Using DC optimisers to connect dissimilar panels in series?

[meetyg]

Hey all,
I have a somewhat unusual setup and would like to maximize my solar panel output.

Currently I have 4x100w panels (2s2p) and 2x455w panels (not in series or parallel).

I currently have these connected to a grid-tied microinverter, with 4 separate MPPT ports.
But, I now want to connect these panels to my 3k 24v AIO which has a max PV input of 145v (up to 2000w).
I'm ditching the grid-tie because I formally don't have an export contract and thus was just a test setup for learning etc...

Vmp of the 2s2p is around 36v and each of the 455w panels around 42v. Isc for the 2s2p is 5A and 11A for the 455w panels (each).

I would like to connect them in series so that I will be in the sweet spot of the single MPPT on the AIO inverter/charger.

However, as they have different voltages and current ratings (which would limit the efficiency of the overall PV output), I was thinking of using DC optimisers as follows:
1 optimiser for each of the 455w panels, and a third for the 2s2p 400w array. All three will be in series. I was looking at some SUNWIN 600W DC optimisers. These can receive up to 60v at input, and limit the output voltage to 41v. So 41v times 3 = 123v, which sounds good to me. This will also handle any higher VOC due to cooler weather (which is usually 5-10 degrees Celsius at the lowest).

As I understand, DC optimisers will allow each "panel" to output its maximum, while bypassing the delta current to the output, generated by other panels in the string.
Basically, in theory it's as if I had 3 similar panels, but one is constantly shaded.
Am I understanding correctly?

To add complication, the 2s2p is at a different angle (almost 45 degrees) than the 2 x 455w panels (10 degrees) to the sun, but all at the same azimuth.

So in short, will this work well?
My other option was to parallel all 3, but to use blocking diodes between them. But I'm not sure how well that would work out.

Thanks for reading my lengthy question. I would like to hear your thoughts!

 

[Hedges]

I think the optimizers are good for about 25% difference in current. Should work within those limits.

You may find that an optimum array of used PV panels is cheaper/better.

 

[meetyg]

I think the optimizers are good for about 25% difference in current. Should work within those limits.

You may find that an optimum array of used PV panels is cheaper/better.

Thanks.
All my panels were bought new. My problem (being setup on a pergola) is usable space.
The 4x100w array was added to fit the wing of the pergola, trying to maximize space.

 

[Hedges]

If multiple orientations on the pergola, that will also cause current differences. Which will exceed 25% at some angles.
Best to avoid that, put multiple orientations in parallel, not in series.

What is minimum MPPT voltage of your inverter?
Vmp of the panels?

Even if series is sweet spot, I think having only same orientation in series will be more important that highest efficiency of MPPT.

 

[meetyg]

If multiple orientations on the pergola, that will also cause current differences. Which will exceed 25% at some angles.
Best to avoid that, put multiple orientations in parallel, not in series.

What is minimum MPPT voltage of your inverter?
Vmp of the panels?

Even if series is sweet spot, I think having only same orientation in series will be more important that highest efficiency of MPPT.

Minimum MPPT is 30v.
Vmp STC is 41v for the 455w panels.
Vmp STC for the 100w panels is 18v each.

So maybe I should put the 4x100w in series (giving 72v) and then the two 455w in series (82v) and then parallel ?
Sounds a bit counter-intuitive having a 910w string in parallel with a 400w string... But hey, that's why I'm here, to learn...

Should I use blocking diodes in between the parallel strings?

 

[Hedges]

900W parallel with 400W is OK.
72V in parallel with 82V isn't great. Close to 15% difference, which isn't too bad when both in full sun. But the 72V string in series, when in shade, will drag it down.

2 strings in parallel, usually no recommendation of fuses. But 450W panels would feed excess current. into 100W.

Ordinarily I don't recommend diodes, not needed. But given lower voltage 72V string and lower allowed backfeed current, could help to put a diode in line with it.

No promises, but I think it could work.

 

[meetyg]

900W parallel with 400W is OK.
72V in parallel with 82V isn't great. Close to 15% difference, which isn't too bad when both in full sun. But the 72V string in series, when in shade, will drag it down.

2 strings in parallel, usually no recommendation of fuses. But 450W panels would feed excess current. into 100W.

