Anyone know what Circut Bob Pease was going on about?

From:
An EDN Blog post about Analog Design tools.

http://www.edn.com/blog/1480000148/post/240014624.html

Bob Pease at a Panel session on DFM went on about a circuit Spice said
couldn't work and held a breadboard up with the circuit he said was in
production.

Anyone know what it was?

Robert H.

 

I think Pease has gone senile. Did everyone see his _religion_ rant
in the last issue?

I went to one of his "seminars", didn't learn a thing, saw a lot of
application notes and data books on National's products.

As for "...Spice said couldn't work...", what does that mean... Spice
couldn't converge on an initial solution? Happens all the time,
doesn't mean that the circuit can't work.

Show me the circuit schematic and I assure you I can Spice it.

...Jim Thompson

 

He usually says that about the LM331.

Cheers,

Phil Hobbs

 

Anyone have a complete schematic?

...Jim Thompson

 

Pease has the problem common to columists, out of ideas but deadlines
keep coming. He's become an embarassment, but Electronic Design has,
too.

HoJo seems to be straining, too.

John

 

Inadequate and incomplete models, of course. And unknown
or unaccounted-for parasitic elements in the circuit.
Happens all the time out there in the real world (i.e.,
the world that's not IC design).

The solution is a painful process to characterize the part,
etc., and create a spice model or subcircuit that matches
over the range of interest. I also find myself making test
PCBs, and taking inductance, capacitance etc. measurements
on that. Lots of bench work, serious bench equipment, lots
of computer work, plenty of knowledge and experience about
what to look for. Given the commonplace absence of this
approach, it's rather easy to throw stones at "spice".

But given the careful, time-consuming approach, I find
spice extremely useful in my push-the-envelope projects.

 

Spice is a reasonable general-purpose program for solving systems of
nonlinear ordinary differential equations. There are some systems that
no program is going to be able to solve, but they're quite rare in
applications. Thus just about anything that can be modelled accurately
with a set of coupled ODEs can be simulated accurately with Spice.

On the other hand, not everything can be modelled with ODEs, not even
every circuit. Some better known non-ODE things, e.g. transmission
lines, have been put into Spice by hand. (A transmission line's circuit
properties aren't given by an ODE because they're nonlocal, i.e. the
output undergoes a true time delay.)

Other classes of problem that can't be modelled as systems of ODES are
transport equations--e.g. the Boltzmann equation for electron transport,
or any problem that involves convective motion, such as the air in a
heat sink.

It isn't great at multiple scale analysis, either, so for instance it
would be very poor at modelling the turn-on behaviour of a laser diode,
in which time scales from sub-femtosecond (the E & H fields) to hundreds
of milliseconds (the thermal transient) all contribute very
significantly and nonlinearly. There are codes for this kind of
problem, but Spice isn't one of them--it has to follow each cycle
laboriously, because a linearized AC analysis won't get the right answer.

And anyway, Pease's main complaint is that people use (generally poor)
computer analyses as a substitute for thought--and have tried to beat
him up with the results over the years. Most of the rest of his
posturing is for fun, I think. You can't really make a serious critique
of computer simulation by throwing a computer off the roof of the NSC
parking garage. Widlar envy, perhaps.

Cheers,

Phil Hobbs

 

Thanks.

As Win said, "inadequate Models...", could cover a multitude of sins.

I wonder how well something like a Flash device's charge storage is modeled
in Spice.

Robert H.

 

*Nonlinear* ODE's? Sounds oxymoronic, didn't know there was anything
ordinary about nonlinear DE's.

Those are actually PDE's...

Actually that is what Will Gear's integration mehod was intended to
handle, widley varying time scales.

SPICE probably could be configured to model this, but it's not going to
hand it to you on a silver platter.

Nope, Pease's main complaint is not that people use poor computer
analyses, it is that they use computer analyses poorly.

SPICE is not about solving, it is about producing numbers, numerical
integration:
http://docs.lib.purdue.edu/cgi/viewcontent.cgi?article=1301&context=ecetr

 

An ODE is a DE with one independent variable, which for Spice is usually
taken to be time, and they can be linear or nonlinear, with constant or
time-varying coefficients. YCLIU.

