Investigating intresting parasitic resonance phenomena

[yefj]

Hello , I have a line geting closer to the line connected to the ports.
as the distancebetween them gets closer the S21 notches get deeper and deeper.
The difference between two peaks is 3GHz.
each notch represent a resonance on the floating line.
at 6Ghz we have two wavelength
at 9GHz three wavelength
at 12GHz we see four wavelength resonances.
given only this data is there a physical way to know that the floating line is 26mm?
Thanks.

 

[Baluncore]

The 26mm strip is a dipole, resonant at: 6G162, 9G159, and 12G120.
Those are not perfect harmonics, so some end effect is in play.
There will also be a resonance at about 3.163 GHz, λ = 94.75 mm in free space.
λ/2 = 47.38 mm in free space. The velocity factor is 47.38/26 = 1.822
Effective Er will therefore be about 1.822² = 3.32 ; but that seems a bit high without a ground plane, as most of the volume about the dipole is air with Er = 1.00

given only this data is there a physical way to know that the floating line is 26mm?

Yes.
Is there a ground plane ?
What is the thickness and Er of the dielectric ?
Then calculate backwards to find the line length.

 

[yefj]

Hello Baluncore, As you can see below we have er=4.1 and ground plane.
At 3.08Ghz we have half wavelength resonance.according to this formula the wavelength is 48.1mm.
Half of wavelength is 24.05mm where in reality we have 26mm.
So basickally we take the difference between resonant notches and put it inside the Wavelength formula shown below.

another very intresting thing I would like to ask is investigating the S21 phase of this stucture.
As you can see in the phase plot we have big slopes and in the places we have resonances we have little jumps.
The slope of the little phase jump which is the group delay is 2.83ns.
How can we transfer the group delay of this sub phase jump into the length of the floating line?
from the calculation below I get 419mm Where did I go wrong?
Thanks.



S2P file is attached.

 

[Baluncore]

There may be some weird effect where you couple a resonant line of length 25 mm to the dipole of 26 mm.

There will be some electrostatic coupling across the circuit board between those two lines, but the major coupling will be through the reflection in the ground plane, which makes it a "backward wave coupler".

 

[yefj]

Hello Baluncore, In every S21 resonant notch I have a small jump in S21 phase.
I tried to convert the slope of this little phase jump into group delay and find the length of the floating line .
i got 480mm . Am I wrong in my interpresting phase strategy to find floating line length?
Is there a way to interpret this smal phase jump to find the length of floating line?
Thanks.

 

[Baluncore]

That jump in the phase looks to be due to a minor resonance in the system. I would make small changes to the model until I could identify what parameter it was most sensitive to. You may find it is related to the extent of the ground plane.

 

[yefj]

Hello Baluncore,regarding the S21 phase below.
Few questions:
1.why the slope gets steeper when the line goes longer?
2.Is there a way I could see this phenomena If I try to look in the resonance frequency electric or magnetic field.
If you could please describe How it looks ?
"through the reflection in the ground plane, which makes it a "backward wave coupler"."

 

[Baluncore]

Maybe it is time for you to study odd and even mode couplers.
The pdf is safe.
http://www.ittc.ku.edu/~jstiles/723/handouts/section_7_6_Coupled_Line_Couplers_package.pdf

 

[yefj]

Yes I will read that .
regarding the other question could you please explain why longer lines creates steaper phase slopes in S21?

 

[Baluncore]

Longer lines have lower frequencies, so they do more turns around the Smith chart per GHz.