AoE x-Chapters, 4x.26, MOSFET current source, nodal analysis

[Winfield Hill]

Here's a new section I'm hoping to complete, so
it can be added to the x-Chapter book before it
goes to the printer in a few weeks. Please look
it over, but don't be too harsh, about its lack
of mathematical vigor. It's closer to our usual
back-of-the envelope approach to calculations.
Fixes for errors, suggestions for clarification,
improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1


--
Thanks,
- Win

 

On Friday, 9 August 2019 21:50:34 UTC+1, Winfield Hill wrote:

Here's a new section I'm hoping to complete, so
it can be added to the x-Chapter book before it
goes to the printer in a few weeks. Please look
it over, but don't be too harsh, about its lack
of mathematical vigor. It's closer to our usual
back-of-the envelope approach to calculations.
Fixes for errors, suggestions for clarification,
improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

loopy noodle analysis?

Brain will have to wait to take it in.

 

[John Larkin]

On 9 Aug 2019 13:50:20 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

Here's a new section I'm hoping to complete, so
it can be added to the x-Chapter book before it
goes to the printer in a few weeks. Please look
it over, but don't be too harsh, about its lack
of mathematical vigor. It's closer to our usual
back-of-the envelope approach to calculations.
Fixes for errors, suggestions for clarification,
improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

I Spice stuff like that. That includes nonlinearities, real part
models, and lets me quantify things like step response accuracy.

Besides, I've forgotten most of that college math.

 

[Winfield Hill]

John Larkin wrote...

On 9 Aug 2019, Winfield Hill wrote:

Here's a new section I'm hoping to complete, so
it can be added to the x-Chapter book before it
goes to the printer in a few weeks. Please look
it over, but don't be too harsh, about its lack
of mathematical vigor. It's closer to our usual
back-of-the envelope approach to calculations.
Fixes for errors, suggestions for clarification,
improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

I Spice stuff like that. That includes nonlinearities,
real part models, and lets me quantify things like
step response accuracy.

"Some designers turn to SPICE to analyze the circuit,
but be aware that the MOSFET’s SPICE model will likely
be highly defective at low currents, see Figure 3x.35,
not to mention possible problems with the op‐amp’s
capacitive‐load model at high frequencies"

The Figure 3x.35 reference is in Chapter 3x.5, posted:
https://www.dropbox.com/s/0ph15moamo0mlt4/3x.5_MOSFETs_Linear-Transistors_DRAFT.pdf?dl=1

This is one of those case where relying on SPICE is
not a good idea. Plus, when you have an analytical
solution to your circuit, you can more easily see
what the trade offs are, and optimize the circuit.


--
Thanks,
- Win

 

[John Larkin]

On 9 Aug 2019 17:55:28 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

John Larkin wrote...

On 9 Aug 2019, Winfield Hill wrote:

Here's a new section I'm hoping to complete, so
it can be added to the x-Chapter book before it
goes to the printer in a few weeks. Please look
it over, but don't be too harsh, about its lack
of mathematical vigor. It's closer to our usual
back-of-the envelope approach to calculations.
Fixes for errors, suggestions for clarification,
improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

I Spice stuff like that. That includes nonlinearities,
real part models, and lets me quantify things like
step response accuracy.

"Some designers turn to SPICE to analyze the circuit,
but be aware that the MOSFET’s SPICE model will likely
be highly defective at low currents, see Figure 3x.35,
not to mention possible problems with the op‐amp’s
capacitive‐load model at high frequencies"

The Figure 3x.35 reference is in Chapter 3x.5, posted:
https://www.dropbox.com/s/0ph15moamo0mlt4/3x.5_MOSFETs_Linear-Transistors_DRAFT.pdf?dl=1

This is one of those case where relying on SPICE is
not a good idea. Plus, when you have an analytical
solution to your circuit, you can more easily see
what the trade offs are, and optimize the circuit.

