Can heterogeneous arrays be put in series?

[scpanish]

I never see heterogeneous arrays in series discussed. I an upgrading my 23 year old system and would like to retain my old panels, currently configured as three 24V (nominal) arrays in parallel. One array has 10 24V panels (10P, each panel Vmp 34.2V, Imp 3.5A), the other two arrays are each four 12V panels (2s2P, each panel Vmp 17.8V, Imp 7.3A). I'd like to make that all one string with it's own MPPT SCC, separate from a larger second string of new panels. The new battery bank will be 48V, so it would be nice to get the string voltage above that. The easiest way would be to connect the existing arrays in series for nominal 72V. But I never hear of this being done. Is there a problem?

Alternatively I can do it the usual way at 48V nominal with the 24V panels 2s5P and the 12V panels as two 4s arrays, all in parallel. But it would be nice for the MPPT to get the higher voltage/

 

[Hedges]

There have been threads on this.

Imp 3.5A, two in parallel for 7.0A
That should be OK in series with 7.3A

Or, 17.8Vmp, two in series for 35.6V
Should be OK in parallel with 34.2Vmp

You have 8x "12V" 17.8Vmp, 7.3A Imp.
You have 10x "24V"34.2 Vmp, 3.5A Imp

You cold connect 8x 12V in series for one string,
4x 24V in series for second string,
4x 24V in series for 3rd string.

Assuming your MPPT can handle the Voc, adusted or cold weather.
I have 12x "24V" panels in series for a 600V max MPPT, so yours probably would need 200V SCC.
A Midnight Classic 150 might handle it (HyperVOC survives higher voltages in cold weather but without operating.)
Do the math yourself to check limits.

That leaves 2x 24V unused.

 

[scpanish]

Thanks for the reply, I considered that 96V option but am trying to not waste those two old but good Solarex MX120 panels that would be left over. And the old MPPT is indeed a Midnite Classic 150, this is NH and even cold mornings would mostly avoid HyperVOC, especially with climate change, but it is very close.

So my 72V option, which uses all the panels and avoids HyperVOC is to keep the 3 arrays in their current form, 24V each, but connect the arrays (not the panels) in series for 3x24V= 72V. For the array 1, 10 24V panels, max current is 3.5A x 10 = 35A, each of arrays 2 and 3 are 12V panels in 2s2p each having 7.3A x 2 = 14.6A , the arrays 2 and 3 in series total 29.2A, in series with array 1 is 29.2A + 35A = 64.2A which is well under the 96V max for the Midnite 150. The panels are pretty compatible electrically (I think) even though the MX120's are polycrystaline and the Sun Electronics 130s are single crystal. They have been working together for a long time.

I don't know what would be wrong with this 'series array' arrangement but I haven't seen serial arrays like this discussed and I'm no EE. There are physical advantages as well, the 3 arrays are separate and the old wiring would mostly be reused, the 12 V panels date to 2006 and use MC4 connectors, while the old Solarex are just wires running through flexible watertight conduit. The combiner arrangement is going to be a bit of a nightmare no matter what I do.

I can do a conventional approach at 48V nominal. But the MPPT would be happier at 72V.

 

[Hedges]

If you do connect any 7A in series with 2p of 3.5A, a failure mode would be if one 3.5A went open-circuit, 7A shoved through other 3.5A would be too much for its bypass diodes.
Simply shading one 3.5A panel would do that. Shading both, 7A would not split evenly through two parallel diodes, so failure seems likely.

So I prefer purely XsYp arrays, OK to parallel strings of different model if they add up to similar voltage, OK to series connect panels of different model if all are similar current.

 

[scpanish]

Good point that I had failed to appreciate. Sounds like I should just do the conventional 48V approach. The old MSX120s don't even have bypass diodes, that was an option in 1999 and since the arrays are in a field and unshaded, I didn't bother. But I would not want a failed panel to take out another panel. I'm going to put in bypass diodes when I rewire and will get them over 7 amps in any case. Have to see what is available.

The 48V approach configures the 10 24v panels as 2s5p, paralleled with the 12v panels configured as two 4s arrays. Then all 3 arrays in parallel.

I've been using them all in parallel at 24v for years, may not be optimal for MPPT but still works fine to power the house. Need to scale everything up for mini-splits and a heat-pump domestic hot water heater.

 

[Hedges]

I'm going to put in bypass diodes when I rewire and will get them over 7 amps in any case. Have to see what is available.

A diode rated for 7A can't carry 7A unless heatsinked well enough. And operated at 25C ambient.
Some data sheets are conservative and some are optimistic. I've seen some particularly absurd ones in surface-mount (after they failed in our design.)
You have to review thermal parameters and design the heat conduction and dissipation scheme.

 

[scpanish]

So just to finish up this thread, all 3 arrays have now been reconfigured for 48V nominal and paralleled, and are working fine (for 2 weeks) with my Midnite Classic 150. I added 10A bypass diodes to the ancient Solarex MX120's. The Solarex are now wired as 3 strings, two with 4 panels as 2s2p and one with 2 panels as 2s. Those strings are paralleled in the equally ancient Solar Electric Specialties combiner, along with the 2 4s strings from the Sun Electric panels. So I'm using all the panels, and performance seems identical to what it was as a 24V system.

The rest of the summer is devoted to adding another 4KW of PV, new Conext Pro inverter, and Growatt SC48120-MPV charge controller. The home-built 29KWH LFP will be rewired to 48V from 24V.

Just FYI, Growatt support never answered questions (via email) about the charge controller. But it was cheap!