Stripline design question

I am trying to refurbish an old KLM CS-2 polarization switcher for a circularly polarized beam antenna (KLM 435-18C) operating at approximately 435 MHz. Part of this refurbishment involves understanding the design, because certain parts are either not made anymore or have become too expensive to replace like for like, so I am in the realm of engineering in a couple of new parts. For this I need to understand the original design.

My question involves the use of a microstrip impedance transformer inside the CS-2. There is one section of strip line that is supposed to be a 1/4 wavelength 37 ohm strip line. I verified the impedance through a design program, so I know it is 37 ohm nominal impedance. However, the length of this strip line section is not 1/4 wavelength, but rather 1/8 wavelength long. I don't understand why. One clue - this section doubles back on itself in a "U" shape. Does folding a strip line back on itself create a much shortened length required for a given portion of a wavelength? It is 50% too short by my calculations. Would the fold make it act as though it were 1/4 wavelength long?

Darryl

Darryl, have you accounted for the PCB dielectric constant? For FR-4 it is about 4 and the physical length of the 1/4 wavelength section will be reduced to 1/2 of the free space value. For FR-4 the length would be 3.4 inches.

Thank you for the reply David.

I did try to account for the dielectric constant. I used an assumed Er of 3.6 and a velocity of propagation of 60%. With these constants the calculated vs measured length of a 1/2 wavelength delay line on the PCB was correct and also using this Er the calculated impedance of the delay line (50 ohms) and the impedance transformer (37 ohms) was correct. The only item that is way off is the length of the impedance transformer. As measured it is about one half of the calculated length.

This transformer is laid out on the PCB in a "U" shape with the left side 17mm long, the bottom being 14 mm and the right side also 17 mm. Total length is around 48mm. If the VoP is 60%, I calculate the 1/4 wavelength should be 172.4 mm (for 435 MHz free space) * 0.60 VoP = 103.2 mm vs the measured on the board of close to 48mm. Looks like I am off by a factor of two somewhere.

I am wondering if the geometry of the strip line being a "U" rather than a straight line would account for the difference in calculated vs measured length. I see via Google and Bing that strip line engineers often fold lines to make designs shorter and smaller. However, a factor of two seems like a lot of reduction for one simple fold with the two "legs" of the "U" 14 mm apart.

This design is off of a older commercial product (KLM CS-2 polarization switcher for a KLM 435-18C circularly polarized amateur radio antenna). I feel confident that their design is correct and my calculations are wrong. I'm just trying to learn where they are wrong. It has me puzzled.

Darryl

Hi Darryl
You aswer your own question in the description "impedance transformer"
the strip line forms a balun witch is a fraction of a 1/4 wavelength.
The length and coupling give the ratio and bandwidth!
THE calculations are not worth learning unless its your job

Cheers Bob