Talk:Mismatch loss

Talk:

The page describes what better is known as "Return Loss". Reflections only occur, where there are waves involved, that are reflected, i.e. meaning returning on a transmission line or waveguide in opposite direction. (Other than using that ambigously confusing misnomer "Mismatch Loss" , the page does make a correct difference, though. It is very valuable. Don't delete it).

So the title should be changed. That includes Links to that title, which unfortunately isn't that easily done. Alternatively, both possible meanings should be described inambigously and a well distinguishing subtitle for each meaning should be used to stop omnipresent confusion.

Reason: Though a transmission line has a characteristic line "Impedance" and can be terminated in a "mismatched" manner, the result on the line is "Return Loss", but should not be synonymed "Impedance Mismatch Loss".

Correctly named "Impedance Mismatch Loss" is caused by terminating a complex impedance source (or a Thévenin equivalent toward the source at a circuit point) with a complex impedance load that is not conjugate complex with respect to the source impedance.

In impedance mismatch there is no reflection involved. Current is unreflected. It only flows through source and load identically and in one way only.

To distinguish these two is necessary, as an old school of teaching claimed that both "in a broader sense" share the same definition equation [1].

That old school of teaching statement was never proven, however. But the contrary can be proven by derivation.

The correct Return Loss equation includes the "Reflection Coefficient" RC, often called Gamma = (Z2 - Z1)/(Z2 + Z1), that is good on Transmission lines. (Their characteristic line impedance by physical limits always is nearly, but not totally, real only. Zo = 50 -j2 Ohms would be realistic, but not Zo = 50 + j100 Ohm)

Impedance mismatch, however, can include any impedances without any limiting nature.

The Impedance Mismatch equation includes the "Impedance Mismatch Coefficient" IMC = (Z2 -Z1*) / (Z2 + Z1) with Z1* meaning conjugate complex of Z1.

To see the misleading result, simply enter in Trevor. S. Bird's definition Z1 = Rs + jXs = 50 + j100 Ohm and Z2 = 100 - j100 Ohm (Resonance compensation of X1 and X2) and see the calculation's result: Negative "Return Loss" (old school equation) in spite of the article's main subject that Return Loss cannot be negative : (false) Return Loss = - 2,76 dB.

If, however, the IMC is used implicitely instead of the implicit reflection coefficient, we get the correct result: Return Loss = + 9,54 dB

("IMC" is not common so far. Please suggest a short and distinguished name instead, if you have a better one. It is not enough to ambigously call both "Gamma".)

Unfortunately even ATIS standardized the "in a broader sense equal" error of that old school of teaching in their glossary for "Reflection Coefficient". [2]

[1] Trevor S. Bird, former IEEE chief editor: „Definition and Misuse of Return Loss“ , IEEE Antennas & Propagation Magazine , Bd.51 , Iss.2, S. 166–167, April 2009.

   The PDF can be found here: https://www.qsl.net/ve2pid/ReturnLossTrevor.pdf

[2] https://glossary.atis.org/glossary/reflection-coefficient-rc/?char=R&page_number=all&sort=ASC — Preceding unsigned comment added by 2001:16B8:2DD3:3300:C9ED:A687:FFC:680D (talk) 11:48, 18 January 2022 (UTC)

According to Walter Maxwell[2] mismatch does not result in any loss ("wasted" signal), except through the transmission line. This is because the signal reflected from the load is transmitted back to the source, where it is re-reflected due to the reactive impedance presented by the source, back to the load, until all of the signal's power is emitted or absorbed by the load.

This is dubious--consider a source->TL (w/Z0)->load where the source is matched to Z0 and the load is mismatched. The wave reflected by the load is fully absorbed by the source due to the perfect source match (equivalently, the source produces less power). It would be wrong to say that the source re-reflects the power in this example. Furthermore, from the perspective of available power from the source, signal is "wasted".

Looking deeper, this seems to be a misinterpretation of the cited textbook. In section 1.2:

This is because all power reflected from the feed-line-to-antenna mismatch that reaches the input source is conserved, not dissipated. The power is returned to the antenna by re-reflection in the antenna tuner (transmatch) at the line input. On the other hand, although the loss from reflections and high SWR is not zero, this additional loss is negligible because of the low attenuation of openwire lines. If the line were lossless (zero attenuation), no loss whatsoever would result because of reflections. (This is discussed further in Chapter 6, in connection with Fig 6-1.)

Maxwell's argument is that non-unity SWR is ok at the antenna feedpoint because it's possible to perform the proper matching at the source (e.g. matching to the Zin of the TL, not the Z0). He is not arguing that load mismatch in the general sense is always re-reflected by the source.

MyIpIsLocalhost (talk) 01:58, 14 August 2025 (UTC)

Agreed (and even then, it may be easier to get a broad-band match with a matching circuit at the mismatch, rather than at the other end of the cable). catslash (talk) 13:41, 14 August 2025 (UTC)

I've edited the first paragraph for precision while trying to keep the point.

It seems like the first paragraph now slightly disagree with the presented equations, which imply that mismatch loss is always nonzero when a termination impedance is different than that of the Z0 of the TL. It could be useful to discuss a case similar to "Mismatch Loss When Both Ports are Mismatched" similarly to this All About Circuits (Although, I need to find a copy of the referenced textbook--I would have expected that total loss depends on the electrical length of the TL, a-la quarter wave transformer).

microwaves101 also uses the simple 1-Γ² expression, however. Maybe we need to clarify whether "mismatch loss" actually maps to "real" loss or not.

MyIpIsLocalhost (talk) 06:51, 19 September 2025 (UTC)