On May 18, 5:42*pm, John KD5YI wrote:
So, we are saying that the load at the line input can be viewed as a
lumped circuit. So now we have a transmitter loaded with a lumped
circuit for further analysis.

It doesn't quite work that well. I gave an earlier example where Wim
got the the s11 parameter wrong by an infinite percentage. The s-
parameter equations for a lumped circuit vs an impedance discontinuity
are nothing alike. Even the IEEE definitions for the two different
types of impedances are different. The interference conditions at the
impedance discontinuity can be proven to be different than for the
lumped circuit replacement.

That's all. It's simple.

Quoting Einstein again: "Everything should be made as simple as
possible, but no simpler." :-)

When you switch to the lumped-circuit model, you are agreeing to
faster than light signal speeds, NO superposition of signals, zero
interference, zero phase shifts through coils, identical current
everywhere, etc. How the heck can you assert and prove there is zero
interference inside a source when reflected energy is flowing through
it?
--
73, Cecil, w5dxp.com

On 5/18/2011 5:58 PM, Cecil Moore wrote:
On May 18, 5:42 pm, John wrote:
So, we are saying that the load at the line input can be viewed as a
lumped circuit. So now we have a transmitter loaded with a lumped
circuit for further analysis.

It doesn't quite work that well. I gave an earlier example where Wim
got the the s11 parameter wrong by an infinite percentage. The s-
parameter equations for a lumped circuit vs an impedance discontinuity
are nothing alike. Even the IEEE definitions for the two different
types of impedances are different. The interference conditions at the
impedance discontinuity can be proven to be different than for the
lumped circuit replacement.

That's all. It's simple.

Quoting Einstein again: "Everything should be made as simple as
possible, but no simpler." :-)

When you switch to the lumped-circuit model, you are agreeing to
faster than light signal speeds, NO superposition of signals, zero
interference, zero phase shifts through coils, identical current
everywhere, etc. How the heck can you assert and prove there is zero
interference inside a source when reflected energy is flowing through
it?
--
73, Cecil, w5dxp.com

So, you're saying that the Smith chart is wrong?

On May 18, 3:42*pm, John KD5YI wrote:
....
I'm not speaking for Wim, but I think we are both saying the following:

* You have a known load
* You have a transmission line with known characteristics
* Is is possible to use a Smith chart to get the impedance at the input
to the transmission line.
* We now know the load applied to the transmitter.

All we need to know we get from the chart. We admit that reflections are
responsible for the impedance transformation from load to line input.
But, we don't need to know anything about the reflection details, energy
content of the line, nor how light would like it.

So, we are saying that the load at the line input can be viewed as a
lumped circuit. So now we have a transmitter loaded with a lumped
circuit for further analysis.

That's all. It's simple.

John

Exactly so, John. Good summary. So long as the transmitter's
bandwidth is small enough that you are always operating practically at
steady-state conditions, the transmitter can't tell the difference
between whatever assembly of transmission lines and lumped loads
distributed along those lines you want, and a simple lumped circuit
that presents the same impedance as the steady-state value of the
jumble of transmission lines out there. (For very narrow-band loads,
you might want to use a lumped equivalent that presents sensibly the
same impedance as the load across the whole transmitted bandwidth, not
just at one point.)

It is NOT that anyone is assuming "faster than speed of light," it's
that we're recognizing that the (HF voice-bandwidth) transmitter is
slower than molasses relative to the propagation times involved in a
couple hundred feet of coax, or probably even a couple thousand feet.
The attenuation per foot of the lines we use is high enough that it's
just about impossible to deviate significantly from steady-state
conditions for the bandwidths we use.

That's certainly not true for pulsed radar signals, or for fast-scan
TV, or for other wideband signals. In those cases, you'll probably
find it pays to insure the line is matched to the load so there aren't
significant reflections, and you may want to arrange the source (PA/
transmitter) to have an output impedance close to the line impedance
so it absorbs any reflections that do happen at the load end of the
line. (If you want to get fancy, you might use a circulator to insure
dissipation of such returning signals.)

Cheers,
Tom

On 19 mayo, 00:58, Cecil Moore wrote:
On May 18, 5:42*pm, John KD5YI wrote:

So, we are saying that the load at the line input can be viewed as a
lumped circuit. So now we have a transmitter loaded with a lumped
circuit for further analysis.

