jump to navigation

At a loss August 3, 2010

Posted by mareserinitatis in electromagnetics, engineering.
Tags: ,

A coworker (in this case, my husband) came up to me and said, “Our customer wants you to change your plots so that they have negative values instead.”

I glared at him and said, “Absolutely not.” Then I paused for a moment and added, “Well, I’ll change them, but I’m changing the labels, too.”

“Good enough.”

Two things occurred to me after this conversation.

First, I need to work on my diplomacy skills. That’s a topic for another time.

Second, I really hate it when people are ‘sloppy’ with their definitions.

The particular values I was plotting were, among others, return loss and insertion loss. If you are injecting a signal into a device, you generally want to know how much actually makes it into the device. Likewise, you’ll want to be able to tell how much of the signal actually makes it through the device and out the other side. We generally take these quantities in decibels because we can then add and subtract them rather than multiply and divide.

I won’t get into the math but instead approach this from semantics. If you have a device that’s losing signal at the input, we say that this device has a return loss. Likewise, if not all the signal that gets into the device comes out at the exit, we say it has an insertion loss. Because all devices are not perfect conductors (having, therefore, an impedance along the length of copper traces, for instance) and there are mismatches between devices, there is always a loss. To indicate this, we say that the loss should be a positive number, and for that to be true, the mathematical definition has a negative sign at the beginning. Saying you have a negative loss is like saying you have a negative debt: having a negative debt means you’re in the black, and a negative loss means your signal is being amplified. If you have a passive device, then you can’t have a negative loss.

As such, I made sure to express my results in terms of positive numbers. Our customer said this made the plots more difficult to read and compare.

I agree that it make make things a bit more difficult to read, but if I swap the sign, I have really plotted the reflection and transmission coefficients in terms of decibels. Therefore, while I am going to change the sign on my plots, I will also be changing my labels.

The problem with all of this is that some journals are very careful about this while others are not. When looking at papers dealing with certain passive devices, I’ve noticed that they indicate their return and insertion losses with negative numbers. This shouldn’t happen. With active devices, it’s more confusing. Do they really have a loss or is their signal being amplified? Usually you can’t tell unless you dig into the details of the paper. Unfortunately, common references like Wikipedia can confuse the issue further.

I’m not sure what to do about the problem other than make sure that my own stuff is straight and hope that others are doing the same.



1. Fluxor - August 3, 2010

I’m afraid your efforts to battle this “problem” is about as useful as the advice of obscure pedantic grammar prescriptivists on the proper use of apostrophes. There’s textbook English and then there’s everday English. It’s all about getting used to the culture and customary jargon of the industry.

It’s like the phrase “I couldn’t care less”. It has now morphed into “I could care less” for a big swatch of the English speaking population. So what’s right? On the face of it, these two sentences are opposite in nature, but in practice, both actually mean the same thing. When a usage with language becomes widespread enough, it becomes “correct” by default.

The same is true when talking about rejection as a negative number. When someone is talking about rejection, it is customary in my line of work to assume that there is a loss of signal, no matter what the sign of the plot is, unless otherwise notified by the author. We plot insertion loss as negative numbers. We also plot power supply rejection ratio as negative numbers. No one ever gets confused because everyone is acclimated to this norm. Some people will use positive numbers, some people will use negative numbers, and some will mix and match their signs with nary a care. *shrug*


mareserinitatis - August 4, 2010

I can’t stand it when people say, “I could care less.” That should establish what type of personality I have. 🙂

In antennas, it seems like people are pretty good about following convention, but obviously some other fields are a bit less careful. I was very surprised, however, that someone would actually ask me to plot it incorrectly for the sake of convenience.


Fluxor - August 4, 2010

I’d say it’s more than convenience; rather, it’s the type of technical language that your customer speaks. If I were to produce an instruction manual, I’d make sure that the spelling is different for the UK and US markets. I can be a hardass and say the English invented English, so we should all do just what they do. But that’s not what the US consumer expects and wants. Same goes for your customers. They want it in their language, not necessarily the language based off of a narrow definition.


mareserinitatis - August 5, 2010

But when you’re discussing technical topics, your language should be based in the mathematical definitions, not whatever pops up on your network analyzer.


2. Fluxor - August 5, 2010

I think that’s debatable. Mathematical definitions are like dictionary definitions. Great most of the time, but incomplete under some specific circumstances. Technical discussions, like regular language, isn’t immune to idiosyncrasies. As the marketing folks like to say, it’s a feature, not a bug! 🙂


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: