Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

CPW tapering, how to do?

Not open for further replies.


Member level 4
Oct 21, 2005
Reaction score
Trophy points
Activity points
Dear colleagues

Please pardon me for a lack of education, but I can not figure out how to realize a tapered transition between completely different coplanar waveguide circuits with different input impedances.

Specifically, I have a ~40 Ohm CPW input on high-ε ceramic substrate (ε>10). And a 50Ohm coax connector (SMA or K) which I attach directly to a quarter-wave CPW transformer on polymer substrate with ε=2.6.
The CPW sections should be attached one against another with tinned metal layers coming in contact (and soldered in this position).
The problem is that central widths (w) and gaps (g) of the waveguides are completely different because of different ε. Even when w_polymer < (w_cerr +2*g_cerr) to avoid short circuit, the reflection is awful and the overall input impedance frequency profile has a saw-shape ( ReZ↑ /\/\/\/\/\ ->Freq )
I tried to smooth the ends like exponential or Klopfenstein transformer. The impedance got a bit smoother, The S11 just a bit better, but not enough. I tried to play with the length of the smooth transition, but no luck. It is not an impedance transition, because ending point is matched geometrically, but with different ε. Therefore making the transition longer makes results worse.

So experimentally I got 3 criteria - 1) the junction is better to have close w and g on both sides (with high end low ε) 2)There should be a taper at the transformer side to match criterion 1. 3) It should be much shorter then λ/4 to avoid transforming to a wrong impedance.

The problem it that I have no idea how to design the length and shape of that taper for the best performance. My current goal is to get a 1GHz band at 10GHz with S11 below -15 dB (at least steady below -10 db).

Can you please explain me how to make such matching, or tell me a book or paper with a solution to this problem?
Last edited:

smth like this: Capture.GIF

Have you tried a really short taper section, about the size of your overlap in the picture?
Then, you can check if the remaining discontinuity is inductive or capacitive, and start compensating.

Not open for further replies.

Part and Inventory Search

Welcome to