They're integral equations, because they're nonlocal. PDEs, e.g.
Maxwell's equations and the heat equation, are just as local as an
ODE--they just have more than one independent variable. I was careful
to say that *the circuit properties* of a transmission line aren't given
by ODEs, because they aren't. The fields inside the transmission line
obey Maxwell's equations, which as you say are PDEs. I'm talking about
the two-port circuit behaviour.

No, it was designed to handle stiff systems, i.e. those whose
eigenvalues are very different in size. Multiple scale is something
quite different. YCLIU again.

It could, if you could wait long enough for it to integrate 10**14
cycles, and could handle the resulting extreme roundoff problems. There
are better methods, e.g. multiple scale analysis. Bender & Orszag's book
on asymptotic analysis is an excellent read on this sort of stuff--and
it has lots of great pictures.

Sure, that's what I said--Spice is an ODE integration package, and what
Pease doesn't like is that many people use it as a substitute for
thought. Anyway, what's the operational difference between wrong
answers from bad models and wrong answers from stupid modellers?

Cheers,

Phil Hobbs

 

The horrible electrical case is time-domain modelling of a crystal
oscillator. You need picosecond time steps (or fs, maybe) to resolve
the actual frequency, and you may have to run for seconds until it
reaches steady-state. Besides, Spice is terrible for measuring
frequency.

I just model them in the frequency domain, and estimate behavior based
on open-loop phase-amplitude plots.


Nope, Pease's complaint is that people use computers.

John

 

Well, maybe...but using them instead of thinking is what he _ought_ to
be complaining about.

Cheers,

Phil Hobbs

 

Pease lives not far from us. It's funny, his front yard, and sometines
the street in front of his house, is always full of rustbucket old
Beetles and VW Microbusses. I guess he's just a traditional sort of
guy.

I've met him a couple of times. He's very nice, one-on-one.


John

 

On Sat, 06 Oct 2007 17:26:46 -0700, John Larkin

[snip]

You just need to know how ;-) I regularly simulate start-up in
crystal oscillators because that amount of time is part of the
specification.

[snip]

The best (most stable) crystal oscillators are "sort of" AGC'd.

...Jim Thompson

 

I've met him a couple of times at NSC seminars on Long Island. The
first time, he was plugging my book (not knowing I was there, of course)
and I introduced myself. We got on fine. Of course, in SF people would
have VW microbuses on blocks in their front yards...I bet they wouldn't
let you have Chevys like us more staid folks.

Cheers,

Phil

 

Most SF "front yards" might have room for one bicycle up on blocks.
RAP lives in a neighborhood that has actual detatched [1] houses and
real grass. One of his Electronic Design columns related his skills in
repairing rain gutters.

John

[1] the standard lot, like mine, is 24 feet wide, and houses and roofs
are in total contact. One could walk almost my entire block, end to
end, on the roofs.

 

So tell me? How can I measure the frequency to, say, 1 PPM in a
time-domain sim?

The envelope response is reasonable to simulate in time domain. Phase
noise and tempco aren't.

Yup. Crystal drive matters. But I mostly just buy oscillators now,
even OCXOs. OCXOs and TCXOs have really gotten cheap lately.

I need a new universal frequency counter. The low-end Agilent looks
nice, but it has a crappy XO, and they want another THOUSAND DOLLARS
for a mere TCXO.

John

 

You don't nor do you care ;-)

...Jim Thompson

 

Oh. Thanks for straightening me out on that.

John

hmmm... I suppose you could heterodyne it against an ideal generator,
and eyeball the difference.

 

Has it got an external input? I've got an ancient (or is that well-aged
and "even more stable"?) GenRad OXCO which I can plug into any generator
or counter that will take an external reference to make it more accurate.

With your rather more serious (than mine) operations, I'd think you
could put in one rubidium or cesium reference (and or some new-ish
method tied to GPS I'm not overly familiar with) and distribute it
throughout your labs/plant/wherever, so long as your counters and/or
generators will take an external reference frequency. Assuming decent
cable termination and appropriate drivers, I don't see that you'd need
more than one "really good" reference.