PV homerun to parallel loads

In general, MPPT controllers do not like to share a single solar array.

The MPPT controller basically does an "I/V" curve of the array (vary current to "solve" Pmp=Vmp*Imp). The older controllers would simply "rescan" the array every 5-15 minutes and use the "new" Vmp-array number (typically around ~81% or so of Vmp-array-stc) and modulate Imp-array to keep Vmp-array approximately correct (call it +/- 5%-10% or better of Vmp-array-measured).

If your MPPT controller is using less than Imp-array-at that instant in time, then Vmp-array will drift up toward Voc-array (which is also temperature dependent).

When you have two MPPT controllers trying to "solve" that equation, and then attempt to control Imp-array and Vmp-array... They tend to get confused.

So, the intent is to share the single solar array with two different loads. The "optimal" solution would be to have all of the solar energy go to the battery bank (since you already have a battery bank), and then send the power to the various loads as needed and let the battery bank buffer the energy.

Of course, that gives you more losses (charging battery, DC to AC inverter, etc.) and wear on the battery bank.

And how much cost would sharing the array save... Say that 50% of the array energy goes to the battery bank, and you choose 10% rate of charge (5% to 13% rate of charge is typical for solar):

• 550 Watt array * 0.77 panel+controller derating * 1/0.10 rate of charge = 4,235 WH of storage
• 4,235 WH storage / 12 volt battery bank = 353 AH @ 12 volt battery bank
• 353 AH * 12 volts * 0.85 AC inverter eff * 1/2 days storage * 0.50 max discharge = 900 WH per day of stored energy (overnight, cloudy weather)

And based on hours per day of sun for summer (fixed array, biased towards summer production):
http://www.solarelectricityhandbook.com/solar-irradiance.html

Wilmington
Average Solar Insolation figures

Measured in kWh/m2/day onto a solar panel set at a 71° angle:
(Optimal summer settings)

Jan	Feb	Mar	Apr	May	Jun
3.17	3.71	4.79	5.61	5.71	5.62
Jul	Aug	Sep	Oct	Nov	Dec
5.50	5.03	4.68	4.46	3.63	3.16

Say >5.0 hours of sun per summer day:

• 550 WH array * 0.52 off grid AC FLA battery system end to end eff * 5.0 hours of sun = 1,430 WH per day "summer average" harvest

And there is the issue that Batteries are the one thing that is easy to "kill"... If the battery bank is not kept happy, the batteries will not have a very long life. Flooded Cell Lead Acid batteries are cheap ("golf cart" 6 volt @ 200 AH ~$100 each, x4 for a 12 volt @ 400 AH battery bank).

Of course, you could use LiFePO4 type batteries (don't care about charging current, actual hours of charging current, and state of charge as much as FLA do)--But much more expensive (and typically need a Battery Monitoring System).

I worry that the battery bank will either be damaged from improper/non-optimum charging and that the "cost" of electricity (for a typical MPPT system, ~$2.00 per kWH (GC batteries last ~3-5 years typically) is a good starting estimate)--And the amount of battery energy per day is variable based on the A/C system usage (even more complexity to energy management).

I don't believe that diodes are needed to isolate your two loads (A/C and Battery Charging), and using an SSR to cycle the solar array voltage to the MPPT charge controller (SSR "ON" when "A/C" compressor is off)--Can sort of work.

But one of the "advantages" of "inverter compressor" A/C systems is that they are variable load. Ideally, your compressor is running 100% of the time at 50% (plus or minus) of rated power. So the SSR does not really work well in this case.

Another possibility is--If you have the ability--To program the MPPT Vmp voltage for both A/C and Charger... Set the charger Vmp manually (or with an offset) X volts over the A/C system... As the array Vmp drifts up (A/C uses less power), the MPPT solar charge controller will draw more energy.

Another cost issue is that "High Voltage" MPPT solar charge controllers tend to be much more expensive (>Vmp-std ~ 100 VDC rating). For example, you would need a ~40 amp (minimum) Solar Charge controller for a 550 Watt @ 12 volt raise:

https://www.solar-electric.com/search/?q=schneider+mppt

https://www.solar-electric.com/xaxwmp80amp6.html (high "600 Vmax" voltage 80 amp MPPT controller ~$1,100)
https://www.solar-electric.com/xaxwmp60amps.html (mid-voltage "140 Vmax" 60 amp MPPT controller ~$500)
https://www.solar-electric.com/c40.html ("low voltage" PWM charge controller 40 amps ~$110)

https://www.solar-electric.com/morningstar-ts-mppt-600v-48-db-solar-charge-controller.html (60 amp 600 Volt MPPT ~$1,600)

Anyway--My guess is that "saving" 550 Watts of solar (at less $1 per watt for larger panels that "Need" MPPT controllers" or ~$2 per Watt for 140 Watt panels that can work with either PWM or MPPT controllers)...

I don't think this will be cost effective... It will cost you more to harvest the excess power from the 1,100 Watt array than to simply build a second 550 Watt array "lower voltage" + Charger + Battery bank. And both the battery and end user will be "happier".

Anyway, just a starting point for discussion. Questions or corrections to my assumptions/guesses?

-Bill

Yep... Simple and reliable. Complexity is the bane of many solutions.

I had to design a large redundant computer system (voice mail system for telephone companies). I would try to use the simplest solutions to meet the requirement.

There were still lots of single points of failure--But nobody really cared as most of the system was redundant.

We even put in a very simple switch that would disable auto-shutdown (over temp, other issues) and the tech publications group refused to document the switch (jumper). Had a system that would not start because of the failsafe--And they mechanically exchanged a $1,000,000 system because it needed to be UP and they could not debug it or let it run without the auto shutdown (not in manual). Turned out to be cable where one pin was not fully snapped in during cable manufacturing.

-Bill