But how does the analytical solution deal with nonlinearities, like
the low current that you mentioned? A bunch of small-signal math
analysies, at different operating points, is not only tedious, they
miss the point, that the circuit will be nonlinear during a single
event. Like, for a current step, an opamp may wind up in the region
where the mosfet is barely turned on, and overshoot later when fet
gain is high.

I personally don't "see" much in an analytical solution. Some people
do. I do see a lot in a waveform or a stepped-parameter set of
waveforms.

Sometimes "good" can't be expressed as an equation. Usually goodness
is a complex tradeoff that can't be quantified: tune it until you like
it and think you can sell it.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Jan Panteltje]

On a sunny day (Fri, 09 Aug 2019 20:44:29 -0700) it happened John Larkin
<jjlarkin@highlandtechnology.com> wrote in
<7ceske9roatjlumf42l7lb3lcj4tqgcmum@4ax.com>:

I personally don't "see" much in an analytical solution. Some people
do. I do see a lot in a waveform or a stepped-parameter set of
waveforms.

Sometimes "good" can't be expressed as an equation. Usually goodness
is a complex tradeoff that can't be quantified: tune it until you like
it and think you can sell it.

Right,
the inventor of the wheel did a great job.
The fight over Pi came later.


J.P.
Weapons of math destruction

In the future it will be all AI neural nets,
no questions asked, and no answers given.

You gotto know a little bit about maaz though (if only to pass the exams)
but beware of strings theory attached,

 

[Jan Panteltje]

On a sunny day (9 Aug 2019 13:50:20 -0700) it happened Winfield Hill
<winfieldhill@yahoo.com> wrote in <qikmac01pg9@drn.newsguy.com>:

Here's a new section I'm hoping to complete, so
it can be added to the x-Chapter book before it
goes to the printer in a few weeks. Please look
it over, but don't be too harsh, about its lack
of mathematical vigor. It's closer to our usual
back-of-the envelope approach to calculations.
Fixes for errors, suggestions for clarification,
improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

do not use '&' in filenames
do not use spaces in filenames


># xpdf 4x.26_Loop\ &\ Nodal\ Analysis.pdf
[2] 11547
zsh: command not found: Nodal Analysis.pdf
# Error: Couldn't open file '4x.26_Loop '

[2] + exit 1 xpdf 4x.26_Loop\

# mv 4x.26_Loop* 4x.26_loop_and_nodal_analysis.pdf

# xpdf 4x.26_loop_and_nodal_analysis.pdf
OK now

# mv 4x.26_loop_and_nodal_analysis.pdf winfield/

 

[Tim Williams]

"Jan Panteltje" <pNaOnStPeAlMtje@yahoo.com> wrote in message
news:qiljlc$u5k$1@dont-email.me...

do not use '&' in filenames
do not use spaces in filenames

Bwahaha, it's 2019. You old *nix fogies are still having problems with
this?!

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Design
Website: https://www.seventransistorlabs.com/

 

[Phil Hobbs]

On 8/9/19 4:50 PM, Winfield Hill wrote:

Here's a new section I'm hoping to complete, so it can be added to
the x-Chapter book before it goes to the printer in a few weeks.
Please look it over, but don't be too harsh, about its lack of
mathematical vigor. It's closer to our usual back-of-the envelope
approach to calculations. Fixes for errors, suggestions for
clarification, improved accuracy, and comments welcome.

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

Nice. I like your making a virtue out of a necessity (hand-drawn
figures).

One point that might be worth a footnote is that Kirchhoff's laws are a
low-frequency approximation, applicable only when radiation and
self-capacitance are negligible. It's surprising how many folks (even
some who know more about antennas than I do) treat them as Holy Writ.

Of course, K's laws are at about the same level of approximation that
allows us to draw schematics and reason about them, but they aren't on
the same level as conservation of charge, for instance.

We can also write that v3 is higher than Vg by the voltage i3 R3
across R3.

I know what you mean, but "higher than" might not be the best phrase
when i3 R3 can be negative. A friend who's a much better writer than I
am uses the acronym RWWATP: Real Writers Write Around The Problem.

We can also write that v3 is higher than Vg by the voltage i3 R3
across R3.

