VHF Outdoor Testing

[gkuma]

Hi I have some testing of an antenna in the VHF range and would like some suggestions on how to properly measure the gain. The setup I have currently is transmitting a 180MHz signal with unknown gain of the antenna being received by a dipole with assumed gain of 2.11dBi(Probably closer to 2). This is 3 meters apart and about 6ft high. Would this be a sufficient test setup, i.e. I would be in the far field region correct?

Thanks

 

[VEBill]

...3 meters apart ... I would be in the far field region correct?

Seems a bit close. If you're outdoors, why not move further apart?

You might want to Google the "three antenna method".

 

[Higgler]

outdoor vhf measurements are tricky, often inaccurate and takes more experience than the typical indoor 2-18 ghz measurements in an anechoic chamber.

If you don't have absorber on the ground between your transmit and receive antenna that's at least 4 feet thick, your ground bounce will give you inaccuracy most likely.

I've made a Friis measurement from 50-600 Mhz between two antennas by placing alot of absorber underneath them (two layers of 2 foot walkway absorber totalling 4 foot height of absorber) in an asphalt parking lot (asphalt has 1.5 dielectric, better than concrete for minimizing reflectivity). Antennas were horizontally oriented, and hence the rf cables came down vertically and didn't interfere with the radiated energy (that's an important item for accuracy in your test too).
Results were more accurate than we expected, i.e. compared to calculated.

 

[gkuma]

Yes I see your point. 6ft high was kinda my worry, I think I should go higher. I measured two dipoles, one TX and one RX and with my setup got -7.8dBm received power. Using the Friss equation I should get around -8.5 if both were 2.11dBi.

Is there anyway to calculate a good test setup, ie distance from TX and RX antennas and height for an outdoor range? Or do I just do my best and use the Friss equation taken note of the reflections that might occur? Would 2 wavelength away be sufficient to measure in the far field?

 

[Higgler]

Can't you get rid of reflections with absorber?
If not, you have to average out reflections and use Friis.

Tie your two antennas with a spacer so they don't move relative to each other, like a foam piece. Use a Vector Network Analyzer and calibrate thru, then onto the two antennas, use averaging to measure S21 and move the antennas around relative to the ground, i.e. up/down.
The ground bounce will vector average out as you move the antenna pair, the direct radiation between your two antennas will be the result as they are tied together and always in phase.

Works well if you get enough random motion on your antenna pair. Set averages to 64 or more.

Some VNA's average on successive sweeps like Agilents, which is convenient, you can slow the sweep speed to 1/2 second and know that you've moved your antenna pair enough to get a change in phase from the ground bounce.
I think others will do 64 quick sweeps on a single freq. point (if that's the case, you need to take a bunch of single sweeps as you move the antenna pair, and vectorially average your data in the computer to average out ground bounce).

 

[gkuma]

I wouldn't have nearly enough absorber. Anywhere to purchase for cheap? Last time I checked it was fairly expensive.

My current setup I am just using a signal generator and function generator to do my measurements. This way I have the cables come down vertically like you mentioned and don't cause a bad measurement. I would need to get longer cables to use the VNA and measure S21 at 3 meters.

Simulation compared to measured gain of the AUT is within 1dBi. I think this test method is good until I decide to have it sent of to be measured in an anechoic chamber.

Thanks for your help.

 

[Higgler]

very few anechoic chambers can measure at 180 Mhz.

If you're within 1 dB of calculated gain, I think you're done.