I have a design drawn up using professional certified Clenergy components including CLNERTL3060 "30º-60º" adjustable rear legs (which I can make do a near-ideal 20º-50º with a slight geometry change), on a timber frame, with sandbags for ballast. I've done the wind loading calculations for that. But that stuff is going to cost around NZ$1000 -- more than the panels! -- and current experience shows it's absolutely not necessary from spring equinox to autumn equinox (at least), so I'm going to leave that until April or so.

In December, 12º is the ideal angle, so my current 6º on the concrete blocks (2 high on one side, 1 high on the other) is closer than 20º would be anyway :-)

PS: are those banana trees in the background? If yes, do they produce edible bananas?

Absolutely! Delicious! The problem is you get 100 all ripe at the same time.

[brucehoult]

But anyway, I'm just using my panels on the lawn. They're never going on the roof (at least not at this rented house). They are organised in two strings of three panels, feeding into two 1200W MPPT controllers at around 100V 12A.

Looks like nothing holding them down? That's begging for wind to make them airborne.

At the moment. They haven't shown any signs of budging in 100 km/h winds (reduced a lot my the nearby shelter) so far. I'm going to tie them down to the concrete blocks before summer storm season. There not a lot of lift possible at that flat angle anyway. It will be a different story when I make up proper adjustable 20º to 50º mounting before winter. There will be sand bags involved. I've already calculated how many.