PWM help please

[Breezy]

I am a new girl in electronics here and I hope this is the right place. I used to play with my grandfather when he worked on building electronics. I learned a lot about building PCB's, but not to much about theory or how electronics works for that matter. I can usually follow a schematic and get something built. I hope to learn more because this is fun for me, building anyway.

I am in college now and do not take any electronics or related courses, so this is just a hobby for me.

I found a circuit from a guy called "Zero Fossil Fuel" that is a PWM with constant current ....... parts???? And I found other similar circuits out there too. But this one has my attention cause I need a circuit that has constant current, but then I found another web site where a man says the "Zero Fossil Fuel" PWM is good, but has problems. This is what he says about the 1st attached circuit:

The Circuit designed by "Zero Fossil Fuel" is better than normal PWM, however when it goes into the Current Limiting Mode, it operates in a Linear mode that causes the MOSFET's to Over-heat. So you require a VERY LARGE HEATSINK to try to keep it from overheating.

This is where I run into a problem, I have no idea what all that means. And he says his design of the "Zero Fossil Fuel" circuit is modified and better working than the original Zero has. I attached both of these schematics to let you know what I am refering to.

And this guy says he has a PWM circuit that does what the Zero one does, but with no heat or any other problems. But he wont let anyone know what that is and you can buy it but it is covered in resin or epoxy or some sort of thing.

So my main question is can anyone explain what Current Limiting Mode is and what does operate in a Linear mode mean. And if you can answer those questins maybe you can give me an idea how to make the circuit so it does not have those problems.

Thank you all, Brianna Rose

 

[BrownOut]

The designer, "Chemlec" is a member of this forum, and comments here from time to time. Maybe you can PM him and ask him to read and comment on this subject.

 

[Breezy]

Thank you, i will do that.

 

[Dx3]

Current limiting is a way to protect a circuit that is normally dealing with voltage. If the need for current gets too high while using normal voltages, the circuit protects itself by limiting the current at a constant value. That usually causes the voltage to stop being as high as it normally would be.

Linear mode means that the switching action has stopped. The circuit is being asked for more than a switched signal can provide, so it stops switching and just does the best it can with the current available.

 

[BrownOut]

Honestly, I don't see any significant difference between the two designs. For both, when there is a requirement for current limiting, the output of U1D will change from close to the positive voltage to close to ground. When that happens, the output of U1C will remain low, and the MOSFET doesn't turn on, or at least the turn on time will be extreemly short, effectively making a current limit. The second design would seem to sense the over current a little faster, due to the 10u capacitor connected to U1D, as opposed to the 47U cap in the first design.

 

[Ubergeek63]

use an LED driver meant for an external switch (FET) - they are meant to do efficient current sourcing. NEITHER of those are switchers, they are either oscillators or linears.

something like an HV9910: https://www.electro-tech-online.com/custompdfs/2010/07/AN-H50.pdf

Dan

 

[BrownOut]

After looking at the circuit a little more, the "linear" problem could be easily solved by reconfiguring U1D as a comparator with hystersis and providing a fast charge path for C4. If the discharge is timed to approximate a ΔV/ΔT equal to the hysterisis/oscillator frequency, then a constant average current can be achieved without using linear mode.

 

[Ubergeek63]

After looking at the circuit a little more, the "linear" problem could be easily solved by reconfiguring U1D as a comparator with hystersis and providing a fast charge path for C4. If the discharge is timed to approximate a ΔV/ΔT equal to the hysterisis/oscillator frequency, then a constant average current can be achieved without using linear mode.

There are simpler solutions however.

Which still makes it a current limited chopper and not a switcher. as such you will still see the squared term in the losses of what ever (resistance) is limiting the current.

30A with a 1 ohm current limiter is 900W peak in the resistor while you might only be supplying a fraction of that to the load. at 3A average to the load you are looking at 90W lost in the resistor that would most all go away if it is limited by an inductor instead, making it a switching regulator.

Dan

 

[BrownOut]

The resistor limiting the current is only .003 Ohms in the improved circuit. At 3 Amps limit, the loss is just a few mW, not a big deal if we're talking an automotive application. Chopper regulators have a wide application class.

 

[ronv]

pwm

Actually neither one is to good. The problem is mostly around U1C. The big fet wants to be turned on and off quickly so it is not in the linear reigion (half on half off) where power dissipation is very high. To do this U1C needs to switch fast (It can't) and supply high current to charge and discharge the fet capacitance (It can only drive 20 ma). There is also a bit of a problem in current limit because the only thing that makes the circuit kind of stable is the big cap on U1D. Imagine this picture in your mind. Pin 9 goes more negitive that pin 10 turning on the fet. This increases the current causing pin 10 to go more negitive. The result is the circuit is neither on or off. You could make it much better by changing out all the op amps except U1D with comparators. This would speed everything up. Then add a fet driver between U1C and the 100 ohm resistor to drive the fet with 120 ma or so to speed it up. Then to add some stability add some hysteresis to the comparator at U1C to make sure the switching time is not to high.
Tell us what you are trying to run with this. Maybe there are some other ideas.

 

[ljcox]

I agree with ronv.

I suggest you download the data sheets for MOSFET drivers such as the MC34151 & TC4420.

These are fast switching and can drive a significant level of charge into (& out of) the MOSFET gate region thus significantly reducing the turn on/off times.