This is true of course, but the following equation needs another
sentence or two of introduction, or you'll lose people. A sentence like
"This ignores the DC output of the amplifier, which is OK because we're
doing a small-signal AC analysis here." It's a bit confusing because
the middle bit is an approximation but the last bit is exact.

The middle bit also assumes that the frequency is low enough that the
inverting and noninverting gains of the amp have the same magnitude, but
you don't say that anywhere. Maybe write the equalities in order from
exact to approximate?

Your wheeling in your colleague Alan Stern to do the math reminds me a
bit of Woody Allen in "Annie Hall", when he pulls Marshall McLuhan into
the scene.
<https://www.youtube.com/watch?v=9wWUc8BZgWE>

In the expression for Zg, you're assuming that the drain is sitting
still. You do say you're ignoring Crss, but a sentence explaining why
that's OK would be helpful. The Miller effect is usually pretty
important in high frequency analyses. You're also ignoring the Early
effect, which can be important even in small-signal situations because
it limits the available voltage gain. (Is there a better name for the
Early-effect equivalent in MOSFETs?)

At the top of P3, Zg = 1/sCiss + R1.

I'd leave out the 'duh's. (I assume you would too, but they might get
overlooked in editing. That commonly happens with swear words in source
code comments, for instance. Very embarrassing.)

We’ll assume the voltage across R1 is equivalent to our current
output.

That assumption needs motivating--it assumes that R3 >> R1, for instance.

If we assume the (alpha+beta)/(alpha‐1) term is about unity, the
equation says the op‐amp’s influence is going away above a cutoff
frequency fc = 1/2pi R2 C2, which is not a surprise.

There's apparently nothing motivating this assumption--it would be
clearer stated the other way round. That whole paragraph is too terse,
I think. Which terms are being neglected, and what additional
assumptions does that involve?

>But this is still pretty fast, current pulses

Comma splice alert!


Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Tom Gardner]

On 10/08/19 01:55, Winfield Hill wrote:

Plus, when you have an analytical
solution to your circuit, you can more easily see
what the trade offs are, and optimize the circuit.

That should be writ large, and engraved in all
university courses.

Unfortunately it is becoming a lost art

 

[Jan Panteltje]

On a sunny day (Sat, 10 Aug 2019 01:25:33 -0500) it happened "Tim Williams"
<tiwill@seventransistorlabs.com> wrote in <qilo0k$eku$1@dont-email.me>:

"Jan Panteltje" <pNaOnStPeAlMtje@yahoo.com> wrote in message
news:qiljlc$u5k$1@dont-email.me...
do not use '&' in filenames
do not use spaces in filenames

Bwahaha, it's 2019. You old *nix fogies are still having problems with
this?!

Tim

You silly widows users still do not understand the world is MUCH greater than your latest auto update security hole.

 

[Winfield Hill]

Jan Panteltje wrote...

https://www.dropbox.com/s/7zl3yi789idg3s8/4x.26_Loop%20%26%20Nodal%20Analysis.pdf?dl=1

do not use '&' in filenames
do not use spaces in filenames

OK, thanks, good suggestion for DropBox links.
https://www.dropbox.com/s/beyu2toxgvvedf9/4x.26_CS_Nodal-Analysis.pdf?dl=1


--
Thanks,
- Win

 

[Winfield Hill]

Phil Hobbs wrote...

[ snip ]

Phil, thanks very much for your comments,
I have made a dozen changes. And Paul
will no doubt rework it further as well.


--
Thanks,
- Win

 

[Tim Williams]

"Jan Panteltje" <pNaOnStPeAlMtje@yahoo.com> wrote in message
news:qim2na$vu$1@dont-email.me...

You silly widows users still do not understand the world is MUCH greater
than your latest auto update security hole.

Windows? No, I mean I don't even hear my Linux friends complaining of shit
that archaic. They used to. I assume it's been fixed. Are you using 90s
Redhat or something? Update that security hole!

Or they still mumble about it but have resigned themselves to hopelessness
over the years...