Ordinarily I don't recommend diodes, not needed. But given lower voltage 72V string and lower allowed backfeed current, could help to put a diode in line with it.

No promises, but I think it could work.

Great!
I guess I'll have to do some testing, to make sure theory meets reality.

I'll probably try with and without diodes and compare.

Since my PV is currently grid-tied via Deye microinverter (with monitoring), I know what to expect in regards to peak production and daily production.
Anyways, I wanted to know if DC optimisers would help.

I thought that I wouldn't be the first to try this, but haven't seen much mentioned online, connecting dissimilar panels via optimisers.
If others have any experience doing this, I would like to hear your observations.

 

[Hedges]

I guess back to your original idea, can you do this with one optimizer for 4s "12V" panel, or maybe 2s2p?

Since they remain mismatched a modest percentage of voltage, it could work.
But you may need a keep-alive transmitter (at least I think Tigo optimizer uses that), which blows the budget.

And SCC for DC coupling to your battery would be another way to go.

 

[meetyg]

I guess back to your original idea, can you do this with one optimizer for 4s "12V" panel, or maybe 2s2p?

Since they remain mismatched a modest percentage of voltage, it could work.
But you may need a keep-alive transmitter (at least I think Tigo optimizer uses that), which blows the budget.

And SCC for DC coupling to your battery would be another way to go.

The SUNWIN optimisers I was looking at don't need any keep-alive transmitters. They are also fairly cheap (around 50$ each).
I guess I could get just buy 3 and experiment.
Not much info on them online (at least not in English), but the specs look good to me.
They don't have any monitoring communication either, which is a shame.
They have a maximum input voltage of 60v, so I can't connect the 100w as 4s to one optimizer. But 2s2p would be fine.

I am trying to avoid the extra wire run from the panels to another SCC. That would also require separate monitoring. The 400w array catches the early morning sun, while the 455w panels catch the noon sun.

Here is what the power production looks like on a typical sunny day (PV1 and PV2 are the 455w panels respectively, PV3 is the 2s2p 400w array):


Maximum daily production during the summer is around 5kwh. Since they are on a pergola, the sun is blocked by the house after noon.
Again, this is currently with a 4 port microinverter, each port being a separate MPPT (currently only 3 ports in use).

As you can see, I'm not getting close to the STC ratings on these panels, maybe closer to NOCT, but that's for another thread I guess...

So to summarize, you are saying that maybe going with the optimisers is the best/simplest way ?

 

[Hedges]

Try a single optimizer (just $50) for 2s2p, 400W.
Can you tilt panels different orientations? 2s one orientation, 2s another, the larger panels can each have any orientation. Reduced current through electronics, runs cooler. Recharges battery sooner, keeps it charged later (if you have more production than battery can store.)

But I'm not sure how optimizers behave in parallel with each other, or other panels. I think they are meant to be in series, see a current, adapt to hit that.

So an all in series setup might match optimizer's functionality. Again I'd try to use just one. Not sure if it can boost voltage (reduce current) or just buck (lower voltage, increase current.) Probably the latter. Have to consider when designing string or array with optimizers.

 

[meetyg]

Well I did an experiment this morning:
Connected each of the 455w panels and the 400w array to the SUNWIN 600W optimisers.
They are pretty simple in that they don't have communications or monitoring, but it seems they do the job (how well, I don't really know, I would like your opinion).
The 3 SUNWIN are connected in series, giving ~123v at max.
Here are the specs:


So today was a bit tricky because of clouds here and there, but I got 830w at peak.
When all panels were connected to the grid-tied microinverter (each on its own MPPT), I saw no more than 970w at peak, on sunny days.

So here are the output charts for today (all panels are shaded by the house around 1-2pm, so that's basically the end of my daily solar production for today.


I really don't know why my panels are not producing close to thier STC (or even NOCT), but that was a problem with the grid-tie too (separate MPPTs).
Currently 20 deg. Celsius outside. The most I saw from the large panels was around 360-380w at peak (each), during summertime.


Anyways, from the voltage and current graphs, what do you think?

BTW, the SUNWINs are buck only (step-down), seem to clamp the max voltage @42v per panel. I like this because it means my voltage won't be too high for the 145v limit on the inverter.

I guess I will need a better sunny day to properly understand the efficiency of this setup. But my gut feeling is that it's optimal for what I have.