The regulator was built on a perfboard and mounted in a sealed plastic
project box, which in turn got an aluminum mounting plate. The connection
terminals are brass. It was solidly built to withstand the harsh
maritime environment and a fair amount of abuse.
When the panel isn't generating, the entire circuit is off and there is absolutely no current drain from the battery. When the sun gets up and panel starts producing at least 10 Volt, the LED lights and the two small transistors switch on. This powers the regulator circuit. As long as the battery voltage stays below 14V, the operational amplifier (which is a very low power device) will keep the MOSFET off, so nothing special will happen and the panel current will go through the Schottky diode to the battery.
When the battery reaches the trigger voltage, which is nominally 14.0V, U1 switches on the MOSFET. This shorts out the solar panel (a condition that is perfectly safe), the battery no longer gets charging current, the LED goes off, the two small transistors go off, and C2 powers the regulator circuit while slowly discharging. After roughly 3 seconds, C2 has discharged enough to overcome the hysteresis of U1, which switches the MOSFET off again. Now the circuit will again charge the battery, until it again reaches the trigger voltage. In this way, the regulator works in cycles, with each OFF period being 3 seconds, and each ON period lasting for as long as necessary for the battery to reach 14.0V. The pulse length will vary according to the current demand of the battery and any load connected to it.
The minimal ON time is given by the time C2 takes to charge up with
the current limited by Q3 to roughly 40mA. This time is quite short, so
this regulator can work down to very short pulses.
The MOSFET can easily be replaced by any type you like, as long as its RDSON is low enough so that its dissipation will remain acceptable at the maximum current delivered by your panel. For D2, basically any diode is acceptable as long as it can safely handle the total current produced by your panel. A Schottky diode like the one shown is an advantage because it will produce only half as much voltage drop as a standard silicon diode, and thus generate only half as much heat. But a standard diode is perfectly suitable if properly sized and mounted. With the components shown, the regulator comfortably handles a 4 Ampere panel. For larger panels, only the MOSFET and diode need to be replaced by larger ones. The rest of the circuit remains the same. No heat sink is required for the power level shown. The indicated MOSFET can handle a much larger panel if fitted with a modest heat sink.
R8 in this circuit is 92k, which is a nonstandard value. I suggest that
you use an 82k resistor in series with a 10k one, which is simpler than
trying to find a special resistor. R8, R10 and R6 define the cutoff
voltage, so it's nice if they are reasonably accurate. I used 5% resistors,
which usually are a lot better than the rated 5%, but if you want to be
on the safe side, use 1% resistors here or pick the more precise 5% ones
after measuring several with a digital meter. You could also include a
trimpot in this circuit, so that you can adjust the voltage, but I would
not suggest this if your application calls for high reliability in a corrosive
environment, like mine did. Trimpots just do fail in these conditions.