PWM generator for half Hbridge

Anywhere in the switching loop will usually do; obviously, you want it in the branches most involved with the current in question (e.g., between

+switch and -switch of a half bridge, not just hanging out there in series with the output branch). The supply side is typically the handiest, as it does not include gate drive ground loops, and is already at a high voltage so doesn't care much about transients and noise.

"Switch" encompasses anything of diode, BJT, SCR, MOSFET, IGBT.

The hardest thing you will find is clearing this advice from your mind: "minimize inductance". There are cases where such a suggestion is possible to follow; it's rarely the case in modern designs. The instruction should rather be, "optimize inductance". Which should feel more intuitively comfortable to the engineer (who must compromise everywhere: actually minimizing something is an alien task).

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams
Loading thread data ...

Hi Joerg,

Joerg wrote: []

That is correct, and for the record, we are only interested at this stage to perform some measurements on the current draw and induced ripple in order to better specify the power supply part, therefore we are not concerned about regulation, an open loop drive will do.

Since the power supply is specified by us and supplied by a subcontractor, we need to make sure we specified it correctly and as of today one of the missing information is about the induced current ripple the PS will see at its outputs. I'm wondering what else it should matter to correctly size the DC/DC and output regulators in order not to have surprises when we plug the HBridges.

Al

Reply to
alb

Hi Tim,

Tim Wescott wrote: []

I assume this is to avoid lousy slew rates that would increase losses at the driver input, but it's just a guess. Why would you choose a Schmitt-trigger gate otherwise?

Uhm, maybe because the RC will smoothen the signal suddenly causing the AND to have a ringing transaction before it really switches to the correct value?

Reply to
alb

I don't know how experienced you are with MOSFET half-bridge drives, but there are other things to watch out for WRT shoot-through. It's of course necessary to use real serious MOSFET driver ICs. We usually slow down the ON drive, in part for the reason-discussed below, but maintain the OFF drive at full force, with a Schottky diode. Any resistor in series with the diode has to be low enough to resist turn-on from gate's Crss current due to the other MOSFET's rapid turn-on and the slewing drain. Finally, there's a painful issue of gate spreading resistance, wherein most of the MOSFET's area may be turned off, but a portion is still on from the internal RC delay. This aspect isn't covered on datasheets; you may need a test jig to measure and eliminate unsuitable MOSFETs.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Hi there,

Winfield Hill wrote: []

Not that expert indeed, I'm learning by doing...and reading lots of appnotes.

I'm unfortunately constrained by the availability of most of those ICs in radhard version. When it comes to hirel there's always an extremely limited choice. For the time being we've found really a few of them.

Our half-bridge is powered between +/-, with 50% duty cycle for 0 current output. So I assume you talk about ON drive for the upper branch and OFF for the lower branch of the bridge.

I see, when the upper MOS gets turned on the drain of the lower will tend to follow, leading to an unintentional switch on of the lower branch as well, hence a shoot-through.

I do I measure for this effect? We've mounted an h-bridge and "it works", but maybe we are overlooking some critical aspects...like the one you mentioned!

IIUC the off transaction is slower due to RC delays and we need to take that into account with the delay insertion, is this correct?

Reply to
alb

Your test fixture will use the exact MOSFET gate drive circuit planned for the converter. Use an low-frequency square wave to drive the gate. Connect a fast load switch to the drain of the DUT. Add an adjustable delay after the gate turnoff, and trigger the load switch. Measure the drain current (which should be zero) with a current probe. Make a plot vs time delay.

The easiest way to "measure" this effect is not to measure it! I haven't made this measurement on many MOSFETs, but the one I've looked at show a surprising difference, some are rather bad and other types have virtually no effect at all. So an easy way to deal with the issue is to look for telltale signs of shoot-through, which shouldn't be present if all the other criteria have been met.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.