ie3d and coplanar waveguide

[sviodo]

I encountered a problem using zeland ie3d:
1. coplanar waveguide - ie3d need gnd when defining ports...
2. if i need to optimize strip width is there any chance to make optimalizatiom symmetrical so the gaps remain the same size?


--
regards,
Sviodo

 

[jian]

No. It is not necessary to have a ground plane for defining port. However, if you do not have a ground, you do need to define differential ports. For the extension ports, they need to be + and -. Yes. You can optimize symmetric structures. The scheme implemented is quite flexible. You just need to associate the change of some dimension with the change of some other dimensions. There are examples in the manual, please check it.

 

[sviodo]

shame on me

it's true that ports don't need gnd and can be differential... however i can't find anything about your second suggestion in manual... i've checked the website and found interactive movie about optimalization - nothing about symmetry though :|

--
Sviodo

 

[loucy]

If there is no ground, how is the +/- voltage being measured? the Ground is in infinity?

 

[jian]

The new manual may not have the example of optim of CPW. We may use other example. However, the manual of some old version does have an example of how to optimize the CPW gap for the 50-ohms. However, the concept is the same as the new example of optim on the IE3D 10.0 manual. Please look at the example for optimizing a hairpin filter in Section 10.9. It is a filter in symmetric shape. It is optimized while keeping the symmetry. The concept will be the same for optimizing CPW with symmetry. In fact, it will be simpler.

 

[jian]

For a CPW with 2 traces on the sides as finite ground, you should define the center conductor as + port and the 2 side traces as - ports. Then, IE3D will calculate the voltage between the + port and the 2 negative ports. It is differential.

 

[sviodo]

thanks for your help - i've found manually how to proceed symetrical optimalization
and i have to admit that ie3d is really flexible software especially coz this is only 2.5D simulator :>

thanks again and have a good day all of you!

--
Sviodo

 

[jian]

Thanks! IE3D is more than a 2.5D EM simulator. As the IE3D 10.0 is, it can handle 3D metallic such as wirebonds and thick traces in RFIC very well. It is even better than full 3D EM simulators. Also, for the IE3D 10.1 to be released by the end of the year, we are going to relase the finite dielectric feature. Many of the full 3D capabilities will be there.

 

[sviodo]

I have to say tha I was really amazed by the analysys speed however if i use finite gnd simulation time is similar to other em simulators :<
for example analysys of cbcpw (simple wavegiuding structure) takes 5sec if backing groundplane is infinite and takes 50 sec if backing gnd is finite :<

I was using differential ports for finite structures...

--
Regards,
Sviodo

see attached file

 

[jian]

Yes. Finite back plane one requires the meshing of the finite back plane. The number of unknowns will be much bigger. It will take longer time.

 

[eirp]

Hello, Sviodo!
Infinite ground plane can be included in green function definitions. Thus no need to mesh ground plane metal...

e.

 

[sviodo]

the only misunderstanding was that i assumed finite gnd and infinite will absorb similar ammount of time/memory
thinking about that again anyone would find this radiculous :>
anyway - ie3d is really helpful

 

[yingyang]

you do not need to define the -1 at the bottom plane....it should be automatically taken as ground in the basic parameter definition.

 

[sviodo]

are you sure that finite gnd (patch on layer 0) doesn't need -1 port to be connected?
i'll check simulation results tomorrow :> but thanks for this remark

--
Sviodo