The first collinear antenna, a length of black coax taped vertically down the inside of an apartment window over an autumn view of the street

My first attempt of a Collinear Antenna. Professional mounting!

Part 1 – Checking the Velocity Factor of the Coax Cable

I have a homebuilt Collinear Antenna on one of my two ADS-B receivers. It’s an eight element antenna, built without any real effort of making sure that all the elements where the right length. So it’s about time to try to do it properly. The easy way is to get Flightaware’s great ADS-B antenna and filter, but it’s fun to build your own as well.

To build a good antenna, you would preferably use a Vector Network Analyzer so you can do proper measurements of your antenna design. With it you can quickly check stuff like SWR and optimize the impedance of the antenna at the frequency it’s intended to be used. In my case 1090MHz for ADS-B. But if you don’t have the money to spend, you have to improvise. I’m going to build a 12 element Collinear Antenna out of cheap RG-58 coax cable. But to calculate the right length of each element, we need to know the Velocity Factor, or the signal speed through the cable compared to light traveling in a vacuum. So let’s do the measurements.

The Cable

I got a long stretch of some old RG-58 coax cable from an old cash register. You could just go with the velocity factor values you find on the internet for RG-58, which the almighty Google says is 0.66. But using this value in my case would be wrong, and the calculations would end up with an antenna way out of the optimal performance, as we shall see.

The Measurement Setup

Here’s what we are going to do. Send a pulse with fast rise time (in my case a pulse with 9.5 ns rise time) from a signal generator at relatively long pulses, in my case with a Silent SDG2122X 120MHz Waveform generator (Actually a hacked SDG2042X.) The pulse goes into an oscilloscope, with a T-BNC connection. Connect the coax cable to the other BNC connector on the T-connector. The pulse first shows up as a quick rise on the oscilloscope, but the signal will travel down the unterminated coax cable and reflect back. This will double the signal when it reaches back to the oscilloscope. So using cursors, we can measure the travel time of the pulse and divide it by two to get the propagation time.

Rigol scope showing the bare 2.06 V pulse with nothing attached, its edges clean and under 9.5 ns

The pulse without the unterminated cable attached.

Ok, and let’s connect the unterminated cable. Now we see the adding of the returning signal from the coax cable.

The same pulse with the unterminated coax connected: the reflection returns partway along and steps the trace up to 2.22 V

Pulse with return echo from an unterminated coax cable.

The pulse width should be long enough because we want to be able to measure short cables. If you don’t have a signal generator with a fast rise time (the quicker, the better), you can build a pulse generator with the swift rise and fall edges of about 1.2 ns for around a dollar. Here’s a video on how to create it.

https://www.youtube.com/watch?v=9cP6w2odGUc

Here are the settings on my signal generator with 8.4ns rise time. I’ve set the output to 50 Ω impedance to match the connections at the oscilloscope.

The Siglent SDG2042X set to a 1 kHz pulse, 2 Vpp, 381.3 ns wide with an 8.4 ns rise edge into a 50 ohm load

 

Calculating the Velocity Factor of the Coax Cable

So now it’s time to use the cursors on the scope for measuring the time for the signal to go to the end of the unterminated cable and echo back.

Scope cursors bracketing the reflection at -6.2 ns and 84.8 ns, a 91 ns round trip, which gives the cable’s velocity factor

So the Δ between the start of the pulses are 91ns. The propagation time is half of that, 91÷2=45.5ns for the signal to go one way through the coax cable. The cable is 950 cm long and the distance light travels in a vacuum in 1 ns is 29.99709 cm.

950cm ÷ 45.5ns ÷ 29.99709ns = 0.696038 Velocity Factor. The factor is 0.696 times the speed of light.

Will the specified value of 0.66 and 0.696 VF make a difference when doing the calculations for the lengths of the sections of the antenna? Let’s check.

The formula for calculating the length of each element is:

λ = c / f

c = speed of light in vacuum 299 792 458 meters per second. f = Frequency, 1090MHz for ADS-B

λ = 299.792458 / 1090 = 0.2750389523 meter wavelength

Length of element = (half wavelength in meter) 0.2750389523 ÷ 2  = 137.5194761468 * (VF)0.696 * (convert to cm) 100 = 9.4888438541 cm

So the length of each of the elements should be 9.49 cm with this cable. But how much would it have differed if I used the official velocity factor from the manufacturer? Then the result would have been 9,08 which is almost 5 % off target. Will this make a difference? We’ll see when I build an antenna. You can find a handy calculator for calculating both ½ and ¼ wavelength for collinear antennas with build instructions here.

In the next installment, I’ll do the assembly of the antenna and make some comparisons against my FlightAware Antenna, and my old bodged together collinear antenna.