(not so) Great cascode paper

Here's a handy PDF on cascode circuit design.

It's a good discussion, but the expression, at least as evaluated, for the impedance at the source of the common-gate transistor is absurd: it is about 1/gm all right, but that is of the order of 100 ohms, i.e. 1/gm2, not 1 micro-ohm as stated!

This leads to the result that the Miller effect is completely eliminated in the cascode. But this is not correct: for two devices of the same type and bias the Miller multiplier is 2, because the gain from input gate to first drain is ~-gm1*(1/gm2) = -1. So the Cgd is effectively multiplied by 2.

This is not the first time I've seen this gotcha trip people up.

...yeah well... Brits, eh?

What do THEY know?!!?! :wink:

Keef

But thank goodness they are NEVER arrogant :razz:

Arrogant? -US...???

HA!!! -you fools!

Keef

It's a good discussion, but the expression, at least as evaluated, for the impedance at the source of the common-gate transistor is absurd: it is about 1/gm all right, but that is of the order of 100 ohms, i.e. 1/gm2, not 1 micro-ohm as stated!

Click to expand...

Thanks for pointing that out, especially since I would never have known this!

...and I've edited the title accordingly :grin:

Page 2 equation 2 derives a Miller-effect -3dB frequency, without any source impedance. It appears that grid-plate capacitance times gain sucks on the plate circuit, which isn't so.

It would be correct for an infinite cascade of identical stages, but that's rare.

Then it goes on to say that a load capacitor introduces a second pole....

Very pretty math and simulations, but I'm missing something.

Yeah I hope the simulator wasn't written by the author of the note ;-)

My god, I can't trust anything these days...

Is this one any more accurate?

http://www.analogzone.com/col_0412.pdf

It covers bipolars, not FETs, though...

Dennis Feucht is never all wet*, in my experience so far. He learned well from the giants at Tektronix before that company went into a dismal slump.

The part about shunt feedback combo circuits is going to get some study, but I'm willing to trust him implicitly. I was fortunate enough to get a hardcopy of his book, the material in which is now reproduced and augmented on the CD's he sells. I wish him well.


* the German comprehending people will get the gag there.

I was fortunate enough to get a hardcopy of his book

Click to expand...

You mean the Handbook of Analog Circuit Design that's available from Innovatia.com? The topics he covers look interesting, and the files are about $25 each. If it's half as good of a resource as something like The Art of Electronics has been for me, I'd gladly shell out the money for the four-volume set. I'm starting to get wary of some of the free information I find online, I'd be happy to support a good source of info with my money.

bcarso;
Is this the "CD he sells"?

http://www.innovatia.com/

Wow he's been a busy guy (or maybe not busy enough). I/m assuming all of that stuff is dennis's but maybe not.

Yes the Handbook of A C D is the one I had in mind. Looks like there are many other cool things to get as well.

I'm going to buy the book. I'm just trying to figure out exactly how to print this thing out...or if I should just leave it in its original form. It's like 1000 pages long. I think the price is very much worth it.