buck converter current problems

[Nicx]

hi, ive been trying to construct a buck converter with output 24V and 10A max. but i would settle for about 150Watts output. my input specs are 30V minimum and 60V max.

when i tried testing it at constant voltage test of an e-load, the current output registered on the e-load is almost the same as the current drawn from the supply. i'm currently testing it with a desktop DC power supply with output 30V-3A. and since my circuit is acting this way, my max power is about 72Watts. i was advised to place two of these DC power supplies in series so i would get the 60V max input and 180Watts of total power input. however when i hooked it up to my circuit, and ran a constant voltage test, my circuit got fried. specifically the pwm controller's Vcc resistor.
can someone suggest how i should debug this? how do i fix the load=supply current problem? my calculations indicate an inductance of 18uH and ive increased mine to 39uH to avoid saturation. im using an iron powder toroid core for the inductor (material -26). ive attached a schematic. i need help badly please. thanks!

 

[ghydda]

Hi,

I'm not sure I understand your schematic.
I cant find the inductor anywhere, and where I expect it should be there's something else.

Please clean up the schematic and explain what those black box components are.

Best regards.

 

[Nicx]

Hi,

I'm not sure I understand your schematic.
I cant find the inductor anywhere, and where I expect it should be there's something else.

Please clean up the schematic and explain what those black box components are.

Best regards.

oops,sorry about that.

 

[ghydda]

Hi Nicx,

The Vcc pin of UC3843 have an abs.max. rating of 30 volt.
So first of all you will need to deal with that before any power conversion can begin.
I suggest running the regulator-chip at a more comfortable 14-18 volts. Let me know if you need a small schematic on how that can be achieved.

Secondly I see no reliable way of getting a pulse transformer to go much above 50% duty cycle. And you need 24V/30V = 80% (assuming lossfree components, lossy components will drive that number up even further).
You could replace the pulse transformer with this: www.irf.com/product-info/datasheets/data/ir2117.pdf
Any other similar high side driver will most likely work just fine. I just happen to like that one and it is fairly cheap.

Thirdly, I believe the UC3843 needs the Isense pin terminated either to an actual current sensing shunt (thus getting current control mode) or to the RT/CT pin (thus getting voltage control mode). But this is just how I remember it, you (or me or someone) needs to check this.

Let us know if this improves things. I suspect you will need a few more tweaks before the converter is running smoothly.

Happy hunting!
/Ghydda

 

[Nicx]

Ghydda, thanks!

i've already biased the Vcc pin using a zener diode, keeping my limiting resistor between source and Vcc. is that enough?
i've gotten the expected waveforms all the way to the switching diode, around 94% duty cycle.
however when i attach the inductor and output capacitor, the PWM signals drop to periodic spikes instead of full pulses. it looks like an opposite duty cycle -- instead of 94% its 6%...
i'll try to fix the Isense pin like you said.
do you have any tips on what i should do about my inductor? its currently measuring 39uH with a powdered iron core. i used an awg-16 (for max of 5A -- 5A output is good enough, i dont really need to reach 10A anymore. )
thanks again!

 

[ghydda]

i've already biased the Vcc pin using a zener diode, keeping my limiting resistor between source and Vcc. is that enough?

Well nearly, the small decoupling capacitor in the circuit which didn't feature the zener is still very much required as your shunt regulator for the driver supply is a very high impedance type and during mosfet turn on you will need a large, albeit short, current spike to swiftly turn on the mosfet. Usually 100nF (as you've used) is addequite, but it all depends on the charge needed to turn on the transistor. Increase it if the supply dips more than a volt during turn on.

i've gotten the expected waveforms all the way to the switching diode, around 94% duty cycle.
however when i attach the inductor and output capacitor, the PWM signals drop to periodic spikes instead of full pulses. it looks like an opposite duty cycle -- instead of 94% its 6%...

This sounds ok. Without the inductor no energy can flow to the output and hence to the feedback sense pin thus the switching regulator cranks up the dutycycle.
When you insert the inductor the feedback recieves energy and the switching regulator adjusts the duty cycle down until equilibrium is reached - in this particular case it's a very low 6% - I suspect you are drawing very little current from the output. This means only very small pulses are needed to maintain regulation.

i'll try to fix the Isense pin like you said.

do you have any tips on what i should do about my inductor? its currently measuring 39uH with a powdered iron core. i used an awg-16 (for max of 5A -- 5A output is good enough, i dont really need to reach 10A anymore. )
thanks again!

It doesn't strike me as bad place to start. Keep it as is until things suggest you should change it (ie. if it gets very warm when running full power).

Cheers,
/Ghydda