String voltage mismatch fixed using an extra PV panel & DC-DC buck converter

[cyd]

The site has 3s3p. Two of the strings are identical, 1125W @ 103.5 Vmp. The third string is 1200W @ 93 Vmp & mismatched by ~10v @ Vmp. As you can imagine the inverter MPPT is tracking wildly.

So I added a forth panel in series with the 1200w string. Except that this single extra panel is wired through a buck converter and two series blocking diodes. The buck converter is set to ~12v and is bypassed by a single diode. (See picture)

The MPPT still tracks wildly but the power output remains relatively constant. The power and voltage charts were sampled for two minutes under varying cloud conditions.

Food for thought anyway

 

[sunshine_eggo]

Testing when the output of a ~3500W array is only 500-1000W.

How does the unmodified 3S3P array perform in direct sun as opposed to "varying cloud conditions?" where erratic output is expected.

How does the 3S2P array perform in those same conditions (omitting the mismatched string)?

What MPPT?

 

[cyd]

Testing when the output of a ~3500W array is only 500-1000W.

How does the unmodified 3S3P array perform in direct sun as opposed to "varying cloud conditions?" where erratic output is expected.

How does the 3S2P array perform in those same conditions (omitting the mismatched string)?

What MPPT?

The two 1125w strings face southeast & the other four panels face northwest.

Output today was 2551 KVA @ 89 Volts about 2pm. The power factor of 0.96 means that actual power draw was 2143W (+ 210W charging the battery bank)

The MPPT is integrated into an AIO easun 5kva unit.