Synchronous buck up to 100% duty cycle

Thread Starter

mwilliford

Joined Oct 6, 2009
2
I'd like to build a synchronous buck converter with about the following specs, notice that I want the ability for Vin = Vout, up to 100% duty cycle.

Vin = 12-100v
Vout = 12-18v
Vin >= Vout (minus the voltage drop from Rdson)
Iout = 40amp
F = 400khz
duty cycle = 0 - 100% (in the case of Vin = Vout, 100% always on)
1.5-2amp high side driver peak current (A guess based on some mosfets I've calculated)

Controlled by a dsp, using a PID routine.

Options and Questions:
If I want the possibility of turning on the high-side (transition from buck to an always on high-side), what are my options for driving this sync buck?

1. If I use a typical high-side driver that calls for a bootstrap capacitor (which I think would normally limit my duty cycle to somewhere well below 100%), can I replace Cboost with a dedicated power supply 10volts above Vout? Somewhere between 22-24 volts for the high side gate supply. How would I wire this, I can't find application notes that discuss such a use of a high-side driver? Can a high-side mofset driver be driven to 100% duty cycle with a dedicated Vboost supply?

2. If not, I have considered using a high-side driver for sync buck operation (with bootstrap cap), then when I want to switch to always on mode, I could have some op amp (with dedicated supply) power the gate. Of course the DSP would shut down the buck pwm and use a ttl output for the op amp. This complicates the design, I prefer option #1 above.

3. I looked for a charge-pump high-side driver that could run at 100% duty cycle, but didn't find anything. Perhaps this is an option if such a part exists.

Perhaps it would help to know why I want to do this:

Reason: At some voltage of Vin (perhaps near 13v), it would be more efficient to directly connect the Vin = Vout. I can limit the loss to Rdson of the high-side (plus overhead of gate supply and dsp), and skip the switching losses.

I'm obviously new to using high-side mosfet drivers with bootstrap capacitors, any help would be appreciated.

Thanks,
Marcus
 

JDT

Joined Feb 12, 2009
658
1. If I use a typical high-side driver that calls for a bootstrap capacitor (which I think would normally limit my duty cycle to somewhere well below 100%), can I replace Cboost with a dedicated power supply 10volts above Vout?
Yes you can. Use a small DC-DC converter to power your high-side MOSFET. It will need to have no more than 12V output. The -ve output of this supply goes to the high-side source, +ve basically across the bootstrap capacitor. You will still need the bootstrap capacitor as you need a reservoir of charge to charge the gate-source capacitance. The peak current may be amps!

You will have to check the datasheet of your high-side driver to be sure that there are no other restrictions to 100% on-time. Sometimes they have capacitive coupling on the chip for the gate drive signal.

It is possible to get an optically-coupled MOSFET driver. Use this with your DC-DC converter. Look at Avago ACPL-3130-300E (Farnell 164-0539)

See:- http://www.farnell.com/datasheets/76297.pdf

I've picked one type a random. There seems to be quite a large range.
 

Thread Starter

mwilliford

Joined Oct 6, 2009
2
I'll look into the opto-coupled high-side drivers, I didn't think of those. Thanks!

Regarding a bootstrapable high-side:
I did plan on having some sort of supply for the high-side driver, however if It isn't has high as the Vgs of the high-side mofset, it will rely on the boostrap capacitor to charge on the off-cycle, correct? This is my understanding of the various datasheets i've read. This limits my duty cycle to < 100%, am I correct? Unfortunately most datasheets don't talk about this, so it is a bit hard to figure out what the possibilities are. So I was thinking of having both a supply for the high-side driver itself AND a higher supply to replace the boostrap capacitor. This higher supply would have a capacitor on the ouput capable of the 2amp surge for 25ns or so. Then I think I could turn on the high-side 100%.

Thoughts if this would work?
 

JDT

Joined Feb 12, 2009
658
By having a separate supply for your high-side driver you are no longer "bootstrapping".

Your separate supply needs to have a high enough voltage to fully switch on the MOSFET. 12-15V will be OK. This is referenced to the source of the MOSFET.

The actual voltage of the MOSFET above the system ground is not relevant.

See attached drawing.
 

Attachments

Top