Hi,
I'm trying to figure out how one would go about building an MPPT. I have attached a document with three hand drawn schematics (bottom of this entry I think) in an attempt to figure out how one of these power trackers work, and I would really appreciate it if someone who has actually built one of these guys could read through my reasoning and tell me if I'm understanding this correctly.

Circuit 1:
The first thing we need to figure out is how to get the I-V curve. We have three variables, V, I, and R. Since we can't actively sweep V or I since we are dealing with a solar panel, the best we can do is to put a pot between the terminals of the solar panel, and then connect the positive and negative terminal to each side of the pot. Then we sweep the pot, while taking current and voltage readings to get the I-V curve. However, this is a rather manual procedure, so lets try to automate it. I propose we put a inductor in the short circuit path. At DC, this would just be a short circuit, but if we put a FET in there and create a AC signal, then we get some resistance, or rather impedance. So this way, we can have a MCU using PWM to vary the resistance value, and hence the amount of current, and now we can record an I-V curve for the solar panel. The equation R = 2*pi*freq*L is correct - I think. Does this hold for 50% duty cycle, or 95% as well?

Circuit 2:
This circuit just adds a bit of interface hardware to connect to the MCU. We add a voltage devider with large resistance in order to be able to tie this to the ADC of our MCU to measure the voltage. We also add a current sense resistor in order to measure the current, and now we are actually able to read our I-V curve. From this we can also calculate the power curve by just multiplying the two values. Now the question is, how do you set up the sampling with respect to the PWM applied to the FET?

Circuit 3:
This is my best guess at what the "actual" MPPT circuit would look like based on the previous two circuits. Again, the question arises as to the sampling frequency with respect to the PWM of the FET - or perhaps more importantly the phase shift between the two. There might be some need for a cap in there to create a real buck converter. The diode is neccessary, I think, in order to complete the circuit when the FET is open - otherwise no current would flow.

So, some final questions:
1, Does this make any sense? Or am I way off in my thinking here?
2, is the voltage tap and current sense resistor in the correct place in the chain (i.e. before FET) for circuit 3?
3, How do I determine the required smapling rate of the voltage and current sense resistor?
4, When applying the PWM, do I change the duty cycle as well, or just the frequency? Can I change the frequency to figure out power max, and duty cycle to set the output voltage, or does that not make any sense?
5, what kind of inductor value should I go with?



Thanks for reading!
Marcus