Mysterious s21 coplanar waguide

[Guest]

Hi everyone,

I have design a coplanar waguide with this parameters : w=3.3mm, h=1.6 and gap=2mm, er = 3.66, via espacing = 2mm and via size =0.3mm.

sma ref : Cinch :142-0791-821

in HFS. CST and measure i get invalid results near 12GHz, it's look like low pass filter.

Can someone help me please

Any suggestions are welcomed.

Thank 

board_measure.s2p

[Admin 2578]

,,,hi buddy...!!!

,,, dimensions are incomplete, you can post a * CST file...!!!

,,,and that's what it is...??

"board_measure.s2p"

 

[Guest]

Thank you for your reply

Please find the cst file

,,and that's what it is...??

"board_measure.s2p"

 

that is s parameter measure  of the board

coplanar_board.cst

[Guest]

I think the problem come from :

- via espacing : 2mm is too large

- internal gnd plane : can make resonnance

- sma transition can make capacitive effects as the picture show i reduce the trace at the sma contact

[clanon 258]

What about the REST of the BOARD you know is COUPLED right...?

and there are missing components , there...a cap maybe...

Edited October 7, 2021 by clanon

[clanon 258]

OK , this is what i got (low precision so take it with a grain of salt)

tell if i'm close or far away from goal...

 

it seems there's ASSYMETRY from 2:2 (and its okay from 1:1

Edited October 7, 2021 by clanon

[Guest]

10 minutes ago, clanon said:

OK , this is what i got (low precision so take it with a grain of salt)

tell if i'm close or far away from goal...

 

it seems there's ASSYMETRY from 2:2 (and its okay from 1:1

goal_measure.s2p

Thank you for your help clanon, i send you attached the s parameters measurement, 

the goal is to get S11<-20DB and not resonnance at S21..

 

[Guest]

19 hours ago, Admin said:

,,,hi buddy...!!!

,,, dimensions are incomplete, you can post a * CST file...!!!

,,,and that's what it is...??

"board_measure.s2p"

 

 

 

goal_measure.s2p

coplanar_board.cst

[clanon 258]

28 minutes ago, a.ndiaye@ said:

the goal is to get S11<-20DB and not resonnance at S21..

 

S11 is high Q (more than okay)@13ghz

S22  (@17ghz) not so high Q 

What's the function of this MIXER...getting and S1 + S2  , S1 - S2...

Sorry it isn't clear for me...

besides the card you show on the first post seems to TAKE SIGNAL out...(and is incomplete)

Edited October 7, 2021 by clanon

[Guest]

34 minutes ago, clanon said:

S11 is high Q (more than okay)@13ghz

S22  (@17ghz) not so high Q 

What's the function of this MIXER...getting and S1 + S2  , S1 - S2...

Sorry it isn't clear for me...

besides the card you show on the first post seems to TAKE SIGNAL out...(and is incomplete)

 

This is used for switching signal, the idea is to pass a maximum power and not loss in the transmission line

but as you see, i lost signal in the insersion loss

[clanon 258]

is it a DUPLEXER...?

and WHAT about the REST of the BOARD...?

Edited October 8, 2021 by clanon

[Admin 2578]

,,,see this...

 

[Guest]

Yes, i see that but i haven't CST studio softwar to modify the design, i used hfss and i have correct the design

 

Optimizing-Test-Boards-for-50-GHz-End-Launch-Connectors.pdf

[Guest]

40 minutes ago, clanon said:

is it a DUPLEXER...?

and WHAT about the REST of the BOARD...?

No it's a RF SWITCH

and this trace line is used for calibration 

[Guest]

 I have change the design and now the gap between the connector and the pcb is 0.1mm

The result is similar to the measure but it's not enougth, i must get s11<-20dB (0GHZ to 15GHZ)

[Admin 2578]

[Guest]

 

There are internal GND plane, the electronic technician who makes the layout forgot to remove it.

there are some via in the layout

 

[Admin 2578]

[Guest]

1 minute ago, Admin said:

 

Without gap, there may be a short circuit

i put a litle gap of 0.1mm and the result is here

[Admin 2578]

,,,it must to be without gap...!!!

[Guest]

15 minutes ago, Admin said:

,,,it must to be without gap...!!!

Have you some idea to remove the gap in fabrication or special sma connector to minimize the gap?

[Admin 2578]

,,, by translation...!!!

[Guest]

7 minutes ago, Admin said:

,,, by translation...!!!

can you explain please?

 

[Admin 2578]

[Guest]

2 minutes ago, Admin said:

I have try different scenario

But physically to  mount the SMA it's will impossible to translate less than 0.1mm, and when i do this the result is not ok for at 15Ghz