Electronics > Projects, Designs, and Technical Stuff

RF attenuation not happeneing as expected

(1/2) > >>


I've got a transmitter and RF power detector with extremely poor antenna's, and using them to validate RF shielding.

I have access to sig gens, frequency counters, vector network analysers, a spectrum analyser but these are obviously very expensive bits of kit, and not something that leave the rack very often. I wanted to build something for low cost to sit in a "dirty" environment and work as a go/no-go.

When I put the DUT in between the two antennas, instead of seeing an attenuation, I actually see a gain. Obviously the DUT isnt amplifying the signal, but more likely coupling the signal across.

My question is why can I put the DUT inbetween 2 antenna's on a VNA and see the attenuation of S21, but on my jig I see the opposite?


Obviously your test jig is measuring something different than the VNA.  Since you don't tell us what your test jig measures (or is supposed to measure) it is difficult to say.  You might also explain what your DUT is.  Since you are measuring attenuation between two antennae, I guess it is some sort of shielding material?


The set-up is to validate RF shielding of different materials.

So I have two set-ups.

1) 2 port VNA with each port have 2 wires (approx 24AWG) with a 50ohm resistor at the end (1 wire connected to the signal, 1 to the ground). Set the VNA to measure S21. Insert the DUT between the two antennae and observe the relative drop in measured power.

2) Test jig is a transmitter and power detector. Each antenna is approx 3.2x48mm copper clad FR4 (with only one side having copper on). Measure the received power without a DUT inserted between the antennae, and obverse the relative drop.

With (1) I can see the drop in received power when inserting the DUT - it drops by over 10dB. However, when I use (2) the received power seems to increase. 

As far as I can tell, the VNA and test jig are performing the same task in a similar method. In fact, if I set the VNA to sweep at a spot frequency (instead of from min-max freq), then the method would be almost identical.

The only difference I can see is the type of antenna I am using - but I wouldnt have thought that would be hugely important as I am not concerned about transmitting far distances or having a particularly well-matched antenna.

*The antenna separation is ~11mm.

What frequency are you using?  Unless your answer is 30 GHz, this is a near-field problem.  Your two wires and a resistor look like a loop.  They generate a magnetic field.  In essence, you have made a transformer.  When you put your metallic/magnetic shielding material between the two plates, you induce a current that screens the excitation. On the other hand, your PCB "antenna" looks like a capacitor plate, and the energy is transferred through the electric field.  In this case, a non-grounded object put between them increases the dielectric constant, and increases the mutual capacitance between the two plates, leading to more signal transfer.  Connect your screen to ground, and it will properly attenuate the signal.

Obviously your screen isn't amplifying the signal.  If you measure S11 you would see that it drops by at least as much as S21 increases, although the change may not be resolvable.  Your object is merely decreasing the impedance mismatch.


Frequency is <1GHz.
I hadnt though of it like that.  So if I change the antenna on my jig to a loop then they would behave in a more similar manner, and I should see the attenuation as expected.

Thanks for your help, and I'll let you know what happens.


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod