Page 1 of 1

the KE1GF 100KHz Voltage Mode PWM Generator

PostPosted: Sat May 29, 2004 7:16 am
by ke1gf
How's everybody doing? I've decided to come up with my own PWM generator and MOSFET driver for learning, experimentation, curiosity and just plain bordome...

here's what I have as of right now

I'm running off a single 8AH 12V sealed lead acid

Image


Image

My bridge/2 (totem pole) driver is able to chage the .004uf gate of an IRFP260N in about .4 uS... think that's fast enough?

Best
-'GF

Maybe

PostPosted: Sat May 29, 2004 5:03 pm
by N9NEO
Hi Bill,

400ns probably isn't fast enough. It sounds like you're only getting about 100ma.

I assume you are going to be running some type of hard switched forward or buck type controller. These modes cause a lot of stress on transistor when switching. The freewheel diode is conducting when you turn transistor on. You have to supply enough current to sweep the Miller charge.

I=Cdv/dt at 0.10A you will only sweep 1A/.004u V/S. This is 250v/uS If you are switching a 160v bus, this will take you about .6us to turn-on or turn-off. This is a large percentage of total time if switching at 100kHz.

Losses can be approximated as Vbus*I*1/2*Tswitch. This then is in Joules which can be multiplied by the carrier rate to get watts.

You will need something bigger. Look at Elantec driver chips that Todd has mentioned in posts here.


I have a board with the resonant full bridge controller on it for you. I knew you were busy at school, so didn't want to bug you. (Graduated Jessica this afternoon at Fitchburg :) I have three of the converters running over here if you want to take a look. I've also got lots of parts kicking around too. I'm out of work next week so lemme know if you can stop over.

73
Bob

But if you're up for a challenge

PostPosted: Sat May 29, 2004 6:19 pm
by N9NEO
A method which would be fairly straightforward to implement and probably better than a bridge or an interleaved forward converter would be a resonant forward converter running at about 500kHz.

You could get by with one or two fets in parallel and a very simple power structure. With a closed loop design *which is basically cookbook* the input bus capacitance would be about 5000uf-10000uf at full pep. A very inexpensive design which is easy to understand and build. Could very well be best all around implementation for Class E type power levels.

The waveform would look like the typical class E waveform. The idea is to keep the off-time constant so that the reverse voltage rings up and then back down like a class E rig. You vary the on-time in order vary the output voltage.

I have a very small forward resonant design like this running at 1mHz. It is fixed square wave drive and is used to get 15v across a 320vdc barrier. Had a clock signal available from micro so it was natural.

73
Bob

PostPosted: Sun May 30, 2004 2:19 am
by ke1gf
Bob, sounds good! Grandpa is turn ing the big 97 this Thursday the 3rd so that day is out... Or mabye you ment the week after this one.


Thanks,
-Bill

Re: the KE1GF 100KHz Voltage Mode PWM Generator

PostPosted: Sun May 30, 2004 6:53 am
by steve_qix
ke1gf wrote:How's everybody doing? I've decided to come up with my own PWM generator and MOSFET driver for learning, experimentation, curiosity and just plain bordome...

here's what I have as of right now

I'm running off a single 8AH 12V sealed lead acid

My bridge/2 (totem pole) driver is able to chage the .004uf gate of an IRFP260N in about .4 uS... think that's fast enough?

Best
-'GF


Hey Bill, looks good.

What are you using for your PWM generator?

Regards,

Steve

PostPosted: Sun May 30, 2004 8:42 am
by ke1gf
Steve, I started off with a 555 timer chip as my pwm generator I used a 2n3904 and a 2n3906 to create a differential time constant in the feedback loop of the chip and I controled the duty cycle via the current to the bases of the transistors... this was kind of cool except as the PW was modulated the frequency would vary... I ended up burning up all my chips trying to get 0...100% duty cycle :D it would of been neat to see it work...


So I grabbed your 7404 oscillator and started building an integrator and comparitor out of the parts I had in the junk box. Basically I came up with a dual voltage supply using 2 7805 regulators so I have 0,5,10,12 volts available to me. I used a single TL074 as not only an integrator but also a comparitor Vcc is 10v Vdd is 0 volts and the non inverting input of the integrator is connected to 5v. I get a good approx 4.5v +/- triangle out and I feed that into the other side of the TL074 acting as a comparitor. so the error voltage range is +.5v to +9.5 0...100%. To buffer and invert I use another 7404 Vcc +10v and Vdd +5v so it floats along on the 5v supply. The inverter drives a single 2n3904 dc bypassed with a 10uf cap... 510ohm to the collector to +12v, emiter grounded and 10k feedback to the base from the collector. This boosts up the TTL to 0v/+12v and I use that amp to drive a 2n3904/3906 half bridge. witch poorly drives a single IRFP260N. See bob's comments.

Well it was fun to play around and get something somewhat working.

I had a nice regulated 40 volts across a cap at a few mA from a +12v batt...

Once I feel confident I'll play around with that UCC25701 that you use in your single chip PWM and use a real MOSFET driver etc, etc, etc...[/code]

PostPosted: Sun May 30, 2004 10:07 pm
by frank carcia
Bill,
I'm impressed, You have returned to the stone age roots of PDM. Gee Steve remember some of those early circuits driving tubes. I have a box full of low voltage high current FETs if you need any. Bob'a set up is very cool I suggest you stand up right and jump past the copper and iron age.
Got beach house, no time to play radio. We installed 1/2 the windows this weekend. Looks like fall before serious radio.
Happy birthday to your GF. gfz