On Mar 24, 10:02 am, Cecil Moore wrote:
Keith Dysart wrote:
On Mar 24, 8:45 am, Cecil Moore wrote:
I made the rough computation of -25 dB before I responded.
It is below the ability of the human eye to detect TV
ghosting.

Would you care to reveal the details of the calculations
so that we can locate the error(s) that lead to a non-zero
answer?

As I said, it was a rough computation based on your 450
ohm series padding resistor and a lot of simplified
assumptions about the configuration of the system.

In order to avoid ambiguity, would you please give us
better specifications for the source? Ramo and Whinnery
warn us not to draw any conclusions about what happens
inside a Thevenin equivalent source so your specification
must be an achievable well-defined real-world source.
How about a class-A linear amplifier with no filters,
no impedance transformation, and no protection circuitry?
Vrms/Irms = 75 ohms Vrms*Irms*cos(0) = 1 watt

The amount of interference cannot be calculated without
magnitude and phase. Shall we assume that the 450 ohm
line is an integer number of wavelengths and lossless?
i.e. Z0-matched to 75 ohms?

If you require detailed calculations, more specified
information is needed.

Standard analysis can tell us quite a bit about the situation
without the need for further information. The information
we have:
- Generator with 450 Ohm output impedance
- connected directly to a 1000 foot line with 450 Ohm
characteristic impedance
- connected to a load with 75 Ohm input impedance.

Remembering that reflection coefficient is RC = (Z2-Z1)/(Z2+Z1) let
use compute the RCs at all the connection points:
RCgen-line = (450-450)/(450+450) - 0
RCline-load = (75-450)/(75+450) - -0.714
RCline-gen = (450-450)/(450+450) - 0

Let the output of the generator into a 450 Ohm load be Vout.

Forward voltage supplied by the generator into the line:
Vfwd = Vout
This is also the Vfwd that reaches the load (for the purposes of
example we are dealing with lossless lines).

The reflected voltage at the load:
Vref = Vfwd * RCline-load = Vfwd * -0.714 = Vout * -0.714

The voltage applied to the load:
Vload = Vfwd + Vref = Vout * 0.286

The voltage reflected from the load that arrives back at the
generator:
Vref = Vout * -0.714

The amount of Vref reaching the generator that is reflected back
to the load:
Vrefref = RCline-gen * Vref = 0 * Vref = 0

There for no ghosts since there is no re-reflection.

A few other things of possible interest...
Power delivered to the load:
Pload = Vload * Vload / 75 = Vout * Vout * 0.0816 / 75
= Vout * Vout * .0011
This is the actual power that flows in the line.

Forwarwd power as indicated by a directional wattmeter in the line:
Pfwd = Vfwd * Vfwd / 450
Reverse power as indicated by a directional wattmeter in the line:
Prev = Vref * Vref / 450

Net power flowing in the line:
Pfwd - Prev = (Vfwd*Vfwd)/450 - (Vref*Vref)/450
= Vout*Vout/450 - Vout*RC*Vout*RC/450
= Vout*Vout*(1-(RC*RC))/450
= Vout * Vout * 0.0011
Which is the same as the load power; this being as expected
since directional wattmeters are useful for obtaining the
net power.

It is also useful to know what we can not compute from the
provided information....

The voltage at the generator output and its phase with respect
to the current since this is dependant on the velocity factor
of the line and the signal frequency. Though if the output is DC
(e.g. a step function) than this can be easily computing for
the situation after one round trip delay.

Generator dissipation since we would need to know the internal
construction of the generator and the actual voltage and current
phase at the output of the generator.

-----

Still, using conventional techniques we have derived quite a
bit about the system without knowing the internals of the generator.

Cecil, are you saying that your techniques can not produce any
derivations without knowing the internal construction of the
generator and the length of the line (in wavelengths)?

....Keith