Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Buck regulator

Status
Not open for further replies.

123duermed

New Member
Hello

I'm making a buck regulator, but I'm having some issues. The regulator is supposed to regulate 50-300V down to 14V. But now when I'm testing it, I gets some problems. First of all the duty cycle gets low very fast. It goes from about 60% to about 5% when the input voltage goes from 22-30V. When the duty cycle get to low, the regulator stops. It seems that the mosfet does not get enough time to switch on and off, when the duty cylce get very low. I'm running the duty cycle at 40KHz, and tried down to 10 kHz, but I'm still having the same problem.

Also the source voltage when the mosfet is off, gets a lot of ringing. Why? Do I need to change something in my buck circuit to improve this? See the picture. I'm also attaching the circuit

20130424_184350.jpg
SPENREG6.png
 
What is the load on the output?

What does the gate drive voltage look like?
 
The load is an motor and a power resistance. Is goes about 1-2 A depending on the resistance (variable). The gate-source voltage on the mosfet looks like the pwm signal, only with the miller effect. but the output gets the ringing vs GND as well.
 
You said 300v to 14v (worst case) which at 80% efficiency is a duty cycle of about 14 / 300 / 0.80 = 5.8%. That is pretty extreme and you will probably have a big issue with inductor saturation in a pure buck design. You will need a really high value inductor.

Normally they use a SMPS with a dual coil inductor (transformer) like a forward converter, that helps with the voltage conversion and means a lower inductance primary can be used.
 
Right now the biggest problem is that the duty cycle is much lower than 14/Vinn. With 25V the duty cycle should be (theoretical) 0.56, but the duty cycle is about 5% or less.
 
he biggest problem is that the duty cycle is bout 5% or less.
You need to step through the PWM IC one stage at a time. (trouble shoot)
What is the voltage at the input of the error amplifier?
What is the Vref at the other input of the error amp?
What is the output of the error amp?
Compare the error amp output to the ramp on the OSC.
What is the duty cycle at the output of the PWM compared to the gate of the MOSFET.
 
try a synchronous design and you'll find the duty cycle remain at nearly the theoretical,

when the current though the inductor is discontinuous, you'll find the output voltage climbs rather quickly, and approaches the input voltage.

https://en.wikipedia.org/wiki/Buck_converter about half way down the page you'll see a graph that shows you the normallized output voltage when operating in discontinuous mode.
 
Discontinuous.......only 10uH, light load.
That's it. Try 100uH or 1mH.
It is ok to be in discontinuous mode. Probably not good to switch between the two modes at 1/3 power level.

What is happening.....The load is light. In a very short period of time enough power is stored on the inductor to hold the load up. In a second very short time the power from the L is moved to the C. The rest of the time "nothing happens". The inductor rings.
 
Hello again.
So I have tried to increase the freqency to 300 kHZ and connected three inductors (10uH) in series to get 30uH. I'm still having some problem getting it to work. When I get over 20V in the hole thing shuts down. Everything stops working. When the input voltage is under 20V I get this result:
20130502_192440.jpg
Blue:Gate of mosfet
Green: Source of mosfet
Yellow: PWM-signal.

But the most confusing is the shutdown after I reach 20V. It looks like the duty cycle don't change. From 15V to 20V the duty cycle should change form about 0,9 to 0,7 right?
 
I can not see clear. Is this the part you are using? uc3823

What is D6 at IR2117? The diode must be fast!
What is C12 470uF? I use ceramic 1uF or 0.1uF.
What is D2 Zener diode? What is supply voltage? 12V?

I see two places on scope where the drive signal is on but the Gate-Source voltage is zero. This is why I am asking about IR2117.
 

Attachments

  • uc3823.pdf
    2.9 MB · Views: 189
Last edited:
The D6 is a STTH100W06C. It's ultrafast.
C12 is a film caps.
D2 is a Zener diode with break voltage of 20V, so the the Vcc is about 18V.
It is a UC3823A.
 
I think your gate drive is the problem because I can see that the gate-source voltage some times is 0 but the output of the PWM is high. It acts like the IR2117 does not have a good supply voltage. The gate-source voltage should have sharp edges but it seems slow.
Please resend the schematic with part values. Also what you sent is not clear. I could not read the IC part number. uc3823???
 
spenreg7.png

I have also tried with an IR2110. I connected Vcc on the UC3823A and IR2110 to a separate power dc supply with 12V. Then I started to increase the input voltage. When i reached about 12V the pwm signal becomes really weird. When this happens the voltage of pin 1 on the UC3823A is about 5,1V which is the same as Vref (which is connected to pin2) So it seems that when I reach this voltage, the pwm signal does not get smaller, but start to act weird.

10V:
20130503_221356.jpg

12V:
20130503_221407.jpg
 
Screenshot-10.pngi am guessing due to lack of a photo of your circuit, that you do not have a low enough inductance path between your 220uf and your 2.2 uf supply capacitors, the mosfet, diode D1, and your gate drivers, which is made up of the supply capacitors, and diode D6 and the ground line between pin 6 of the ir2117 and the source of the mosfet.

this is why the weirdness shows up at around the same voltage that you're supplying the gate driver with (12 volts)

i recommend reducing the frequency to 30Khz, getting everything to work and then increasing the frequency when you have everything figured out.

in the attached photo, is where i think your mosfet is actually turning on. it is hardly even getting perhaps one volt beyond the gate turn on transition, by the time the pwm pulse is almost over, so it turns off quickly afterward.
the majority of the signal you see is noise, again due to a high inductance path, possibly the majority of which is in your oscope probe.
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top