Im working on a project, Im reading the voltage and current values off a solar panel, using, an AVR microcontroller .. Im also charging a 12V 4.5AH dry battery .. (pls see figure)

When the battery voltage goes to 13.2 V , I want to be able to disconnect the battery, so it doesn't overcharge .. I think i need a high side p-channel Mosfet switch .. using the IR2118s mosfet driver with it..

Please advise is this approach correct .. and is there a better solution ... can I do it without the driver , and which MOSFET is most suitable for the job ..

P-channel mosfets have high on-state R ... so that might be a problem .. so thinking about N-channel..?

what do ye say ..

 

You could use your processor to make a charge pump if your have the available I/O pins. A charge pump will double or triple depending on stages a voltage (minus losses). This could be use to generate a adequate gate drive for an N-Channel FET.
You could also use your processor ADC or if it has an on board comparator to control the switch by sensing the Voltage level.

By the way the IR driver you mention is a high side fet driver it uses a bootstrap capacitor and level shifting to generate Vgs for an N-Channel Fet.

A P-channel would be the easiest if you want the simplest solution.

EDIT

One thing I forgot to add is if you use the IR driver you will need a chargepump anyways if you want the switch to stay closed for extended periods of time you can refer to an978 pg18.Instead of a timer just use your processor for the chargepump.
http://www.irf.com/technical-info/appnotes/an-978.pdf

 

Why does the switch have to be high-side? If it were low-side, you could run a logic-level, N-channel directly off the MCU. Using low-side mosfets makes life easier.

John

 

Why don't you use just a simple relay?

 

The switch has to be high side, because I want to be able to have two charging states, apply a charge, or use PWM for float charge...

Please explain how the voltage tripler works .. and As I would want the switch to be closed for extended periods of time..

and whats the science behind P-channel being easier and simpler ...

 

For P- channel you wouldn’t need a charge pump. To enhance a P- channel you pull the gate below the source so its simple to do for a high side switch. If speed isn’t an issue which in your case it isn’t you could use a simple BJT off your processor to control your P-channel fet.

This is a simple way to drive a P-channel FET.

When the BJT is on (saturated) the Fet is conducting when the bjt is off the fet is blocking. I drew in the “body diode” so you are aware of it. This is a parasitic diode which happens from the process of constructing the mosfet it is internal to the mosfet. So you have to be careful that the drain is more negative then the source or the body diode conducts. When you’re looking for your fet you want to insure it can handle the maximum voltage plus margin, can handle the current and can dissipate the power and you want as low a Qt as possible. The last one helps to reduce the requirements on your driver.

A mosfet is typically described as a voltage controlled device, which is true but it does have an effective input capacitance which the driving source sees. This capacitance is what needs to be charged up to switch OFF a P-channel fet or ON a N-channel fet. One usually uses Qt which is specified in the data sheet to estimate switching times based on peak driver current and drain voltage. In your case you aren’t switching the fet repeatedly your using it as a switch (ssr) so you would have to be sure if you use a weak driving source which results in slow transitions your switch stays within it’s SOA. The SOA graph is in the data sheet speced for 25 degrees.

edit ; For completeness
Obviously if the effective capacitance requires charging to perform a state change on your Fet it needs to be discharged to invert that state. This is why you will see sink/ source used in mosfet driver data sheets.


If you were to use an N-channel fet you need to pull the gate above the source in the case of a logic level FET by 5V for full enhancement. Logic level fets are abundant at your voltage levels.

This is an example of a charge pump doubler I have running off a PIC I built it for a HS driver. The PIC is outputting about 10kHz 50% duty for the charge pump.

The resistor R5 represents the load in my case the driver current. The pulse source V1 is your processors output. PIC’S can only run off 5V so I’m using the extra BJTS to start at an initial 12 volts and double it rather then triple it off the PIC. This has been modified from the original application for a quick example for you so if someone sees errors chime in. You could use shottkys for smaller losses.

The output voltage almost doubled.

 

Here is a discrete High side driver. I didn’t waste time calculating resistor values or selecting the best FET but it should give you an idea on how you could do what you want. It should give the ability to pulse the fet,or keep it open or closed for extended periods. The pulsed sources are your processors outputs.

Here is the output
Top trace is the fet gate source voltage. Bottom trace is the drain current. Pulsed at about 10kHz 50% duty.

You could also use the charge pump to supply Vcc for a fet driver like a low side fet driver rated for at least 25 volts Vcc.

But the above should work. You have to go through it and tune resistor and capacitors so it’s optimized. You want it to pump up the voltage as quickly as possible, so the fet isn’t stressed. You could in software provide sufficient delay for the Caps to charge up before you start switching the Fet for example. Also don’t use the IRF741. I just used it for the sim you should use a logic level FET for at least 30Vds to 60 Vds area with about 10A or better continuous rating, with minimum gate charge (Qt). Its a bit of a balanceing act you dont want to dissapate to much power in your charge pump and drive circuit. But you want quick transtions on your fet.

 

thanks a lot for going into all that detail ..

 

thanks a lot for going into all that detail

I just finished doing something similar. Only for 200VDc using a PIC and a high side driver.

 

making an Inverter or Buck / boost stage ?

 

I have also posted something regarding my solar tracker project in the alternative energy section. Please have a look and answer some questions there...

 

http://www.electro-tech-online.com/alternative-energy/89695-very-very-good-single-axis-tracker.html


Here is the link , sorry its on another forum. i hope posting a link to another forum is not forbidden. Pls remove the link if its illegal/unethical etc etc

 

I want to sk AchMED for more details about this circuit

Is there any equations for calculating the values of resistors and capacitors ?!?!
Is there any limitations for the switching frequency ?

 

I want to ask AchMED if you please send your schematic for 200 VDC ?!?!