Now this is just for the fun of learning to problem solve using discrete components.

But according to what sgt.wookie was sharing, I need to charge and discharge the gate capacitance quickly, to avoid keeping the mosfet in it;s linear region.


The only reason I used large resistances is only to keep the battery current drain as low as possible. during motor run time.

So attempt #1.

I will try to design a quick charge and discharge circuit for the gate capacitance, that will shut down after a set time to switch over to a gate bias voltage with the original high value resistors, during motor run time.

Option 1: an external capacitor to give a jump start, through a lower bias resistors.

Option2: a monostable (oneshot) circuit to introduce low bias resistors at the gate during switching of the mosfets from conduction to non conduction.

Option#3:yet to be defined.

So as to not run the risk of hijacking this thread I'll post my results on a new thread.

 

Well like you all said, this motor cannot work for the previous schematic diagram.

But after adding the resistor into the correct way, the DC motor perform well.
It can rotate the direction what i expect.

But problem still is the heat.
Anyone can provide some formula to calculate the current (Ib,ic,ie) for PNP 2907n which is suitable to saturate/cutoff and can endurance the heat for long period?

The schematic below show my circuit and it's works fine for 10 minutes like that, more over the (PNP)transistor is 2 hot.

 

The general circuit looks about right.
The big question is the motor current?

If it's only about 100mA, you should not have any real problems.

By the time you get to 500mA, each transistor is dropping about a volt even with enough drive to saturate.

For both types of transistor, the data uses 50mA base drive for a 500mA load even to get that saturation level; if you are operating at anything like that current, R6 & R7 should be reduced to 100 Ohms to get sufficient base current on the upper transistors.

If they still get too hot, you may need to change to higher rated transistors (eg. TO220 case type power devices) . These should have a lower saturation voltage at the current you are using, so produce less heat than the small transistors.

 

Along with what RJenkins said comes another consideration.

If you are driving the bridge from a uC like a PIC, the 100 Ohm resistors you are using for R1 and R3 will result in excessive source current from the uC's output, which may overheat/damage it. To calculate the minimum base resistor value for a uC, you can use a rough approximation of:
Rbase >= (Vcc - Vbe) / uC_max_source_current
Assume 0.7v for Vbe for starters. Vcc is what you're using for the uC supply.

So, Rbase >= (5v - 0.7) / 20mA = 4.3v/0.02A = 215 Ohms. 220 Ohms is the closest standard value.

This also affects the collector current, which will be roughly 10x to 15x the base current.

You might instead investigate using logic-level MOSFETs