Building battery charger, can't figure out where pwm fet should be??

Discussion in 'General Electronics Chat' started by rudyauction8, Mar 13, 2015.

Jan 27, 2012
252
2
I've been trying to figure this out for hours. I attached 2 pictures of 2 different circuits showing why I can't use either one. Any advice?

File size:
19.8 KB
Views:
48
• 2.png
File size:
7.1 KB
Views:
45

Jan 27, 2012
252
2
I forgot to mention the battery voltages are between 7 and 16 volts.

3. #12 Expert

Nov 30, 2010
16,298
6,811
If you want to use an N channel mosfet, you do it like this:

• N-ch PWMing.png
File size:
1.4 KB
Views:
39

Jan 27, 2012
252
2
I can't measure the battery voltage without turning the mosfet all the way on to get a ground reference, which if the lithium battery is full or too small to handle the current from the 20 volt power supply could be dangerous. Any ideas here? I'm using a pic12f675 and using the adc to measure the voltage.

5. #12 Expert

Nov 30, 2010
16,298
6,811
You are just gong to have to do some level shifting with normal transistors to get a 5 volt signal to control a 20 volt supply.

File size:
2.1 KB
Views:
30
6. BobTPH Active Member

Jun 5, 2013
782
114
Why are you using PWM to charge a battery?

bob

Jan 27, 2012
252
2
Thanks. I'll start working on it today. However doesn't the fet require 6-7 volts between gate and source? If so my power supply needs to be that much higher than the highest battery voltage?

Because pumping 20 volts at 15 amps into a 300mah 3s lipo will cause it to explode and I need the 15 amps for my larger packs. PWM will be used to limit current and manage the constant voltage portion of the charge.

8. BobTPH Active Member

Jun 5, 2013
782
114
PWM will not limit the current to the batteries. That is why I asked. It will limit the average current, but the peak current will be the same as if there was no PWM, and you will probably start a fire. Batteries must be charged by a constant current.

Edit: PWM is usable in the context of a buck converter that controls a constant current source.

Bob

Jan 27, 2012
252
2
output filter capacitors will smooth out the pwm. using twin 470uf and single 0.1uf tantalum with pwm freq of 20khz.

10. BobTPH Active Member

Jun 5, 2013
782
114

Bob

Jan 27, 2012
252
2
OK wow. Not helpful. I have several commercial chargers but they don't have the functionality and power I need, and I am more than capable of keeping my charger projects safe.

Jan 27, 2012
252
2
You ever heard of switching power supplies? They use pwm and filter caps and their output is very smooth, the voltage spikes that make it past the caps are not powerful enough to harm batteries or the circuits they drive.

Just saw your edit. Thats exactly how I will use pwm, a buck converter that limits current into the battery, the amount of current let through based on the batteries size and charge level.

13. Papabravo Expert

Feb 24, 2006
10,145
1,791
If you came up with those circuits yourself we have every reason to be concerned. If you got them from somewhere else you need to find better sources of circuit ideas. People can and do get hurt doing silly things with voltage and current and chemistry. There is even the posthumous Darwin Award for scientist/engineer/inventors who like to tinker with stuff. BTW if we manage to save your life then I would judge that to be very helpful, even if a bit harsh.

http://en.wikipedia.org/wiki/Darwin_Awards

14. #12 Expert

Nov 30, 2010
16,298
6,811
Only if you insist on using an N channel mosfet...which again brings up the question; If you can't tell the difference between an N-channel and a P-channel transistor, why do you think you can do this safely?

Jan 27, 2012
252
2

Those circuits are not my actual design. I made them in less than 5 minutes to visually explain the problem I'm trying to figure out. Think of them as a very basic idea, with only the actual components I'm thinking about and none of the rest of the circuit.

Jan 27, 2012
252
2
Any cons of just using a pnp power transistor? Besides the efficiency?

17. Papabravo Expert

Feb 24, 2006
10,145
1,791
So it might be more efficient to research actual circuit designs rather than trying all possible combinations of hooking up components in the vain hope that you will stumble on one that works. You might want to start with a basic principle of N-channel FET operation that Vgs must be greater than Vgsth, the gate threshold voltage, in order to turn the device on. That one fact will prevent you from doing some things that have no chance of working.

18. crutschow Expert

Mar 14, 2008
13,016
3,235
That works only if you have a series inductor to limit the peak current as is used in a buck regulator.

19. #12 Expert

Nov 30, 2010
16,298
6,811
A bi-polar transistor would get pretty hot at 15 amps.