It doesn't quite work that well. I gave an earlier example where Wim
got the the s11 parameter wrong by an infinite percentage. The s-
parameter equations for a lumped circuit vs an impedance discontinuity
are nothing alike.


Hello Cecil,

Would you please remind me to the example where I was completely wrong
with S11?

Even the IEEE definitions for the two different
types of impedances are different. The interference conditions at the
impedance discontinuity can be proven to be different than for the
lumped circuit replacement.

That's all. It's simple.

Quoting Einstein again: "Everything should be made as simple as
possible, but no simpler." :-)

When you switch to the lumped-circuit model, you are agreeing to
faster than light signal speeds, NO superposition of signals, zero
interference, zero phase shifts through coils, identical current
everywhere, etc. How the heck can you assert and prove there is zero
interference inside a source when reflected energy is flowing through
it?

Did you ever DESIGNED some serious electronic hardware? I am not
pointing to using a recipe book or troubleshooting/repair.

--
73, Cecil, w5dxp.com

Regarding your helical; I don't have Eznec. Maybe you can use some
screenshots from it, put some comment to it and put that on website,
so we can view it.

Wim


PA3DJS
www.tetech.nl

On 19 mayo, 11:05, Wimpie wrote:
On 19 mayo, 00:58, Cecil Moore wrote:

On May 18, 5:42*pm, John KD5YI wrote:

So, we are saying that the load at the line input can be viewed as a
lumped circuit. So now we have a transmitter loaded with a lumped
circuit for further analysis.

It doesn't quite work that well. I gave an earlier example where Wim
got the the s11 parameter wrong by an infinite percentage. The s-
parameter equations for a lumped circuit vs an impedance discontinuity
are nothing alike.

Hello Cecil,

Would you please remind me to the example where I was completely wrong
with S11?



Even the IEEE definitions for the two different
types of impedances are different. The interference conditions at the
impedance discontinuity can be proven to be different than for the
lumped circuit replacement.

That's all. It's simple.

Quoting Einstein again: "Everything should be made as simple as
possible, but no simpler." :-)

When you switch to the lumped-circuit model, you are agreeing to
faster than light signal speeds, NO superposition of signals, zero
interference, zero phase shifts through coils, identical current
everywhere, etc. How the heck can you assert and prove there is zero
interference inside a source when reflected energy is flowing through
it?

Did you ever DESIGNED some serious electronic hardware? I am not
pointing to using a recipe book or troubleshooting/repair.

--
73, Cecil, w5dxp.com

Regarding your helical; I don't have Eznec. Maybe you can use some
screenshots from it, put some comment to it and put that on website,
so we can view it.

Wim

PA3DJS
www.tetech.nl

remove ED from designed....

Wim

On May 18, 6:13*pm, John KD5YI wrote:
So, you're saying that the Smith chart is wrong?

The Smith Chart is a tool - a blank graph. How could it be wrong?
Like any tool, it has limitations and can be abused.

On May 19, 4:05 am, Wimpie wrote:
Would you please remind me to the example where I was completely wrong
with S11?

----50 ohm--+--1/4WL Z0=100--200 ohm load

s11 is 0.3333 at point '+'. Put it in a box and s11 magically becomes
0.0?

The first s11 is a physical reflection coefficient, the second s11 is
a virtual reflection coefficient. The virtual 50 ohm impedance is
lossless. All the power is dissipated in the 200 ohm resistor at a
reflection coefficient of 0.3333.

Did you ever DESIGNED some serious electronic hardware?

No, but being a good designer has nothing to do with the present
academic exercise. W8JI is a good designer yet concepts like yours led
him to "measure" a 3 ns delay through a 100 uH air-core 80m loading
coil when the actual delay time is closer to 21.5 ns. That's what
happens when one relies on the lumped-circuit model and ignores
reflected energy. The relative phase of a standing wave doesn't change
with length which gives the illusion that the signal is traveling
faster than the speed of light, i.e. zero phase delay.

I will turn the coil example into a brainteaser and post it to my web
page.
--
73, Cecil, w5dxp.com

On May 19, 4:05*am, Wimpie wrote:
Regarding your helical; I don't have Eznec. Maybe you can use some
screenshots from it, put some comment to it and put that on website,
so we can view it.

Here it is: http://www.w5dxp.com/teaser2.JPG
--
73, Cecil, w5dxp.com

On 19 mayo, 15:04, Cecil Moore wrote:
On May 19, 4:05*am, Wimpie wrote:

Regarding your helical; I don't have Eznec. Maybe you can use some
screenshots from it, put some comment to it and put that on website,
so we can view it.

