Solar power installations are generally designed as if the load[s] will be steady around the clock.
In practice, actual loads may be heavier in the daytime (as when working with power tools) or at night (as when watching movies or running a forced-air furnace).
Due to the nature of battery charging there are better times and worse times to run non-essential loads. If done attentively (or automatically) one can run electrical loads that do not affect the battery's State of Charge at all.
This is especially true for Lead-Acid chemistries that do not take much current in Float or late Absorption stages.
It is common for Lead-Acid batteries to consume C/100 or C/200 amps to Float, meaning that during Float you will have 99% or more of your panels' output available for loads.1)
During Absorption the charge current is tapering down as the battery accepts less; the more it tapers down the more is available for loads.2) At the very beginning of Absorption ~0% of “bonus” output is available. At the very end of Absorption ~99% is available. Mid-way through Absorption3) ~49% of the panels' output is available for loads.
There will generally be more “surplus” power available when using MPPT charge controllers v. PWM,4)5) though the difference is less dramatic during periods one would run opportunity loads (see below). Siphoning off power can actually help PWM controllers run cooler by reducing the ON-OFF switching activity that generates heat.
The best time to run non-essential loads is when there is surplus current over what battery charging needs. This means during Float, or after current has started dropping off in Absorption.
For lead-acid batteries the loads are best applied in Float stage, or once past the beginning of Absorption. At the start of Absorption the controller still needs near-Bulk levels of current for charging. At the end it needs almost no current. Halfway through Absorption about half of the system's peak power will be available for loads. See the image at the top of the page for an example of how current demand drops in Absorption.
For LiFePO4 set the LVD to Vfloat; there is no actual Absorption stage _per se_ in fractional-C charging. We use these setpoints only because our traditional lead-acid controllers work that way.
You can start the loads manually (ie, start using the power when you have extra). This is error-prone but is free and requires no equipment (except your own memory).
The most precise way to do it automatically is to use a charge controller which turns on the LOAD output only when the batteries are in Float stage. Charge controllers with this feature tend to be expensive.
A free way to automate opportunity loading is to set the normal controller's LOAD or external Low Voltage Disconnect to shut off below Vfloat.
Since the LVD only knows voltage and Vabs > Vfloat this approach will
To run loads heavier than the rating of the LVD or LOAD output use a relay between the LVD and the load.
This setup looks like: [LVD or controller LOAD output] –> relay –> load
A 12v timer could give the system a chance to make some progress in Absorption before starting up opportunity loads.
In this approach the LVD is set just below Vfloat as above, but activation of the loads is delayed by some amount of time. Observation of the system during charging will suggest how long it normally takes takes charging amperage to drop off (ie, when the system has surplus current).
Suggested delay for conservative opportunity loading == the time from [passing Vlvr setpoint during Bulk] to [completed charging].7). This delay might be 2-3hours.
Suggested delay for aggressive opportunity loading == the time from [passing Vlvr setpoint during Bulk] to [far enough into Absorption that enough current is available to power intended loads]. This delay might be 30-60 minutes, depending on the charging rate.
This setup looks like: [LVD or controller LOAD output] –> timer –> relay –> load