Is there a way to reliably measure antenna return loss outside a lab?
Assume I'm a complete beginner at RF.
Is there a way to measure return loss of an antenna, in such a manner that I can reliably reproduce the measurement later on?
What I'm talking about is producing the antenna characteristic graph to show what frequency it was adapted to and how wide it is. That is, the classic frequency vs dBm graph with a dip at the expected center frequency.
I've never quite managed to do this in a satisfying manner. I can do it in two ways, either the manual way which involves using a spectrum analyser with tracking generator and a 50ohm directional coupler. I connect the tracking generator to the input of the directional coupler, then measure how much energy that bounces back. I also have access to an antenna measurement instrument that does all of this automatically.
Using either method, I get a graph that looks somewhat correct. The antennas are typically either 433MHz or 902MHz 0 gain omnidirectional with a width of +-/50MHz from center at most.
However, if I nudge the setup and place the antenna slightly differently, or just leave it and do it again another day, the energy dip can move some +/-30MHz. I've tried to use a fixture so that the antenna sits mounted & grounded the same way every time, but still there's considerable variations.
I'm not using any signal damper, could that be a problem? Am I wrong thinking the spec should be able to deal with its own tracking generator?
Or am I naive to think I can do this accurately outside a lab? Will EMI really affect measurements that much?
2 answers
You are accessing this answer with a direct link, so it's being shown above all other answers regardless of its score. You can return to the normal view.
It sounds like you are doing the measurements right. However, I expect the problem is in the space around the antenna.
At 434 MHz, the wavelength is 690 mm or 27 inches. Everything out to about a meter or 1½ m should be considered near field, and can effect the antenna directly. This includes whatever table it is sitting on. Even with a table that's transparent to RF, it needs to be a meter or more above the floor. Usually dry light wood is a good material for such a table.
Even then the surroundings can still matter. If there is something near by, like a flat piece of metal, that reflects the radio waves, it can still matter even if it's out of the near field. This is because the reflection has a fixed phase offset from the transmitted signal.
If you have access to an anechoic chamber, try your tests there to see if they become more repeatable. If so, then you know your environment is the problem.
If you don't have access to an anechoic chamber, try an open field test. See if you can suspend the antenna several meters above the ground. If possible, put some RF-absorbing material on the ground below the antenna. Failing that, move the antenna up higher so that the reflection off the ground is more attenuated.
Local transmission, especially at ISM frequencies, are probably not much of an issue. The signal you are feeding the antenna should be many many times larger than anything the antenna might pick up from elsewhere.
The RSSI detector in the receiver is the best field tool. In Windows I had a tool (Wifi Radar?) that read the Broadcom IC RSSI and displayed a time plot of the results as I changed laptop orientation a couple of degrees and the results might change from -74 to -84 dBm and result in occasional errors at 54 Mbps.
The 2nd best tool is a 3dB splitter to measure the return peak voltage with a diode and DMM but requires calibration.
- When I was doing the same in ‘77 I found that the return loss was depending on people walking into the lab and thought I had just invented a remote intrusion alarm from the reflected waves at multiple wavelengths. This was after tuning the antenna for RL>20 dB so the directional coupler was very sensitive. It simply used a Schottky diode on the return path to a DMM to measure the reflected wave after calibration with lab equipment. You could use a stripline directional coupler or a hybrid 3 dB splitter or anything direction to measure return loss for tuning to a minimum.
You can also have a thermal sensitivity problem as well as a sensitivity to the shift in effective wavelength from reflections.
3 comment threads