Good Cheese - Student Edition

Why can't you do the same thing in CMOS? Have one pole, equivalent to Miller capacitance off the output pin. [1] Then, added load capacitance just slows down that pole. The trick is to not bury the loop compensation on an inaccessable node, which makes the loop go 2nd order, but make it part of the load.

It's an easy way to sense the load impedance and adjust the loop dynamics. I'll simulate it, as an LDO, if I have time. Of course, the error amp has to be wideband and uncompensated, which in my case means using a fast enough opamp that its internal rolloff doesn't affect the loop. Actually, maybe I can get a similar effect by just changing some parts values in my circuit, using the fet's Cdg as the dominant pole. On a board spin, I'd probably add a cap. Then adding load C just slows the 1st-order loop down, but the C also stiffens the transient load response, so it all works out.

All released to the public domain, just in case.

[1] in an IC, "equivalent" would probably have to be synthesized, since you probably don't want any huge caps on-chip.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

Interesting article...

"When Bode Plots Fail Us"

...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

| 1962 | I love to cook with wine. Sometimes I even put it in the food.

Doesn't work quite as nicely in CMOS as it does in bipolar... gate potentials move a lot, base voltages are, for all intents and purposes, "constant". So you have to play cascode games in CMOS to get good behavior. Either way it costs a pin, which customers complain more about than the external cap :-(

And all guaranteed to work as well as Obama's budget ?:-)

Not without taking up more area than the capacitor itself.

Here's an interesting article about Bode Plots versus measuring output impedance (real and imaginary parts)....

...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

No, ignore that. My circuit is a follower. I'd have to flip the fet over to do the Miller thing.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

I knew what you meant ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

Interesting article. That's not what I thought it would be about -- but they make good points.

They mention the Nyquist plot in passing -- when I do close-loop design using frequency-domain measurements, I always make both a Bode plot and a Nyquist plot: the Bode plot to tell me about performance and to let me quote gain and phase margins, but the Nyquist plot to make sure that the absolute stability margin is good all around.

I thought, after reading the title, that the article would be about the shortcomings of frequency-domain design in time-varying nonlinear systems (like, say, switching power supplies), or maybe about looking at impedance plots to see if it goes negative anywhere (which tells you that somehow, somewhere, there's a passive load that will make your power supply oscillate).

But even if it didn't talk about what I expected, and even if I only skimmed it, it looks informative.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

I usually stay in the time domain, and shock circuits with steps and see if they ring or flail around. Frequency response is small-signal linear, but the interesting cases involve slew rates and other nonlinearities.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

One thing I added to my simulation routine, quite a few years ago now, is to examine stability of loops during power-up... ramp the supply up and watch for gotchas... stable loops often are _not_ stable somewhere along the power-up sequence. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

Yes. And JL has a good point about slopes and whatnot -- you can have a system that's dead stable for small signals, but completely falls apart for large excursions.

Which doesn't mean that you should toss the frequency-domain stuff out the window: it just means that you should take the small-signal stuff with a grain of salt, and either back it up with time-domain stuff, take large-swing frequency response data, or both.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

Data has been added to...

It now includes transient simulations of Larkin's claimed "in-production" schematic, before and after... original compensation, then compensated as shown in his later "equivalent" LTspice schematic.

The original flavor, no matter what, is unstable.

The LTspice version is stable... PSpice matches perfectly with the LTspice simulation. It always will match, choose versus HSpice, Synopsis, Mentor, whatever... perfect matches.

Of course the great and glorious question is why does Larkin have this need to fudge the data all the time.

Maybe some of you Larkin boot-lickers can say?

BTW, Larkin boot-lickers in particular are invited to drop by... might even provide beer if you want to hang your hat on something else that's non-physical... maybe Jamie could be sure to bring along his magical hub dynamo ?:-)

WARNING: All visitors are subject to video-recording, and later posting to You Tube. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

I located the original post that uses what Larkin "claims" as his LDO architecture...

From: Bob Engelhardt Newsgroups: sci.electronics.basics Subject: A couple of simple questions about a simple op amp circuit Date: Sun, 01 Apr 2012 11:34:26 -0400 Message-ID:

This is certainly stable when configured as a current sink (resistance in the source path).

36 days later Larkin claims it as his own...

From: John Larkin Newsgroups: sci.electronics.design Subject: Re: Why no depletion mode LDOs? Date: Mon, 07 May 2012 20:42:14 -0700 Message-ID:

...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

You've really lost it, Jim. What a disappointment.

John S

No, I save that for very entertaining, special events. For you, I am willing to bring my propeller hat. I've been told it is a real party popper.

And besides, maybe you're a little confused here? I work with Dynatrons, that would be the equivalent to the toy you speak of. I wouldn't mind bringing that along however, the cost of transportation, re-assembly and power lines to operated it, would be just out side my budget, not to speak of the possible incineration to your estate that would most likely occur from the multi-Million volt output at just under

100ma.

Don't think the camera equipment would survive my presents.

Jamie

What presents are you bringing? I want some.

[snip]

My apologies! I got the wrong name on the Hub Dynamo fiasco. Who was that anyway? ...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

| 1962 | I love to cook with wine. Sometimes I even put it in the food.

Every party needs a popper.

So, fans and heat sinks?

I wouldn't mind

Good! You've got him surrounded now.

BTW, Jamie has his own camera:

and does quite well with the wide angle stuff.

Yes I do, and I invite any one to my club that wishes to join.

I even carry a membership card in my wallet...

The fun starts when things get hot and sweaty.

Jamie.

Ohhh, one of *those* kinds of parties? ;-) I somehow doubt Jim would accept, though he would willingly accuse numerous people of being of that, inclincation...

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

:)