Coplanar Waveguide Isolation

[Surrealist]

coplanar waveguide gap isolation

Let's say I want to design a chip with two CPWs side by side. Each CPW has a width of 2s and gap of g.

If I require excellent isolation between these two structures, is 3 * (2s + 2g) enough of a separation?

Also, what is the minimum coplanar ground width in order to use an infinite approximation?

 

[Azulykit]

sonnet coplanar waveguide

Are you strictly CPW or is there a ground under the CPW? This sounds like a good case for a parametric study and I suspect that via's may be in the mix too.

Just the thing for electromagnetic software analysis. I would use HFSS (because that is what I have) .

I am reluctant to provide guidance without knowing more and doing some study.

 

[kspalla]

Isolation depends on the substrate parameter too.
What is the target?
Use simulator to get exact value.

 

[Surrealist]

Hi guys,

Thanks for your reply. I would like to design a TRL calibration standard on a 400um sapphire substrate. There is no ground plane.

I am mainly wondering if there is a "general rule of thumb" for the design of parallel CPWs. For instance, can you guarentee at least 80dB or 90dB of isolation if the CPWs are N wavelengths apart, N * (2s + 2g) apart, or N times some other physical parameter apart?

Also, I don't have HFSS or any other program that costs money. I've tried simulating CPWs in Ansoft Designer SV, but I could never get reasonable results with comparsion to hand calcuations. I tried Sonnet Lite and simulated CPW that was within 5% of the values predicted by hand calculation, but with only 16MB of memory allocated for this free application, I couldn't simulate anything larger than a single CPW.

 

[Joe_User]

Finding an EM tool that can measure 80 to 90 dB of isolation is going to be tricky. I think Sonnet is probably one of the few EM tools that can do this. If you are running up against the 16 MB limit of Sonnet Lite, then maybe its time to cough up some money to upgrade. If you don't have the money, then have you tried submitting your circuit to their forum? The forum is monitored by the Sonnet Tech Support staff and they might be able to help you with getting the circuit to use less memory.

The first thing I would try (to reduce your memory requirement so it runs in Sonnet Lite) would be to use the Speed/Memory setting under Analysis | Setup.