Jump to content

Talk:Return loss

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

I moved the discussion out of the "See Also" section to this discussion page

[edit]

The equation for return loss has a typing error. Z5 in the denominator should read Z2.Iqbal Syed 13:46, 11 November 2006 (UTC)[reply]


Other sources give the return loss = -20log(reflection coefficient).
Where the reflection coefficient = ZL - Z0 / ZL + Z0, the inverse of what is shown here.

  • The equation is correct as it is. See note 5 in the article. Also notice that the formula in the article has a positive result for the log() function and multiplies this by positive 20. The result of log(ZL - Z0 / ZL + Z0) will be negative for most cases, and when multipled by negative 20 you will get the same positive result. I believe the formula in the article is more correct and more general. Brianonn 15:57, 14 January 2007 (UTC)[reply]

Error

[edit]

(Moved from top) The first equation has an error. The coefficient in front of the log term should be 10 rather than 20.

You're right, fixed. Edgepedia (talk) 04:34, 30 October 2008 (UTC)[reply]

Not sure "sign" is worded correctly.

[edit]

A "loss" figure in dB must indeed be stated as a positive number when it is reported and described as loss in a wordy context, e.g. in specifications or in any list of discrete data items. In a calculation it must be used with a negative or positive depending on the context. An output _level_ calculation is normally approached as a gain calculation. Any loss factor in such an addictive (logarithmic) gain calculation is in fact a negative figure. What would be a proper expression to clarify this?

Spamhog (talk) 12:41, 7 October 2010 (UTC)[reply]

Arguably, if it is being treated as a gain, it should be called "return gain" rather than "return loss". However, it is unarguably true that most published papers will graph return loss as a negative number. In fact, some journals insist on this as a matter of house style. Wikipedia reports the world, it is does not correct its mistakes. The practice probably arises from s-parameters or some similar function to avoid having to include a minus sign in the matrix definition. SpinningSpark 17:08, 8 October 2010 (UTC)[reply]

broken reference

[edit]

Reference (1) is a broken link. Need new link or delete reference. Jpt61 (talk) 21:59, 4 January 2011 (UTC)[reply]

References with broken links should not be deleted, they remain the source of the information even if they cannot currently be accessed. Sometimes sites are only temporarily down, and often an archive copy of the page can be found at places such as The Internet Archive or WebCite. Even if the link cannot be immediately repaired, it should be left in place as it gives a clue to later editors who might be able to trace it. In this particular case it is almost certain that the paper is available at IEEE Explore, although I haven't looked. SpinningSpark 12:45, 5 January 2011 (UTC)[reply]

I take issue with the 2nd paragraph

[edit]

I take issue with the following section, which is poorly written, does not properly state the usage situation in the industry, and ignores what is being done (by knowledgeable practitioners in the art) to rectify matters. In doing this it seems to be trying to get us to acquiesce to ignorance of basic engineering principles:

<quote of bogus paragraph> Properly, loss quantities, when expressed in decibels, should be positive numbers.[note 1] However, return loss has historically been expressed as a negative number, and this convention is still widely found in the literature.[1] Taking the ratio of reflected to incident power results in a negative sign for return loss; RL'(dB) = log10 (Pr/Pi) where RL'(dB) is the negative of RL(dB). Caution is required when discussing increasing or decreasing return loss since these terms strictly have the opposite meaning when return loss is defined as a negative quantity. </quote of bogus paragraph>

Firstly, here's the full citation and abstract for ref [1]: Trevor Bird, IEEE Antennas and Propagation Magazine, April 2009, 51:2, pp. 166-167. Presumably provided as a (obviously much needed) public service, the full text of this article is here: http://ieeeaps.org/aps_trans/docs/ReturnLossAPMag_09.pdf

ABSTRACT

<quote> As Editor-in-Chief of the Transactions, I have noticed over the past year or so that the occasional incorrect use of the term return loss has now grown into a flood of misuse. Perhaps over 30% of all antenna papers submitted to the Transactions in the past twelve months have used return loss incorrectly. The reason for this is uncertain. To remind everyone of the correct terminology, I review the definition of return loss, briefly outline the history of the term, and give some examples of current misuse. </quote>

Clearly the editor-in-chief of the IEEE Transactions is disapproving of the "flood of misuse" of the sign of RL. While he does not say so, the implication is that authors of misusive (c.f. abusive) papers are sent back to their word processors to edit the signs. EE Communications Systems teaches the correct signs for dB values since only that way can calculations like path loss and link loss reach the correct result. If -20 dBm experiences a path loss of 6 dB and we use an 8 dB gain amplifier, the received power is (-20) + (-6) + (+8) = -18 dBm. In such calculations, whenever we see the word loss, we apply a - sign, and whenever we see the word gain, we apply a + sign.

