gate driver

I did that some gross number of years ago, let me see here,

Ah, my original is dated 2006, so 9 years, in fact :)

Saw one submission to the Little Box Challenge that needed CMCs on the gate driver isolators, since GaN at 500V is swinging something retarded like 500kV/us slew rate.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

The balun is one of those super wideband Minicircuits things. This should work for mosfets too.

Your circuit is cute, except that the output is shorted to ground, so it can't output long pulses or DC.

So is yours (over a slightly narrower range of frequencies), or arguably even worse: my example doesn't cost dissipation (only reactive power), but how big do those 25 ohm resistors have to be? ;-)

Hmm, your gate driver\\\\keeper is also supplied wrong: there's no logic ground pin on that thing, so it just sees an input twiddling between 12 and

15.3V (assuming your comparator/receiver is a CMOS output type, which isn't specified), which is well into V_IH, so nothing happens.

You'll need to do a nice ugly level shift if you want to get any negative going voltage out of that PHEMT..

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com

I don't see a short to ground. The 25 ohm guys don't dissipate much power at my expected pulse rates. Both caps could be smaller if needed.

It seems to work (simulate) with 1 ns rise time for any pulse width to DC. Note that the e-phemt acts like a jfet in that the gate conducts around +0.7 volts.

I asked MiniCircuits about the inductance and leakage inductance of their balun, but they didn't answer. Maybe they don't know. I'll have to get some and measure them.

Sure, it needs a level shifter. My actual design has one, common-base PNP.

I expect the output to range from maybe -5 to +10.

This is just a circuit fragment, the point being to overlap a fast, direct, AC coupled gate drive with a slower DC coupled drive. Doing that, separate AC and DC paths, split and then merged, is interesting. It's easy here because the gate is clamped, by the schottky and the inherent gate diode, so the merger can be pretty sloppy. It's trickier to split an accurate analog path.

For a mosfet, the schottky could be a zener.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Hey, this is fun.

It applies the split-path ACDC thing to the level shifter. Same concept, fast lo-Z AC path and a slow, low current DC holdup. The DC current can be tiny.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Pretty. I assume this is a highish-power pHEMT. How fast does it go?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

I might use an Avago ATF-50189. They are really nice parts, and Avago seems to keep stuff in production forever.

If you drive the gate hard, these will turn on/off at, say, half an amp, in under 1 ns. The best gate drive turns on the g-s diode near abs-max current.

The phemts are really clean switches, having near zero Cgd. One cool thing about phemts is that the tab is the source, not the drain like mosfets, so the drain current doesn't ground-bounce the gate drive.

I might try transformer-coupled drive too, instead of the balun.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

THANKS for EXCELLENT contrast, _no_ grey on grey like others seem to love.

Fun.

Why the massively asymmetrical coupling caps?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Irfanview cleans up photos nicely.

Well, you've got to get scientific and simulate it and fiddle. You want to drive the lower winding hard to induce the proper EMF into the upper winding, which the wimpy LVDS receiver is trying to drive.

As mentioned, that same Minicircuits part might work better as a transformer, instead of a balun. The DC holdup injection is a bit messier, but can still be made to work.

So I'm thinking something like this...

which has a more complex AC-coupled path. Or at least different.

I'm waiting for some transformers to try this. It may be a damned fine oscillator.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com