[Theoretical question] Reflection at the input of the amplifier
For a sinus, reduced amplitude. A square-wave will most likely be distorted. Actual length of coaxial cable can affect result a lot.
But why? Shouldn't all energy from the source be reflected back from the amplifier?
For high impedance, no. Infinite impedance, yes.
But then again, in most of the low frequency electronics, in.e. audio cirucuits, input of the amplifier is never matched to the TXLine, but we see no such effects. I know that people will say that on low frequency wavelength is to long compared to the cable length, but still, reflection can be calculated, right?
Yes there are same electrical rules and limitations for LF circuits but at least on same PCB is wavelength related problems not that big. A reason that impedance matching is less important is due to that available gain for bandwidth upper frequency is much higher for a transistor working at LF compared to RF. See GBWP, gain bandwidth product. http://en.wikipedia.org/wiki/Gain%E2...dwidth_product
Also for long cables, is the problem not same as for RF due to this. A typical LF amplifier feeding a cable is designed as voltage sources with internal impedance that is a fraction of load impedance. Due to this is, is available voltage at the load, which can be a speaker, independent of load impedance, as long as reactive and resistive losses in cable are minimal compared to load impedance.
For real long cables can also impedance matching be of importance at LF for keeping distortion low. An example are professional microphones and cables, which often are matched for 600 Ohm, as well as receiving amplifier input. For further improvement are they designed as a balanced transmission line due limitations related to grounding. Similar problem does also exist for RF.
Another example, maybe not as common these days, are the long unshielded wires hanging on telephone-poles. They are regularly impedance corrected for a certain impedance with some kilometers distance apart.
Great explanation for the second part of the post, thanks. However, for the first part: I don't buy it. How is GBWP affecting reflection? If input impedance of the amplifier is high enough (few kOhm), reflection is approximately 1, so the power is being reflected and almost nothing enters the amplifier. Therefore almost nothing should be seen at the output of the amp.
That is probably likely, a long wire in a heavy mismatched system, and distortion will occur. It is not that hard to do the calculation how much voltage that will be transferred.
Assume a typical small signal line amplifier with internal impedance of 100 Ohm connected to a typical shielded LF wire with an characteristic impedance of 200 Ohm, 10 meter, and a termination impedance of 10 kOhm.
As termination is an voltage follower and as line amplifier will shortcut any reflections will most of the voltage reach load unaffected of cable impedance. Loss will be measurable within frequencies below 20kHz, but that is all.
As total cable length is electrical short, will reflections be almost in same phase as original signal so it will not cause oscillations or ringing.
A LF line voltage amplifier is often designed with low internal impedance, its voltage output amplitude is minimal depending on its load impedance, within wide limits. One way to design an LF amplifier to have low internal impedance is to have high amount of internal feedback. Internal feedback do cost of available GBWP. A RF amplifier can not afford that amount of internal feedback, as nothing will remain for providing power gain, so its internal impedance will be an muck bigger factor of how much power that can reach the load. Due to its higher internal impedance can reflections cause instability or even oscillation. If there is a transmission line of length in between, will its impedance also be important in same way as for an telephone line. This length can cause reflections with a phase that differ very much compared to what the amplifier for the moment tries to deliver, can it cause much higher amount of reduced power gain will reach load.
Your "theoretical" amplifier is yet unspecified.
As already mentioned in your previous thread, if you specify high input impedance, I'll assume respectively very high power gain. Then there can be sufficient output signal. In other words, you are stumbling upon your unrealistic assumptions. Talking about about real transistor amplifiers would be a different thing. A "very high input impedance" is realistic at least up to a few 100 MHz.
Guys, thank you very much for your help. I think I got it somehow, I am not sure if that is like you had in mind, but it makes a sense to me:
So we have voltage source Vs with 50 Ohm internal impedance, frequency is irrelevant. We connect it to the 50 Ohm transmission line. The source sees this line as a 50 Ohm load and because of simple voltage divider it makes, we will have voltage of Vs/2 at the input of transmission line. This voltage will propagate through TXLine on which other end we have an amplifier with a high input impedance compared to 50 Ohm. Reflection will be 1 so the total voltage at this end of line will be now 2*Vs/2 = Vs. Output of the amplifier will now be gain*Vs. IMHO, this should work for any TXLine and amp system regardless of frequency or line length. Only in RF amps we will never have such high impedance, rather some complex value, and reflection will therefore be some complex number different than 1.
What do you think?
A line with characteristic impedance of 50 Ohm, is not a 50 Ohm load if line length not is infinite long.
In most cases, with low loss coaxial cable, will most part of the power either be delivered to terminated load or reflected back to transmitter, which then will be an load for its own generated power.
Edit:
Sorry, did probably misunderstand you a bit, "The source sees this line as a 50 Ohm load" is correct, until source finds out that line not is infinite long or have a perfect matched load, as no reflection then will occur.
Yes. It's a usual termination scheme ("source side termination") for DC - medium speed analog and digital signals, usable up to e.g. 100 MHz.
At higher signal frequencies, besides finite load impedance, the non-ideal 50 ohm source impedance suggest to advance to a both side termination scheme.