Preventing reflections from edge of microstrip grid

[nimmerlein]

I have a "circuit" which is a (fairly large, 25x25 cm) square grid of 1mm wide microstrip lines on a 1mm thick substrate (PTFE with epsilon of 2.2). Operation frequency is around 20 GHz.
Reflections from the edges of that grid, which are just stubs of a few mm length where the microstrips have been cut off, unfortunately have appeared to be a bit more of a problem than first anticipated.

So, now I am looking for a good way of preventing these reflections. "Good" would ideally mean that I would not have to use lumped components (I would need many of them because obviously there are a lot of stubs to be terminated which would (a) probably be pricy and (b) be a pain to solder on, especially since the ground plane of the circuit is large which makes this difficult) and even more ideally would not require to modify the circuit design (because then I could just retrofit it to the board I already got made). I was thinking along the lines of adding some sort of lossy material (foil? foam?) to the edges, tapered in some way to match it to the microstrips, but I am of course happy about any kind of suggestion.

 

[biff44]

if you do not want to add lumped lossy elements at the edges, make it a little bigger and put 3 cm wide swath of absorptive flexible media on all four edges.

something like this:

Eccosorb™ 5G MeF 1

It is very flexible and has a high service temperature (165°C). It is recommended for all mmWave applications in the 18-40 GHz. It has a UL94 V-0 rating.

www.laird.com

 

[nimmerlein]

if you do not want to add lumped lossy elements at the edges, make it a little bigger and put 3 cm wide swath of absorptive flexible media on all four edges.

something like this:

Eccosorb™ 5G MeF 1

It is very flexible and has a high service temperature (165°C). It is recommended for all mmWave applications in the 18-40 GHz. It has a UL94 V-0 rating.

www.laird.com

If the absorber is heavily loaded, will this not be quite a sharp impedance mismatch at the edge of the absorber resulting in a lot of reflected energy there?

 

[volker@muehlhaus]

Reflections from the edges of that grid, which are just stubs of a few mm length where the microstrips have been cut off, unfortunately have appeared to be a bit more of a problem than first anticipated.

How did you test/confirm that it's reflection from the edges? Your setup sounds unusual and there might be other causes for resonances.

 

[biff44]

you are entering the realm of empirical testing.
set up the antenna, and measure it. then add the absorber, and measure it again. did it fix it? if yes, good, glue it down and call it a day. if not, try different shapes for the absorber. here is a shape where, if there is a strong reflection from the edge of the absorber, the sawtooth shape would mitigate that somewhat

red is your existing antenna
gray is the added absorber material