Hello Cecil,

On 27 abr, 20:13, Cecil Moore wrote:
On Apr 27, 10:30*am, Wimpie wrote:

Depending *on the frequency resolution of your VSA, the frequency of
the injected signal can be well within 1 kHz of the carrier, so LC
filters in the PA will not distort the measurement. *In case of a 100W
PA and injection of about 100 mW, the difference in wanted signal and
signal to be rejected is 30 dB (not that large).

Would any competent optical physicist suggest that it is valid to
study the conditions associated with interfering coherent light waves
inside an interferometer by introducing an incoherent light source
into the system? Why would any competent RF engineer suggest that the
system source conditions associated with interfering coherent RF waves
can be studied by introducing an incoherent test signal?

As this slightly off-carrier frequency signal behaves like a load with
very low VSWR with a cable in between that extends with constant
speed. In other words, the amplifier sees a constant VSWR, but with
changing phase. Small frequency difference results in slow phase
change of VSWR.


Maybe you should read the postings from Tom also (K7ITM)

Wim
PA3DJS
www.tetech.nl


73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

On Apr 27, 1:43*pm, Wimpie wrote:
In other words, the amplifier sees a constant VSWR, but with
changing phase. Small frequency difference results in slow phase
change of VSWR.

From the IEEE Dictionary:

"impedance -

(1)(A) The corresponding impedance function with p
replaced by jw in which w is real. Note: Definitions
(A) and (B) are equivalent.

(1)(B) The ratio of the phasor equivalent of a steady-
state sine wave voltage ... to the phasor equivalent
of a steady-state sine wave current ...

(1)(C) A physical device or combination of devices
whose impedance as defined in definition (A) or (B)
can be determined. Note: This sentence illustrates
the double use of the word impedance ... Definition
(C) is a second use of 'impedance' and is independent
of definitions (A) and (B)."

The pinging experiment seems to be measuring a physical impedance (1)
(C) the nature of which is unclear. When the amplifier is outputting
power, it seems that the source impedance would be a V/I ratio (1)(B)
which doesn't respond to incoherent signals. Seems to me, you guys are
pinging something other than the source impedance.
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

On 27 abr, 23:18, Cecil Moore wrote:
On Apr 27, 1:43*pm, Wimpie wrote:

In other words, the amplifier sees a constant VSWR, but with
changing phase. Small frequency difference results in slow phase
change of VSWR.

From the IEEE Dictionary:

"impedance -

(1)(A) The corresponding impedance function with p
replaced by jw in which w is real. Note: Definitions
(A) and (B) are equivalent.

(1)(B) The ratio of the phasor equivalent of a steady-
state sine wave voltage ... to the phasor equivalent
of a steady-state sine wave current ...

(1)(C) A physical device or combination of devices
whose impedance as defined in definition (A) or (B)
can be determined. Note: This sentence illustrates
the double use of the word impedance ... Definition
(C) is a second use of 'impedance' and is independent
of definitions (A) and (B)."

The pinging experiment seems to be measuring a physical impedance (1)
(C) the nature of which is unclear. When the amplifier is outputting
power, it seems that the source impedance would be a V/I ratio (1)(B)
which doesn't respond to incoherent signals. Seems to me, you guys are
pinging something other than the source impedance.
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

Hello Cecil,

You may try to figure out how the signal injection method functions
(it is a form of active load pulling).

Can you agree with:

it doesn't matter whether:
-power reflect towards the amplifier is caused by load mismatch, or
-power is sent towards the amplifier by means of a phase synchronized
source.

This source is phase synchronized with the PA's exciter, so we have a
steady state system.

We assume small load mismatch (or low injected power towards the PA)
so that the operating point of the PA just changes slightly (to allow
linear approximation).

Now we insert a coupler between the amplifier and the load. This
coupler will measure the forward voltage generated by the PA, plus the
reflected part of the voltage that originates from the phase
synchronised source. Depending on the phase relationship, it can be
more or less then the forward voltage of the PA. If the PA shows 50
Ohms, the coupler's output would not change due to the signal
injection (as no signal is reflected by the PA).


We note the forward coupler's output voltage (both phase and
amplitude).

Now we change the phase relationship between the exciter and the
source that transmits some power toward the PA. Lets change 180
degrees and keep the amplitude the same. We again note the coupler's
output voltage (both phase and amplitude).

The voltage that is reflected by the PA equals half the complex
voltage change because of the phase change. Off course you have to
correct the readings because of the coupler loss. If you know the
signal that is send toward the PA, you can now calculate the complex
output impedance of the PA for small load change around 50 Ohms.

Instead of changing the phase of the source manually, you can do that
continuously and note the couplers output continuously. If you change
the phase of a signal continuously (with certain constant rad/s), the
result is a decrease or increase of frequency.

Assuming some reflection by the PA, the complex output from coupler
rotates around a certain point. That certain point is the result of
the PA's output power and the rotating vector is the result from the
injected signal that is reflected by the PA (back towards the load).

With a VSA you can discriminate between the voltage component from the
PA itself and the reflected component (with slightly different
frequency). With a normal spectrum analyser, you can only determine
the magnitude of the PA's reflection coefficient (or VSWR as you
like). Given the dynamic range of today's equipment, you can inject a
very low level signal that may mimic load mismatch well below VSWR =
1.1.

With respect to the impedance concept, we as amateurs do not use
steady state signals, as they contain no information. We modulate them
and are still using the impedance concept, despite the definitiones
you showed.

As long as the signal that is injected is well within the pass band of
the PA and it sufficiently low to allow linear approximation, the
concept of superposition and concept of impedance still holds. But if
you feel more confident with the manual phase change, or using two
known loads with known slight mismatch, I have nothing against it.
But if you have a VSA, some couplers and signal source at hand, it
may save lots of time.

With kind regard,

Wim
PA3DJS
www.tetech.nl

Wimpie wrote:

With a VSA you can discriminate between the voltage component from the
PA itself and the reflected component (with slightly different
frequency). With a normal spectrum analyser, you can only determine
the magnitude of the PA's reflection coefficient (or VSWR as you
like). Given the dynamic range of today's equipment, you can inject a
very low level signal that may mimic load mismatch well below VSWR =
1.1.



One needs to have an analyzer with a narrow band detector for this, though.

Inexpensive analyzers like the TAPR VNA have untuned detectors, so the
PA's main signal will screw things up.

The N2PK uses a form of direct conversion detector, so your test signal
would have to be far enough away from the main signal so that the LPF in
the detector filters it out. The original N2PK design uses, I think, a
100 Hz filter, and the adc samples at 15.36 kHz with a digital filter.
The overall BW is something like 5 Hz, so putting your test signal 1kHz
away would probably work.

On 4/27/2011 6:33 PM, Wimpie wrote:
it doesn't matter whether:
-power reflect towards the amplifier is caused by load mismatch, or
-power is sent towards the amplifier by means of a phase synchronized
source.

This source is phase synchronized with the PA's exciter, so we have a
steady state system.

We assume small load mismatch (or low injected power towards the PA)
so that the operating point of the PA just changes slightly (to allow
linear approximation).

Thanks Wimpie.

It sometimes takes a bit to get me to realize this type of thing isn't
hard, it's actually simple. My old brain tries to obfuscate things from
itself sometimes.

It Calc 100. Make the difference small.

tom
K0TAR

On Apr 27, 6:33*pm, Wimpie wrote:
With respect to the impedance concept, we as amateurs do not use
steady state signals, as they contain no information. We modulate them
and are still using the impedance concept, despite the definitiones
you showed.

Trouble is, the impedance in IEEE definition (1)(B) doesn't *cause*
reflections. If the actual source impedance matches IEEE definition (1)
(B), the presumption that source impedance will *cause* a reflection
is invalid.

Walter Maxwell argues that the actual source impedance of an RF
amplifier is in reality a V/I ratio, i.e. it agrees with IEEE
definition (1)(B).
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

On 28 abr, 14:39, Cecil Moore wrote:
On Apr 27, 6:33*pm, Wimpie wrote:

With respect to the impedance concept, we as amateurs do not use
steady state signals, as they contain no information. We modulate them
and are still using the impedance concept, despite the definitiones
you showed.

Trouble is, the impedance in IEEE definition (1)(B) doesn't *cause*
reflections. If the actual source impedance matches IEEE definition (1)
(B), the presumption that source impedance will *cause* a reflection
is invalid.

Are you familiar with the concept of S-parameters where you determine
impedance by measuring of reflection coefficient?

