0
\$\begingroup\$

Suppose I have a transmission line of length \$L\$ with characteristic impedance \$Z_0\$ and load voltage \$Z_L\$. If it is connected to a ideal sinusoidal voltage source \$V_s\$, then the voltage across the load \$V_L=V_s \frac{2Z_L}{Z_L+Z_0}\$. However, I think that when L becomes very small (compared to wavelength of the source) , the line essentially acts as a short and I should get \$V_L=V_s\$ across the load. I do not understand how this approximation can be obtained by reducing the length since \$V_L\$ got from transmission line model seems to be independent of the length (and of source frequency).

\$\endgroup\$
  • \$\begingroup\$ What about Zs?? \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Feb 14 '18 at 18:31
  • \$\begingroup\$ Can I not consider \$Z_s=0\$? \$\endgroup\$ – praveen kr Feb 14 '18 at 18:32
  • \$\begingroup\$ Sure then you always get twice the rated voltage for a short line unloaded \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Feb 14 '18 at 18:33
  • \$\begingroup\$ How is a lumped circuit model valid then at small lengths of wire? How can circuit theory be done with impunity at low frequencies? \$\endgroup\$ – praveen kr Feb 14 '18 at 18:39
  • 1
    \$\begingroup\$ A matched loaded generator with 1V out will always be 2V with no load for short lengths \$< \lambda /10\$ Or you start with 1V no load and when loaded at any length \$< \lambda /10\$ with 50 Ohms it will be 0.5V... same thing.. We only consider waveguides when rise time approaches prop delay or using tR=0.35/f with same criteria For lengths less than this Zo = sqrt(L/C) , the L is the dominant reactance like ground lead in 10 M probes and C is important for high impedance like probes load capacitance on coax. But we dont use Return Loss. Unless you are exceeding lambda criteria. ok! \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Feb 14 '18 at 23:19
1
\$\begingroup\$

Background: The equations derived in transmission line analysis are based off of standing waves. The idea is that the various impedance in the line and load will create reflections that will add and subtract until steady state is achieved. Therefore the equation you have used is based on wave reflection and standing wave theory. The "z" component which is often lost in these simplified forms is the distance from the load (and becomes negative) towards the source. This component is often set to 0 in order to find the voltage at the load.

Problem: The equation you have suggested comes with various assumptions you must consider. The main assumptions is that the forward and reverse propagating waves exhibit the same reflection and thus can both be interchanged via a coefficient. This allows us to use a multiplier to simply calcuate the reverse wave as a fraction of the forward (this is the same theory we use for optic reflections). This factor/coefficent is called the reflection coefficient and it is found using the Zo and Zl of the system.

$$\Gamma = {{Zl - Zo} \over {Zl + Zo}}$$

Where \Gamma is the reflection coefficient

Solution: So yes you could go back to the original derivation and use length and location to show that it can be used for what I could call zero-length-transmission-lines but we can actually use what you have to get to it.

The point is to remove the idea of "reflections" from the equation you have. The way we can remove it while still holding true to the original derivations is simply to state that Zo = Zl and thus no reflection will occur at the junction. If you substitute Zl = Zo into the reflection coefficient you will see it becomes zero and thus there is no reflection. Your equation stated above will also simplify to VL = Vs.

Considerations: I would recommend looking into the theory behind the derivation of these equations to better understand how they should/shouldn't be used. Also, often we use different models for what seem to be the same thing simply because certain models provide assumptions others don't. This can be seen in the consideration of capacitance in medium-length transmission lines that is not found in short-length transmission lines.

\$\endgroup\$
  • \$\begingroup\$ I do not understand where the assumptions start breaking down. I think the telegrapher's equation holds even as the line becomes shorter, and in addition, the load characteristic along with KCL gives me \$\frac{V^++V^-}{Z_L}=\frac{V^+-V^-}{Z_0}\$ where \$V^+\$ and \$V^-\$ respectively are incident and reflected waves. \$\endgroup\$ – praveen kr Feb 14 '18 at 21:28
  • \$\begingroup\$ Putting \$Z_L=Z_0\$ is not what I am looking for actually. In fact, I am trying to understand why we never need to bother about reflections and the characteristic impedance of the line when the length is very short. \$\endgroup\$ – praveen kr Feb 14 '18 at 21:29
  • 1
    \$\begingroup\$ For a short line, the load impedance doesn’t rotate through a large enough angle on the Smith Chart to significantly change its value. \$\endgroup\$ – Chu Feb 15 '18 at 0:50
  • \$\begingroup\$ Perhaps my wording was incorrect: you are right these equations can be used in short lengths, but we must consider that they come from a basis of reflections and standing waves. In power transmission these reflections can be ignored at distances when the line is not long enough to create significant standing waves or phase changes (as @Chu mentioned about smith charts). Even at the microchip level, we even consider looking at reflections in tiny circuits because the frequency and length of the conductor relative to the power transmitted becomes significant enough to create standing waves. \$\endgroup\$ – Commanderson Feb 15 '18 at 13:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.