Designing a 3 mode smart solar battery charger with MPPT

[hwgeek_1221]

I'm trying to design a solar charger with MPPT feature that runs in 3 modes (bulk, absorption and float) depending on the state of the battery. The logic is implemented in a fast micro controller and the dc-dc converter utilised is a buck converter. After doing my homework (lots of readings and research) I still have a couple of questions that I find confusing.

1. Correct me if I'm wrong on this. Do we only use MPPT during the bulk mode? If yes, we use it only if the charging current generated is less than the current threshold?
3. How, and by what means can I calculate the duty ratio of my transistor during the absorption and float mode in a manner to achieve a constant charging voltage with a maximum threshold charging current? Same thing also for the constant current charging.
5. While digging answers for question #1 I found that TI manufactures solar battery charge controller IC (**broken link removed**) that has the capability of regulating the charge voltage and the current voltage with high accuracies. They mention that they use internal loop compensators to do this. Digging in the high level schematic they have I found that the reference voltage feedback VFB and the current feedback SRN & SRP (SR is a current sensing resistor) are inputted to a compensation network. Does anyone know how this compensator network was designed? (see picture below) I would like to implement something similar in software (I'm not asking how to implement it in software).
6. When changing the duty ratio to achieve regulation, the charging voltage and current can't be predicted because changing the duty would take us to another point on the I-V of the PV which is not predictable. Or is it?

​

 

[Warpspeed]

MPPT is only important where the load is greater than the available power from the panels.
There may be 2Kw of panels installed, but on a grey foggy rainy day, there may only be 30 watts available, and the charger has to put up with that, because that is all there is !

Now I have been looking at this whole MPPT thing for quite some time, and I have made some interesting observations.
At least with the 24v monocrystaline panels that I am using, peak power ALWAYS seems to occur at around exactly 30.0 panel volts.

I have one connected to a power meter which drives a PWM buck regulator feeding a big load resistor, and the PWM is arranged with closed loop feedback to hold its INPUT at 30.0 volts.

I have crossed over the inverting and non inverting inputs to the error amp, so that if the input voltage tries to rise, the PWM duty cycle and the load increases, thus regulating the input voltage, not the output voltage of the regulator.

Amazingly this works very well at any time of day, I can rock the 30.0v up and down, but power always peaks at 30.0v and its not a sharp peak either, more of a cusp, and over a couple of volts the power hardly changes.

At sunrise I watch the voltage climb from zero up to 30v only then the PWM starts up, and I start seeing 1 watt, 5 watts, 30 watts etc as the sun rises.

The system I am using is a simple hardware system, but the main PWM loop is controlled by the dc output voltage of the regulator.

The MPPT part senses the solar panel voltage, and feeds that into what is normally the soft start pin on the PWM chip.

This is a maximum duty cycle limiter, and it overrides the main output voltage control loop such that the solar panels can never be overloaded.
It rapidly pulls back the duty cycle at exactly 30v input and that is where it sits if the system is fully loaded or overloaded.

Even with a dead short across my output load resistor, the solar panel voltage sits at 30v.

It works just like the current limit on a bench power supply, the output voltage is regulated, up to the overload point where the supply then goes into current limit and the output voltage collapses. But this MPPT current limit is tied to the solar panel voltage.

It works like a charm. And its very simple.

In either hardware or software build your standard multi state battery charger state machine, and arrange it so that the output is reduced whenever the solar panel voltage is pulled down to the max power voltage, which you will have to find for your particular panels by some experimental testing.

Temperature in theory should also probably be taken into account, but in practice I have not seen any huge difference between the winter and summer peak power voltage. The difference, (if it is there) is well within the "cusp" of max power and is not noticeable in practice.

 

[srizbf]

MPPT is only important where the load is greater than the available power from the panels. ...

....It works just like the current limit on a bench power supply, the output voltage is regulated, up to the overload point where the supply then goes into current limit and the output voltage collapses. But this MPPT current limit is tied to the solar panel voltage.

It works like a charm. And its very simple.

very crisp explanation of MPPT operation for users from Warpspeed

 

[hwgeek_1221]

Sorry, I didn't really get all the answers I was looking for. How are we able to get constant current and constant voltage charging?

 

[Warpspeed]

You cannot always get constant current charging from a solar system if the sun is not shining.

And you cannot always expect to see constant voltage charging when the battery is pretty much fully discharged.

In either hardware or software build your standard multi state battery charger state machine, and arrange it so that the output is reduced whenever the solar panel voltage is pulled down to the max power voltage, which you will have to find for your particular panels by some experimental testing.

 

[treez]

what you should do, for dull days, is have a big capacitor at the output of the panel, and let this slowly charge up as current dribbles out of the panel, then when the cap voltage gets so high , put it through the switching converter and into the storage battery.

When its shiningg bright, I belive as discussed above, that you need to regulate either the input current to the smps, to the maximum power current of the panel, or regulate the input voltage of the smps....it sounds weird, because we are all used to regulating output voltage of an smps, but you can of course close the loop round input current , or input voltage, as long as you ensure you have overvoltage protection at the output