# What's the best way to determine the impedance of a transmission line along its length?

Question asked by drkirkby on Sep 7, 2017
Latest reply on Sep 19, 2017 by DavidVW

I've made a piece of transmission line from metallic conductors with an air dielectric. I made it for an obscure amateur radio application. My frequency range of interest is 1.8 to 144 MHz. Here's a photo of it

The inner conductor is currently made of silver steel, as that is all I had, but I will get some beryllium copper later.

I've made the dimension such that this should be 50 Ohms, but of course there are a number of imperfections.

• To my knowledge there's no analytical solution to the impedance of such a line, so I used computer simulations to calculate the impedance.
• The N connectors at the end are not precisely 50 Ohms. Huber and Suhner quote 50 +/- 2 Ohms.
• There a discontinuity in diameter as the inner conductor joins to the N connector.
• The channel was machined at home by a friend using an old milling machine, so would not be as good as if done with a modern milling machined.
• There's no easy way to get the centre conductor dead in the middle. Moving it to one side will lower the impedance.
• The conductors are aluminum and silver steel, not the perfect conductors assumed in simulations. (I will later get some beryllium copper for the inner, and get the lot silver plated.)

Of course, I know that I could get much better performance with an airline from a verification kit, but that would be unsuitable for my application. Anyway, a plot of the VSWR shows it varies with frequency, with a maximum value of 1.1 at 671 MHz. Unfortunately it is nearly 1.1 at 144 MHz, which is one of the frequencies I want to use it at!

What's the best way to see how the impedance varies along the length of this transmission line? I assume time-domain would be the way to go, but I'm unsure of the best way to do this, or the best settings. Should I use a short on the end of the transmission line, or a 50 Ohm load as here?

The data was collected to 2 GHz on an 8753ES with the time-domain option (010), but I have a couple of other options.

• Collect to 20 GHz on an 8720D, which also has option 010. But really that's outside my range of interest, but I know higher frequencies give better spacial resolution.
• Save a Touchstone file, then import that into a PNA-X, which also has the time-domain option. For various complex reasons, I can't make the measurement on the PNA-X.

Any suggestions of the best way to look at the impedance of this, with a view to finding out where the imperfections are? Then perhaps I can improve things.

Dave