Hello!

I have built a charging circuit for a battery. I used a P-channel Mosfet for a hi side switching needed for the charging. However, I was only able to find a P-Channel Mosfet with full conduction at 10V and my input is only 5V (from uC). Thus, the charging current drew from the power supply is very small.

Finding a logic level P-Channel Mosfet seems hard that's why I tried to use a PNP BJT (TIP32). Turns out that it indeed drew a lot of current from the power supply which is limited to an output of 3A only. However, it turns out that the TIP32 and TIP31(hotter) gets hot fast and while monitoring the battery charge, it seems that it even discharge a little.

I don't know what is the problem but I'm guessing it is due to the resistors, maybe I'm missing some at some point but I don't know where.

here is the schematic:

C1 - simulation only for the charging of the battery (the battery usually have an initial voltage of ~3.5V)
Battery - the supply for charging the battery

input from R9 is from a uC

 

Here is how I would do it:

Note the Vce(sat) of the PNP and the power dissipation in it (blue trace). Driven from a 3.3V or 5V Port.

btw: Throw away that toy and get a real simulator

 

Your circuit drew high current because there was no resistor to limit the base current from the PNP through the NPN when ON (which is not needed for a MOSFET). MikeML's circuit shows the added resistor (R1). He added R2 to shunt away any leakage current and insure full turn-off of the PNP when the NPN is off.

 

You should using 2N2955 to replacing the Tip32, and plus heatsink at least 50mm*50mm.
The rating current of transistor at least should be using 5 times of the output current.
If your current not too high then you should adding a resistor to limited the current.

Or using the circuit as attached, you can also using IRF540,IRF640,etc...

If you using IRF1010E then you just need a small heatsink or no.

 

Thanks for the quick reply!

Looks like that the only critical difference between our circuit is the 47Ω? Is that a result of your computation? and also with the presence of the 47Ω, the input to TIP42 couldn't be completely pulled down because of the voltage drop at the 47Ω resistor?

I was previously using Multisim for simulation but changed to ISIS since I thought it's better. What then is your simulator?

 

...
Or using the circuit as attached, you can also using IRF540,IRF640,etc..

The OP only has 5V to turn on the FET, and he does not have a "logic-level" FET.

 

I thought my R11 is similar with Mike's R1? Although I didn't think that it will limit the current from PNP's base to NPN and was just there as a pull up resistor. I also don't get how it can limit the PNP to NPN current.

 

The OP only has 5V to turn on the FET, and he does not have a "logic-level" FET.

I just suggested that if using my circuit then it will be no need to use a logic level mosfet.

 

I just suggested that if using my circuit then it will be no need to use a logic level mosfet.

But he doesn't have 12V.

 

But he doesn't have 12V.

Yes, you are right.
When the requirement is not clear, sometimes we only can suggest or give some ideas, unless the contents are very clear to settle down each value.

 

I thought my R11 is similar with Mike's R1? Although I didn't think that it will limit the current from PNP's base to NPN and was just there as a pull up resistor. I also don't get how it can limit the PNP to NPN current.

As you note, R11 is simply a pull-up resistor. Your circuit has nothing to limit the current when the NPN is ON. Mike's resistor is in series with the PNP base to limit the current.

 

Sorry for not understanding it again. Here's what I did to limit the current on NPN.

I can't connect my R11 to the charging supply as what Mike's circuit is since it originally is only at 4.2V.
The R2 indeed limit the current because now both of the transistor is not heating up. However, the charging is really slow. It's about 5-10 times slower than directly connecting the power supply to the battery. I don't know if obtaining the same charging speed is unachievable.