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Design
Website: https://www.seventransistorlabs.com/

 

[Tom Gardner]

On 10/08/19 14:47, John Larkin wrote:

On Sat, 10 Aug 2019 08:11:49 +0100, Tom Gardner
spamjunk@blueyonder.co.uk> wrote:

On 10/08/19 01:55, Winfield Hill wrote:
Plus, when you have an analytical
solution to your circuit, you can more easily see
what the trade offs are, and optimize the circuit.

That should be writ large, and engraved in all
university courses.

Unfortunately it is becoming a lost art

Learning classic circuit theory and analysis is critical to doing
original circuit design, but it's just the starting point.

Design is the opposite of analysis. And most all the interesting stuff
is seriously nonlinear.

What the theory can do is provide insight, guide creative fiddling.
Ultimately most of us solder parts to boards to make stuff that works,
not publish papers.

Completely understood and accepted. Anything significantly
non-linear virtually requires number crunching rather than
standard analysis.

Nonetheless, an analytical solution to a /simplified/ model
can yield valuable insights. The classic simplified model
in physics is exemplified by "...assume a spherical cow...".

There are many similar things in electronics, e.g. simple
model are used to estimate EMI/EMC between one comms system
and another. Imperfect? Of course; it never matches reality.
Useful? Yes.

 

[Tim Williams]

"Phil Hobbs" <pcdhSpamMeSenseless@electrooptical.net> wrote in message
news:qilqcm$rtg$1@gioia.aioe.org...

One point that might be worth a footnote is that Kirchhoff's laws are a
low-frequency approximation, applicable only when radiation and
self-capacitance are negligible. It's surprising how many folks (even
some who know more about antennas than I do) treat them as Holy Writ.

Well, they're still valid locally. You wouldn't use a transmission line (as
such) and assume the currents through one wire of it are equal at both ends;
but you can assume so for the two terminals of each port. Likewise, the
assumption breaks at frequencies where higher order (non-TEM00) modes are
active; but at those frequencies, we don't call it a transmission line
anymore (well, descriptively perhaps, but not theoretically).

More accurately, the applicable locality is proportional to wavelength times
approximation tolerance. In the lambda-->0 limit, it's simply the
conservation of charge for a differential volume, or, rewritten a bit...
Ampere's or Faraday's laws I think?

This kind of discussion can quickly get pedantic, but it's a good idea to
have somewhere. It's beyond the scope of this section; but, the
transmission lines section, I believe, has yet to be seen? That's the
perfect place. Just dropping a footnote to it here, would be nice. (Also
in the section on R/L/Cs and their equivalent circuits, which arise for
similar reasons; and probably other places?)

If we assume the (alpha+beta)/(alpha‐1) term is about unity, the
equation says the op‐amp’s influence is going away above a cutoff
frequency fc = 1/2pi R2 C2, which is not a surprise.

There's apparently nothing motivating this assumption--it would be
clearer stated the other way round. That whole paragraph is too terse, I
think. Which terms are being neglected, and what additional assumptions
does that involve?

Regarding the formatting here -- it may feel better with omega_1 and omega_2
terms (1/RC), instead of alpha and beta (frequency-dependent gain terms), or
maybe included in them, or maybe doing omegas first and then instead of the
rational expression with alpha and beta, a shorter H is used? I'd have to
play around with the algebra a bit to see what's neater. Anyway, that might
arguably be just an editing thing as well?

Another assumption that I didn't see supported: ignoring the current in the
sense resistor. We all know what we're doing, at least in the usual (high
output current) case, but you're also talking about microcurrent sources,
for which the resistors will be similar value, or the shunt resistor might
even be larger. In that case we cannot ignore the contribution.

A direct consequence is, instead of treating the opamp as an integrator,
it's an integrator+1 term -- the +in signal gets feed-forward into both the
gate and the shunt, giving weird DC and non-minimum-phase AC terms/errors.

An inverting configuration would be more honest here, if inconvenient; but,
I wonder also if this would have any potential improvements for dynamics?
May be worth testing. I suppose a noninverting integrator configuration
might also be possible (a variant on the Howland charge pump, with C load,
but also with inverting feedback), again with the possibility of different
performance.