The word "loss" is, in effect, a negative sign, just as the word "gain" is associated with a positive sign. Also, we learn in Algebra 101 that +x+=+, +x-=-, -x+=-, and -x-=+. Thus a gain of -6.7 dB is a loss of 6.7 dB, and is better stated as such, since doubling up on signs obscures the meaning, rather like the double negative in English... Furthermore, in English we don't say "I made a deposit of -$6 to my bank account." or "I withdrew -$6." Thus, the reasons for stating a gain of -6.7 dB as a loss of 6.7 dB should be clear, and indeed is an issue of clarity.

Therefore is someone tells me they have a RL of -9 dB, red flags go up...are they really talking about a passive reflection? Credibility goes down with simple errors such as signs.

The definition

RL = log10 (Pincident/Preflected) dB

always gives a positive value for *passive* reflectors (such as transmission line, antenna, or termination mismatches).

Equivalent definitions of return loss are:

RL = 20 log10 | (VSWR + 1)/(VSWR - 1)| dB

RL = (20 log10 e) tanh^-1 |1/VSWR| dB

To quote Mr. Bird: <quote> Frequently, however, [return loss] is confused with reflection coefficient that has been expressed in dB. The logarithm is taken of the magnitude of the reflection coefficient but this is incorrectly referred to as return loss; the result is still reflection coefficient albeit in decibels. The difference between the two is a minus sign...

Many recent microwave and antenna papers and several well known books carelessly use return loss. I won’t name any for fear of embarrassing the authors. Suffice to say it is common-place to see plots captioned and labelled return loss when in fact they are really describing reflection coefficient. I have even had some reviewers asking authors to change the correct form to the incorrect one ie. change a positive sign to a negative one even though the authors labelled the plot correctly as return loss! </quote>

Trevor Bird finished his white paper with the following strong statement (containing my emphasis):

<quote> More broadly and beyond return loss, correct use of technical terms is vital for promoting consistency and avoiding misunderstanding. Through our publications, and the Transactions in particular, the AP Society strives for the best possible publications and, therefore, it is vital that authors aim for accuracy and consistency. Next time you submit a paper, please carefully check your usage of return loss and reflection coefficient; misuse of these terms may delay publication of your paper. </quote>

Wrong...

[edit]

Can someone fix in the "Sign" paragraph : "...return loss is the difference in dB between the incident power sent towards the Device Under Test (DUT) and the power reflected..."  ? ----> The formula below that text is a ratio, not a difference. MB75.227.120.151 (talk) 17:39, 17 August 2014 (UTC)[reply]

No it isn't. The formula is 10log of a ratio. Which is the difference of two 10log quantities. Which is the difference in dBs. SpinningSpark 18:38, 17 August 2014 (UTC)[reply]

Electical

[edit]

This recent edit expresses reflection coefficient (RC) in terms of 'impedance towards the load' and 'impedance towards the source', whereas the Wikki entry on RC uses the termination impedance (which can be taken to be 'the impedance towards the load') and the characteristic impedance of the line. I am not convinced that in the context of transmission lines, the former expression is more useful than the latter, and propose that this paragraph be replaced with one which uses the expression for RC given in the Wikki entry on Reflection Coefficient.

"When the source and load impedances are known values, the reflection coefficient is given by

{\displaystyle {\mathit {\Gamma }}={{Z_{\mathrm {L} }-{\bar {Z}}_{\mathrm {S} }} \over {Z_{\mathrm {L} }+Z_{\mathrm {S} }}}}{\displaystyle {\mathit {\Gamma }}={{Z_{\mathrm {L} }-{\bar {Z}}_{\mathrm {S} }} \over {Z_{\mathrm {L} }+Z_{\mathrm {S} }}}} where ZS is the impedance toward the source and ZL is the impedance toward the load. Note that this formulation for \Gamma is inconsistent with the formulation expressed in reflection coefficient."


I have now deleted the passage quoted above. It refers to the calculation of reflection coefficient, and is not needed here. — Preceding unsigned comment added by G4oep (talkcontribs) 18:56, 5 January 2022 (UTC)[reply]