Here it is:http://www.w5dxp.com/teaser2.JPG
--
73, Cecil, w5dxp.com

Hello Cecil,

Your circuit (lumped inductance example) with 100V input into 72uH
with 2570 Ohms load):

From lumped circuit simulation (Beige Bag PSPICE, version 4
professional):

I_source = 32mA, -35 degrees
I_load = 32mA, -35 degrees

This agrees with hand calculation, all phase with respect to input
voltage.

From simulation, but now a pi filter C=6pF, L=72u, C=6pF, load = 2570
Ohms
Simulation carried out with same PSPICE package without using
transmission line sections:

I_source = 38mA, -1.5 degrees
I_load = 38mA, -44 degrees.

Total required time for setting up the simulations and guessing the
parasitic components to simulate the actual inductor behavior: about
15 minutes.

As you can see good agreement without using any of the photons, speed
of light, momentum and other issues, just lumped circuit simulation
where some parasitics are added. Of course a can make a better match,
but this doesn't contribute to the discussion.

I hope that some followers or contributors will do the same simulation
in a lumped circuit simulator, so that we don't arrive in a discussion
that I am cheating.

Wim
PA3DJS
www.tetech.nl

On 19 mayo, 14:03, Cecil Moore wrote:
On May 18, 6:13*pm, John KD5YI wrote:

So, you're saying that the Smith chart is wrong?

The Smith Chart is a tool *- a blank graph. How could it be wrong?
Like any tool, it has limitations and can be abused.

On May 19, 4:05 am, Wimpie wrote:

Would you please remind me to the example where I was completely wrong
with S11?

----50 ohm--+--1/4WL Z0=100--200 ohm load

s11 is 0.3333 at point '+'. Put it in a box and s11 magically becomes
0.0?

The first s11 is a physical reflection coefficient, the second s11 is
a virtual reflection coefficient. The virtual 50 ohm impedance is
lossless. All the power is dissipated in the 200 ohm resistor at a
reflection coefficient of 0.3333.

Cecil,

It is very simple, the 1/4 lamba line (100 Ohms) looks into a 200 Ohms
load, so seen from that line, VSWR = 2, hence resulting in RC=0.3333.

The 50 Ohms source looks into a 50 Ohms load (you can use the quarter
wave formula). This equals VSWR=1, so RC=0. I think I wasn't wrong!
As mentioned before, a source (whether PA or small signal) doesn't see
the difference between a lumped 50 Ohms load or your quarter wave line
with 200 Ohms load. Whether or not it is "physical" or "virtual" is
also not relevant, just the complex V/I ratio (we call that impedance)
counts.

The RC inside the line is of no relevance for the PA.


Did you ever DESIGNED some serious electronic hardware?

No, but being a good designer has nothing to do with the present
academic exercise.

It may be of importance w.r.t. selecting the right model to solve
technical problems. A good example what can happen when selecting
over-the-top approaches is this thread.

W8JI is a good designer yet concepts like yours led
him to "measure" a 3 ns delay through a 100 uH air-core 80m loading
coil when the actual delay time is closer to 21.5 ns. That's what
happens when one relies on the lumped-circuit model and ignores
reflected energy. The relative phase of a standing wave doesn't change
with length which gives the illusion that the signal is traveling
faster than the speed of light, i.e. zero phase delay.

I will turn the coil example into a brainteaser and post it to my web
page.
--
73, Cecil, w5dxp.com

Wim
PA3DJS
www.tetech.nl

Hello Cecil,

I would like to return back to the topic with a brainteaser also. You
posted links to some document written by Walt. The first reference
(QEX, May-June 2001, http://www.w2du.com/QEXMayJun01.pdf ) shows
some measuring results using load pulling in table one.

I have a source 100Vp, 4 MHz, sinusoidal, in series with a capacitance
of 796 pF (that is a capacitive reactance of 50 Ohms). Would you be so
kind to determine the output impedance of this source using load
pulling (for example using 51.2 Ohms and 44.6 Ohms). Of course I have
no problems if you (or somebody else) use a simulator to save time.

If you (or somebody else) feel uncomfortable with a zero ohm voltage
source, you may add 1 Ohms in series with the capacitor.

Did you (or somebody else) expect the calculated result based on load
pulling?

With kind regards,


Wim
PA3DJS
www.tetech.nl