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Design
Website: https://www.seventransistorlabs.com/

 

[John Larkin]

On Sat, 10 Aug 2019 03:05:57 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

Comma splice alert!


Cheers

Phil Hobbs

From wiki:


Brother Cadfael is the main fictional character in a series of
historical murder mysteries written between 1977 and 1994 by the
linguist-scholar Edith Pargeter ...


That's funny. Her continuous comma splices make the books almost
unreadable.




--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[John Larkin]

On Sat, 10 Aug 2019 08:11:49 +0100, Tom Gardner
<spamjunk@blueyonder.co.uk> wrote:

On 10/08/19 01:55, Winfield Hill wrote:
Plus, when you have an analytical
solution to your circuit, you can more easily see
what the trade offs are, and optimize the circuit.

That should be writ large, and engraved in all
university courses.

Unfortunately it is becoming a lost art

Learning classic circuit theory and analysis is critical to doing
original circuit design, but it's just the starting point.

Design is the opposite of analysis. And most all the interesting stuff
is seriously nonlinear.

What the theory can do is provide insight, guide creative fiddling.
Ultimately most of us solder parts to boards to make stuff that works,
not publish papers.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Jan Panteltje]

On a sunny day (Sat, 10 Aug 2019 06:52:59 -0700) it happened John Larkin
<jjlarkin@highlandtechnology.com> wrote in
<9uitke92j0216tpuapdrdrodfl1krbefnu@4ax.com>:

On Sat, 10 Aug 2019 03:05:57 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

Comma splice alert!


Cheers

Phil Hobbs


From wiki:


Brother Cadfael is the main fictional character in a series of
historical murder mysteries written between 1977 and 1994 by the
linguist-scholar Edith Pargeter ...


That's funny. Her continuous comma splices make the books almost
unreadable.

Never heard of Cadfael, but did watch Catweazel in the seventies.
https://en.wikipedia.org/wiki/Catweazle
he is the inventor / embodyment of "elec-trickery" (electricity) and the "telling bone" (telephone)
https://www.imdb.com/title/tt0063881/mediaindex

Have to think of his way of seeing things a lot...

 

[Tom Gardner]

On 10/08/19 15:56, John Larkin wrote:

On Sat, 10 Aug 2019 15:38:39 +0100, Tom Gardner
spamjunk@blueyonder.co.uk> wrote:

On 10/08/19 14:47, John Larkin wrote:
On Sat, 10 Aug 2019 08:11:49 +0100, Tom Gardner
spamjunk@blueyonder.co.uk> wrote:

On 10/08/19 01:55, Winfield Hill wrote:
Plus, when you have an analytical
solution to your circuit, you can more easily see
what the trade offs are, and optimize the circuit.

That should be writ large, and engraved in all
university courses.

Unfortunately it is becoming a lost art

Learning classic circuit theory and analysis is critical to doing
original circuit design, but it's just the starting point.

Design is the opposite of analysis. And most all the interesting stuff
is seriously nonlinear.

What the theory can do is provide insight, guide creative fiddling.
Ultimately most of us solder parts to boards to make stuff that works,
not publish papers.

Completely understood and accepted. Anything significantly
non-linear virtually requires number crunching rather than
standard analysis.

Nonetheless, an analytical solution to a /simplified/ model
can yield valuable insights. The classic simplified model
in physics is exemplified by "...assume a spherical cow...".

There are many similar things in electronics, e.g. simple
model are used to estimate EMI/EMC between one comms system
and another. Imperfect? Of course; it never matches reality.
Useful? Yes.

Circuit design starts with the topology problem: what is the schematic
that we want to analyze? Where does it come from?

In my EE school, when I talked about designing things I was told
"Undergrads don't design; that starts in graduate school" so I didn't
apply for grad school.

Bizarre. Objectionable. My alma mater continues to have a
completely different tradition, 40 years later. They have
extremely well equipped labs, and the undergrads use them
at any time for any project, including personal projects.