bill_zunderman
New Member
Since I haven't seen much about this - I figured I'd write up my experience for anyone else to follow. Using a victron MPPT controller on a van with a 48/51VDC system is a tough thing to do. It requires a lot of panels in series - a minimum of 4 "12V" panels in series for charging to commence. Even if you had the minimum of 4 panels, then shading on 1 small cell on any panel would pretty much stop all charging. I had installed 4x100W and 2x160W panels in parallel in the old setup, so I would've lost that extra 60W in those panels if I'd just put them in series. I wanted a DC based solar optimizer solution, but I couldn't find anyones similar experiences searching the Internet.
I'd read mixed reviews on Tigo, and they seem to list a 25% mismatch limit that my setup exceeds. In hindsight they would probably work just fine as shaded panels produce much less than 75% of unshaded panels all the time. I did get some unlocked/IndOp enabled Solaredge p320s from aliexpress, as people claim that you can mix and match any panels with solaredge without any issues. They did work good, but they output around 37V at no load and slowly adjust down to the panel Vmp when loaded up. The main issue was the Voc of 4 units ran around 154V, which exceeded the input voltage of the 150/35 MPPT controller I'd gotten. I didn't want to buy a giant/expensive 250V MPPT and operate at 5% capacity just to run 6 solaredge optimizers. With 3 solaredge modules in series (2x160W, 2x100W, 2x100W panel setup) the Victron MPPT would drive the string Vmp around 70V. The solaredge unit with 2x160W panels would basically run in passthrough, and the units with 2x100W panels would drop the output voltage to boost the current to match the output current of the 2x160W panels. While it worked, I was still hoping to get an optimizer per panel since shading varies all the time on the roof of a van.
I ended up going with some alibaba MPPT controllers - 6 x Sunwin 600W units. They seem relatively new to the market. There's not many reviews on them anywhere, but they appear quite basic in their circuitry compared to Tigo and Solaredge. They come fully encased in a plastic housing more similar to Tigo, and at least based on surface temperature they appear to be quite efficient. I'm only running 100W and 160W panels, so not really pushing their 600W rating either. Their sales did advise me that mismatching panels was not optimal, but I'm basically starting off suboptimal with trying to reach 63V with 6 mismatched panels in series on the commonly shaded roof of a van. Cost per unit was only like 25USD each, but shipping and alibaba fees ran an extra 100USD. So far they've worked well, but I've only been testing them for a couple of weeks. I did some modifications to greatly reduce their bulky mc4 cables and package size. Will update if I have any issues.
Victron basically says not to use optimizers with their controllers. There shouldn't be any technical reason why it wouldn't work. The IV curve is very different coming off an optmizer as there's an inverted knee where the optimizers will adjust the current while preserving most of the power. I'm imagine victron is just warning that their MPPT might not perform optimally with solar optimizers. Victron seems to run the MPPT algorithm very quick and PV cells would probably respond very quickly. The current matching buck driver in these optimizers probably isn't nearly as quick to respond. I would've expected string current to match the Imp of the 160W panels, but the Victron MPPT algorithm seems to operate in the 64-68V range which pushes all 6 optimizers into buck mode. About every 10-15 mins the MPPT does a global scan to check higher/lower voltages to ensure it wasn't optimizing at a lower output local mppt, but every time this runs it usually lands right back in the mid-60v range. I'm sure it's losing a couple of watts here and there, but having all 6 panels actively contributing and keeping the string voltage high has provided the greatest daily production numbers in all my testing thus far. It'd be cool if I could measure the efficiency, but it would probably require professional calibrated gear to accurately measure the 1% loss here and there.
Reason for all this - In the stock setup in my van, the MPPT actually did DC boosting on the output allowing a 15vdc string voltage to get boosted to 58.2vdc peak absorption voltage. Volta relabeled a Kisae MPPT controller that had limitations that I wanted to ditch - lots of heat, noisy fan, clicking relay, 10A output limit that clipped power, no telemetry, fixed 14s NMC charging profile, and only a dry contact remote charging control. With the old MPPT I had 6 x 32 cell "12v" panels in a 2S3P setup. Since the 3 strings were in parallel, I maximized the panels to 720W with 2x100W, 2x100W, 2x160W. This mismatch worked fine in parallel, but needing to reach 63V+ for a traditional buck based MPPT required putting the panels in series.
