output current problems with buck converter

[Nicx]

hi i'm supposed to be constructing a buck converter which will provide around 160 W to 200 W. when i test it on an e-load at CV mode, the output current is negligible (0.2 A) compared with the expected value. i designed my circuit using a UC3843 PWM controller. can anyone help me? i dont know where in my circuit to look anymore. thanks.

 

[jmpac]

Have you put the output capacitors which ensure regulation loop stable?

 

[jgk2004]

Also what is your duty cycle like at your switch pin, before the inductor. With increasing your load you should see your top switch (PMOS) time increase.(time on) Is it doing this? Also like Jmpac stated is it stable. You can see if its stable by looking at the duty cycle as well. It should be constant for constant load.

Jgk

 

[Nicx]

thanks for replying!

Jmpac, i selected a large capacitor value way above the computed minimum, and paralleled 2 caps for lower esr. errr...how do i select the cap to ensure stability? thanks!

jgk2004, that's another problem. i programmed the pwm controller to have a maximum duty cycle of around 96%, so i followed the component selection indicated in the data sheet. the frequency at the oscillator pin is correct, but the duty cycle at the output pin of the UC3843 is the opposite of 96%. so im getting very thin pulses and long pauses and additionally, the frequency is halved of the one at the oscillator pin. i tried changing the components for a lower duty cycle but its still the same thing. ive also tried replacing the pwm controller but it doesnt seem to be the problem.

 

[jgk2004]

Ok so let me get this straight. Are you setting your PWM controller to a SET duty cycle? Do you not have any feedback to sense the load changing? thus adjusting the duty cycle?


If you have a feedback for your duty cycle control, I would say if you can sweep your load resistor you should see the duty duty cycle change with a correlation to your resistor. So if you increase your resistor you would see your duty cycle decrease which would mean your Pmos top switch would be on less of the time. If this happens you have the right feedback. If its always the same duty cycle maybe you have positive feedback... which would mean your output voltage would also not be correct.
If you gave more information i could maybe help little more. What is your schematic?
What is your input voltage? what is your output voltage. what is your current out needed range. Are you using voltage mode feedback or a current mode sensing?

Jgk

 

[Nicx]

i temporarily grounded the current sense pin for simplicity since ill be testing it for a specific voltage first. here are my specs:

Vin ranges from 28 to 60 V
Vout is 24 V
Iout ranges from 6 to 10 A (or even 8 A is good)

 

[jgk2004]

One more question. If you sweep your input voltage from 28 to 60 volts does your output stay at 24V? I do not think its a good idea to temporarily ground your current sense pin. I think this is your problem why the output current is so small..... Also did you try sweeping the load resistor and looking at the duty cycle?

let me know the results.

Jgk

 

[Nicx]

Jgk. thanks for your suggestions. in my previous circuit, the voltage was regulated at around 23.7 V. then i was able to find the problem with my pwm signal. i adjusted the compensation part of the chip and i was able to get the maximum duty cycle waveform.
so the input to my fet gate is fine, however, the voltage is not regulated anymore.
on constant voltage testing, my output current is significantly higher but still not high enough. can you suggest a way i can go about the feedback (for regulation)? thanks!

 

[jgk2004]

Well, you would need to have a circuit which adjusts your PWM duty cycle depending on current load, which would of course cause your voltage to drop. You can do it with current sensing (a current mode) which would be monitoring what current is being charged up and feed to your output from your inductor, or voltage sensing (voltage Mode) which would be sensing your output drop and compensating for it. It really depends on how you want to get this information back to your PWM and what is supports. Hope this helps

Jgk