Walter Maxwell argues that the actual source impedance of an RF
amplifier is in reality a V/I ratio, i.e. it agrees with IEEE
definition (1)(B).
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

Wim

On May 2, 5:33*am, Wimpie wrote:
Are you familiar with the concept of S-parameters where you determine
impedance by measuring of reflection coefficient?

Exactly how do you determine the s-parameters for a single-port black
box? It is my understanding that an s-parameter analysis requires an
input port and an output port to be able to measure the parameters.
Where is the input port on an RF source?

What you seem to be measuring is the effect of one or more physical
impedance discontinuities existing in an environment of interference.
Is what you are measuring the actual dynamic source impedance? If I
understand correctly what Walter Maxwell is saying is that whatever
combinations of physical impedance discontinuities from which you guys
are reflecting your test signals, it/they are *not the source
impedance* which is a V/I ratio that originates in the source. A V/I
ratio and a physical impedance discontinuity do not yield the same
reflection coefficients for a 2-port device.
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

Hello Cecil,

On 2 mayo, 14:37, Cecil Moore wrote:
On May 2, 5:33*am, Wimpie wrote:

Are you familiar with the concept of S-parameters where you determine
impedance by measuring of reflection coefficient?

Exactly how do you determine the s-parameters for a single-port black
box? It is my understanding that an s-parameter analysis requires an
input port and an output port to be able to measure the parameters.
Where is the input port on an RF source?

I used a 2 port VNA frequently for antenna measurements, with the
difference that a single port calibration takes less time than a full
two port calibration.


What you seem to be measuring is the effect of one or more physical
impedance discontinuities existing in an environment of interference.
Is what you are measuring the actual dynamic source impedance? If I
understand correctly what Walter Maxwell is saying is that whatever
combinations of physical impedance discontinuities from which you guys
are reflecting your test signals, it/they are *not the source
impedance* which is a V/I ratio that originates in the source. A V/I
ratio and a physical impedance discontinuity do not yield the same
reflection coefficients for a 2-port device.
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

I would recommend you to measure something yourselves, or put it into
a simulation. You will see that it doesn't matter whether you use a
deltaV/deltaI setup (complex values, not magnitudes) or reflection
(time varying phase of VSWR) measurement. Just try to explore other
paths and discover other insights.

It doesn't matter when you are measuring a single port device that
contains a generator in it (as long as your VNA setup is able to
distinguish between the output generated by the single port device and
the reflection towards the VNA).

There is similarity with measuring antenna impedance (single port
measurement) when close to (broadcast) stations. Your antenna is a
generator in that case. You can't use the non-coherent type of VSWR
analyzers as the detector detects the signal from the broadcast
station also. However when using a device with a coherent
(multiplying) detector you can, as the detector doesn't respond to the
output because of the (broadcast) station.

With Tom's HP89410 setup, the injected signal for S11 can be well
within the modulation bandwidth of SSB (that means well within 1 kHz
of the transmitter's carrier frequency).

When you have good understanding of diode detectors, you can even do
it without a VNA by using heterodyning and put your focus on the phase
and amplitude of the beat frequency.

With kind regards,


Wim
PA3DJS
www.tetech.nl
without abc, PM will reach me very likely

Wimpie wrote:
Hello Cecil,

On 27 abr, 20:13, Cecil Moore wrote:
On Apr 27, 10:30 am, Wimpie wrote:

Depending on the frequency resolution of your VSA, the frequency of
the injected signal can be well within 1 kHz of the carrier, so LC
filters in the PA will not distort the measurement. In case of a 100W
PA and injection of about 100 mW, the difference in wanted signal and
signal to be rejected is 30 dB (not that large).
Would any competent optical physicist suggest that it is valid to
study the conditions associated with interfering coherent light waves
inside an interferometer by introducing an incoherent light source
into the system? Why would any competent RF engineer suggest that the
system source conditions associated with interfering coherent RF waves
can be studied by introducing an incoherent test signal?

As this slightly off-carrier frequency signal behaves like a load with
very low VSWR with a cable in between that extends with constant
speed. In other words, the amplifier sees a constant VSWR, but with
changing phase. Small frequency difference results in slow phase
change of VSWR.


to the device under test, this isn't much different than a electrically
controlled line stretcher (a classic automated load pull setup... see
the stuff from Maury Microwave, for instance)

It's a very clever technique. A variant is used in antenna ranges where
you have a probe with a mismatch in it.