I'd read mixed reviews on Tigo, and they seem to list a 25% mismatch limit that my setup exceeds. In hindsight they would probably work just fine as shaded panels produce much less than 75% of unshaded panels all the time. I did get some unlocked/IndOp enabled Solaredge p320s from aliexpress, as people claim that you can mix and match any panels with solaredge without any issues. They did work good, but they output around 37V at no load and slowly adjust down to the panel Vmp when loaded up. The main issue was the Voc of 4 units ran around 154V, which exceeded the input voltage of the 150/35 MPPT controller I'd gotten. I didn't want to buy a giant/expensive 250V MPPT and operate at 5% capacity just to run 6 solaredge optimizers. With 3 solaredge modules in series (2x160W, 2x100W, 2x100W panel setup) the Victron MPPT would drive the string Vmp around 70V. The solaredge unit with 2x160W panels would basically run in passthrough, and the units with 2x100W panels would drop the output voltage to boost the current to match the output current of the 2x160W panels. While it worked, I was still hoping to get an optimizer per panel since shading varies all the time on the roof of a van.
I ended up going with some alibaba MPPT controllers - 6 x Sunwin 600W units. They seem relatively new to the market. There's not many reviews on them anywhere, but they appear quite basic in their circuitry compared to Tigo and Solaredge. They come fully encased in a plastic housing more similar to Tigo, and at least based on surface temperature they appear to be quite efficient. I'm only running 100W and 160W panels, so not really pushing their 600W rating either. Their sales did advise me that mismatching panels was not optimal, but I'm basically starting off suboptimal with trying to reach 63V with 6 mismatched panels in series on the commonly shaded roof of a van. Cost per unit was only like 25USD each, but shipping and alibaba fees ran an extra 100USD. So far they've worked well, but I've only been testing them for a couple of weeks. I did some modifications to greatly reduce their bulky mc4 cables and package size. Will update if I have any issues.
Victron basically says not to use optimizers with their controllers. There shouldn't be any technical reason why it wouldn't work. The IV curve is very different coming off an optmizer as there's an inverted knee where the optimizers will adjust the current while preserving most of the power. I'm imagine victron is just warning that their MPPT might not perform optimally with solar optimizers. Victron seems to run the MPPT algorithm very quick and PV cells would probably respond very quickly. The current matching buck driver in these optimizers probably isn't nearly as quick to respond. I would've expected string current to match the Imp of the 160W panels, but the Victron MPPT algorithm seems to operate in the 64-68V range which pushes all 6 optimizers into buck mode. About every 10-15 mins the MPPT does a global scan to check higher/lower voltages to ensure it wasn't optimizing at a lower output local mppt, but every time this runs it usually lands right back in the mid-60v range. I'm sure it's losing a couple of watts here and there, but having all 6 panels actively contributing and keeping the string voltage high has provided the greatest daily production numbers in all my testing thus far. It'd be cool if I could measure the efficiency, but it would probably require professional calibrated gear to accurately measure the 1% loss here and there.
Reason for all this - In the stock setup in my van, the MPPT actually did DC boosting on the output allowing a 15vdc string voltage to get boosted to 58.2vdc peak absorption voltage. Volta relabeled a Kisae MPPT controller that had limitations that I wanted to ditch - lots of heat, noisy fan, clicking relay, 10A output limit that clipped power, no telemetry, fixed 14s NMC charging profile, and only a dry contact remote charging control. With the old MPPT I had 6 x 32 cell "12v" panels in a 2S3P setup. Since the 3 strings were in parallel, I maximized the panels to 720W with 2x100W, 2x100W, 2x160W. This mismatch worked fine in parallel, but needing to reach 63V+ for a traditional buck based MPPT required putting the panels in series.
