Yes I connect the panel directly to the battery. Again I'm learning as I go but would I found is that as long as your panel(s) don't produce too much current, then your fine charging this way. Keep in mind I'm talking about lead acid batteries, AGM's included. I'm sure it's better to go through the reduced current top off charge cycle. I wouldn't try to charge a LIPO battery this way. What I observed is that as the battery nears full charge the resistance increases and less current moves into the battery anyway. So there is some dynamic current reduction that way. I just put the ammeter in circuit and watched/logged what happened.

I had a diode on my controller but took it out as the panels I'm using have them already.

I'm just doing a pump for now. The idea is that the fountain runs when the sun is shining, but you can turn it off and just charge the battery if you want it to run at night.

Are you going to shine the light on the fountain or put it in the water or something? That looks like a good LED driver. There's a good article on this site about driving LED's. I think it's in the DC power section. You should take a look.

Yes I'm using a linear regulator, the L7805C. The switching regulators are more efficient but add a lot of parts and cost. Let me know if you find one that works well for you.

I'm building a panel that will put out 4 volts at 4 amps. I'll run the pic right off the panel then step up the voltage (using the PIC PWM and and inductor) so that it boosts the voltage for the maximum output needed to charge the battery or run the attached load. I think it should be pretty efficient as the battery will never be pulling the panel voltage down. I'll just boost the voltage to whereever it needs to be.

 

I'm building a panel that will put out 4 volts at 4 amps. I'll run the pic right off the panel then step up the voltage (using the PIC PWM and and inductor) so that it boosts the voltage for the maximum output needed to charge the battery or run the attached load. I think it should be pretty efficient as the battery will never be pulling the panel voltage down. I'll just boost the voltage to whereever it needs to be.

I will be really interested in that. I'd like to learn a bit more about PWM.

One thought I had was to use it for a voltage and current regulator but I would not know the first thing about building that type of circuit.

 

A simple boost converter looks like this:

http://en.wikipedia.org/wiki/Boost_converter

The PWM would be controlling the switch.

A simple buck (step down the voltage) converter:

http://en.wikipedia.org/wiki/File:Buck_circuit_diagram.svg

I think you could limit the current simply by running it through a transistor that you switched with the PWM output pin. If you vary the duty cycle then you're voltage doesn't change but less current is able to pass through.