Solar Charge Controller Circuit

[sonar_abhi]

Hello All,

I have been designing a solar charge controller for 12 volt battery. The circuit is shown below. The parameters of the buck converter are as follows:

D1: Schotkey diode MBR1645
L: 8uH
Ci: 470uF
Co: 2000uF
Can anybody please validate the above design. The parameters are as follows:

Vin(min): 15V
Vout: 12 V
Iout: 15 A
Switching Frequency: 50kHz
Iripple: 30%
Vripple: 10%

The 1kHz PWM wave feeds the Dickson Charge Pump for driving the gate of the IRF44Z.

 

[ericgibbs]

Hi sonar,
It works reasonably well in LTSpice simulation.
I have simmed the gate charge pump as +20v.
The Solar panel sweeps from 15v thru 17v.

Connect the solar 0v to the switcher 0v.!

E

 

[MikeMl]

Begs the question: Why do you need a buck converter, when a simple on-off switch would do?

 

[sonar_abhi]

Thanks Ericgibbs, I tried to run the simulation in proteus, but that was showing a false value, took too long to stabilise and also had to put 8mH instead of actual 8uH for the simulation to work (hence can't really call it a simulation )

Can you suggest anything to reduce the output ripple?

@mikeMI,

I am controlling the PWM with a micro controller that tracks the voltage of the battery via a voltage divider (not shown in the circuit diagram)

Also, the voltage and and current of the panel is monitored via a voltage divider and a current sensor (also not shown in the circuit diagram)

When the voltage of the battery is below 14 V, the PWM is at maximum duty cycle which charges the battery at Vmp of the panel. When the battery voltage rises upto 14 V, the duty cycle is reduced so as to keep the battery at trickle charge.

At the same time, a P&O algorithm monitors the panel power and if the panel is being loaded, it adjusts the duty cycle to keep the panel at Vmp.

A simple on/off switch wont perform this.

 

[ericgibbs]

hi sonar,
Would you please confirm your switching frequency.?
What Output voltage and current do you calculate at a 50/50 duty cycle.?

E

 

[sonar_abhi]

The switching frequency is 50kHz.

Duty cycle I have calculated using D.C = Vout X Converter Efficiency (Assumed to be 90%)
-----------------------------------------------------------------
Vin
For a Vin (minimum) of 15V and the Vout as 12 V, the D.C comes to around 72%

Below this value of Duty Cycle, the current flowing through the system is essentially 0, hence didn't calculate for 50% duty cycle

 

[MikeMl]

....@mikeMI,
...When the voltage of the battery is below 14 V, the PWM is at maximum duty cycle which charges the battery at Vmp of the panel. When the battery voltage rises upto 14 V, the duty cycle is reduced so as to keep the battery at trickle charge....
A simple on/off switch wont perform this.

Using MPPT tracking for charging lead-acid batteries is a total waste of time.
If you simply connect a ~20Voc solar panel to a 14V LA battery, the panel is so close to its MP voltage, that the switching losses in a MPPT tracker totally cancel out the few percent gain that you could potentially achieve by operating the panel at a volt or two higher (nearer its actual MP voltage...

Charging a battery implies that at some point in the solar cycle, the battery achieves full charge, and the available panel power after that point is lost anyway..., so why worry about saving a couple of % during the actual charging time? The whole MPPT thing only makes sense if you are generating AC power and putting it back on the grid (i.e., you have a load which absorbs all of the available power, which a battery does not).

I just use a simple on-off switch (PFET) to connect-disconnect the panel from the battery. A simple Schmitt-trigger voltage sensor controls the gate of the PFET. Algorithm is for a SLA is dirt simple: Connect the panel to the battery when the battery voltage drops below 13.2V; disconnect the panel from the battery when the voltage reaches 14.7V.

In my case, there is an intermittent load (green house thermostatically-controlled vent open/close actuator, water misting pump) that occasionally draws down the battery voltage, and starts a new charge cycle. If none of the loads turn on, and the battery just sits full, it takes several hours for its terminal voltage to drift down below the cut-in voltage, at which point the panel is reconnected to the battery until it reaches the cut-out voltage. That would happen about once per solar day, and it only take a couple of minutes, assuming none of the loads come on. If they do, they trigger a recharge cycle that takes longer...

 

[sonar_abhi]

I agree with what you say Mike, but the thing about charging batteries is that if you actually just bang turn on/off the charging after it reaches a higher cutoff voltage, say 14.7 V, without providing the equalization/float/trickle charge, you are just charging the batteries to their 80% capacity. At the same time the battery life is also reduced.

I totally agree that in most of the cases MPPT is a total waste of time and energy but a charging algorithm must include either a PWM or MPPT sequence so as to provide a trickle charge at constant voltage of ~13.5 V. Without that battery life is shortened.

In my case, I am using a MPPT because I will be coupling the charge controller to an inverter which will power daytime loads so the incoming power is always utilised

 

[MikeMl]

... but a charging algorithm must include either a PWM or MPPT sequence so as to provide a trickle charge at constant voltage of ~13.5 V. Without that battery life is shortened.
...

By the time the battery the battery needs to be floated, it is drawing so little current that MPPT makes no sense there, either. Just use a linear regulator to make the float voltage. The series pass element in the regulator dissipates almost no power. Initially it is "shorted", to connect the panel to the battery with no heat produced. After the switch-over to float mode, the battery current is so small that heatsinking requirements are minimal...