high current 0 to 40 volt pwm and v-reg power supply / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

high current 0 to 40 volt pwm and v-reg power supply

previous 1 2 3 4 next

Move Thread

LAN_403

Neet Studio

Tue May 22 2012, 01:17AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

Okay... If you only want PWM, but no regulation. Not sure why you would even need to have a voltage feedback with the LM339 etc. Just connect it up like Test circuit on TI's datasheet on figure 1 and use DTC to vary the duty cycle.

Now the comment about the removal of output cap, MOSFEET driver, MOSFET sizes still stands.

You are going to be driving 10nF of gate capacitance with 5 IRFP250!! Remind me to buy stock at IRF.

Neet Studio

Tue May 22 2012, 02:08AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

What you have for gate driver right now won't work. This circuit should make the output as least closer to the 555 "200mA" push-pull you had before.

See part of my working circuit attach (idea stolen from Art of Electronic). My part also has similar to the TL494 collect & emitter of the output driver (DRV_C, DRV_E).

When Pin 5 is high, D4 conduct and drives the MOSFET.
When Pin 5 is off. R1 pulls the pin low, thus D4 is reverse bias. Q2 now pulls MOSFET gate low. You can probably get sink around 200-300mA peak with this. Size your signal/switching diode/transistor appropriately.

I still think it is insufficient to drive 10nF of gate capacitance with a 200mA push/pull...

brandon3055

Tue May 22 2012, 03:47AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

As I said before the reason I can't run feedback straight into the PWM is because the positive output didn't change the negative dose unless you mean negative feedback but I don't think that would work because the PWM dose not run on the 40v supply
But as i said I am still open to new ideas (it haven't actually built anything yet)so i may consider a buck converter as soon as i figure out how they work the basic circuit I saw on the link you posted looks exactly the same as a boost converter so in the difrence in the driver circuit? Also I don't have any mosfets drivers accept 1 tc4420 which I originoly boaght for a micro Sstc that I never got around to building would that drive 4 or 5 irfp250's

Neet Studio

Tue May 22 2012, 04:24AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

As for the TC4420, the datasheet claims that "High Capacitive Load Drive Capability: 10,000 pF" which is exactly what you have (10nF). So you have the right part.

There are a couple of ways of doing this.

>because the positive output didn't change the negative dose
You could use a differential amplifier with an external opamp and 4 resistors to subtract the +40V rail and the switched ground rail. See figure 4:

If you pick the correct ratios of the 2 pairs of resistors, you can have both the voltage scaled down to the proper levels as well as them subtracted like a differential probe on a scope and the output is now properly referenced to ground You can now run them directly to the error amp and make a proper feedback circuit. All of this with better bandwidth, less parts and no optoisolators.

The standard buck topology once you have understood it could easily be flipped upside down to make it to work with switching the ground side.
------------------------------------------------- -----------------------
Another option is to use your MOSFET to switch the positive rail.
- you'll need a charge pump (or of you have an extra winding off the transformer) sitting on top of the +40V to provide +50V or so to power the MOSFET driver
A charge pump can be built with one phrase of the TL494 output, plus the pseudo push-pull circuit I had, with 2 diodes and 2 caps. This is similar to those "Bootstrap" circuits for the upper H bridge.

- an external transistor with 2 resistors to pull it up to the +50V rail. (because the TL494 is only rated for 40V not 60) 2 resistor because you don't want the external transistor to short the schottky diode (see below) directly to ground across 40V.
- and run the MOSFET driver floating on the +40V rail. Clamp its input with a schottky so that it doesn't drop below its ground and blow the MOSFET driver.

brandon3055

Tue May 22 2012, 04:34AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

I think I understand most of that but i don't quite understand the setup of the driver ic but if I switch the positive rail I cant emagin it would be to.hard to set up a gdt driven by the gate drive ic in a push-pull configuration
Also do you think i would be better off building a buck converter?

Neet Studio

Tue May 22 2012, 04:41AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

Or a gate transformer with your gate driver chip and not deal with level shifting etc! Forgot that I am on 4hv.

You are much better off with a proper buck converter as now you can actually use it as a high current DC supply (with some filtering or post regulation) could replace the linear supply. Unless you are going to use the PWM as a driver solely for flyback or something else in mind.

Neet Studio

Tue May 22 2012, 05:01AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

Okay. Just for completeness. This is what I talked about - charge pump, level shifter and gate driver. Not tested, but that's what my design would start from there. I would run the output in "parallel/singled ended mode" by grounding pin 13.

but yeah a GDT with the gate driver would work well enough without dealing with charge pumps and level shifting. lots of examples and people know how to do that here. You can switch the +40V rail with N-MOSFET and reference all the feedback stuff to ground without other parts.

brandon3055

Tue May 22 2012, 05:22AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

I am just going to use this as a general purpose power supply I think i will go with the idea of using a gdt and switching the positive rail but I will make a new schematic when I get home gust to make shore we are both thinking of the same thing

Edit: also i would still like to learn how a buck converter works so if you know a good sight with some example circuits please leave a link thanks

Edit: also about how I paralleled the outputs on the 494 I wasn't shore about that myself but i have seen a fuw other circuits that do that and seen to work ok and i think i also remember reading in the datasheet that the outputs can be run I parallel to get 500ma output but i may be mistaken

edit: is this what you had in mind
]tl494_vevision_2.zip[/file]i also found this dose this mean that to run the outputs in paralell the output control pin needs to be conected to ground?

Neet Studio

Tue May 22 2012, 11:55AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

>also found this dose this mean that to run the outputs in paralell the output control pin needs to be conected to ground?
Yes. That's why I said you had it connected as "wired-or" for opposite phrase for your previous rev.
Connecting to ground runs both outputs in phrase, so they are both switching at the same time in parallel.

The output should be connected to the MOSFET driver directly not via a resistor.
You'll need a pull down resistor (to ground) on the output of the TL494. I'll say 10-20mA is a good start. It can go higher, but you are only driving a MOSFET gate driver chip.

So ~1K - 500 ohms, work out the wattage.

Reading material: go to TL494 webpage on TI

and read up on app note materials there
"Understanding Buck Power Stages In Switchmode Power Supplies" would be a good start on the math and theory.

I haven't tried this, but it sounds like something useful:

>Power Stage Designerâ„¢ Tool helps design the power stage of the most commonly used switchmode power supplies. It is also a great assistance for getting a deeper understanding of voltages and current flows inside converters.

There should be lots of app notes of Buck mode regulators on TI, Linear Technology, Maxim, Analog Devices etc. They have analog engineers writing these full time (enough said). IEEE papers are probably way over your head and do require a subscription.

Linear Technology mainly uses current mode (vs Voltage mode in TL494), but their app notes are probably one of the best. Read up on their layout part about the high frequency/high current paths for the power components etc. Also treat the gate driver path as high frequency/high current.

One thing I like to point out is that you'll find that at some point in your calculations, the power losses would go up rapidly and some of the components are starting to get out of hand. This is the reason why you see multiphase power supply above 20-30A and they use synchronous rectifier at above 10A.

I'll say up front that the goal of 50A / 40V is not a easy one. It'll be a lot of designing and learning, but hey that's the fun part, right?

brandon3055

Tue May 22 2012, 01:14PM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

thanks for all your help i will have a good look at the design tool im shore it will cume in handy but do you have any sugestions in wether to use a single output on the gdt or one for each mosfet and how many windings should i use im thinking about 20 to 30 torns my core matireal is L8 and 25MM

previous 1 2 3 4 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob