Microstrip vs. Grounded co-planar

[Jim cage]

Hi,

Can anyone tell me what is the difference between Microstrip and Grounded co-planar?

If you take a grounded co-planar waveguide and separate the two ground conductors (near the center trace) to infinity, why wont you get a microstrip line?

as I understand, microstrip is a trace where the ground is located below the trace.
Coplanar has a center conductor, 2 ground trace distanced at a GAP and a ground plane below.

Regards,
Jim

 

[flatulent]

The coplanar with ground plane is slightly better than ordinary microstrip. The main advantage is that you can have shunt elements to ground without the parasitics of vias. This is very important when you get into the 4+ GHz region. The US company Hittite always has coplanar with ground in their suggested layout application notes.

 

[Jim cage]

1. But why when you enter a gap distance of infinity you are not getting the characteristic impedance of a microstrip?

(use ADS(Linecalc), TXLine or Appcad and you will see that if you take co-planar with a ground plane and have a gap distance of 1000000mil for example between the center conductor and the other two you are not getting the characteristic impedance of a microstrip)


2. Another questions, when you manufactor a PCB there is always a stage where they fill out the entire CS with ground on the empty spots. this means that the microstrip is now turning to co-planar right? (near each conductor there is ground now...)


Regards,

 

[Azulykit]

There are two very useful books on various PCB transmission line forms. Simon and Waddel (I have paper copies). You might want to run them down.

Your descriptions are correct. MICROSTRIP is a thin trace over a larger (often much larger groundplane). COPLANAR WAVEGUIDE is a trace with groundplanes (or sidegrounds) on both sides and all in the same plane. GROUNDED CPW is similar, with a trace and sidegrounds (groundplanes) ....and....a groundplane under the trace and sidegrounds. Usually the sidegrounds and lower groundplane are connected with "stitching" vias. This is what makes the convenient connections that Flatulent describes.

If the gaps are increased significantly, grounded CPW will (generallly) have the same impedance as microstrip.

Just a warning, the thickness and dielectric constant play into this in a big way and one often sees unexpected impedance trends when just the gap is changed. I guess one could think of flooding the area around a trace as converting microstrip to grounded CPW. I would expect to see the addition of vias as well if that is the objective.

 

[RF-OM]

The major difference between microstrip and coplanar structures is in modes. Coplanar is better when it is real coplanar (when slot is narrower than substrate height). The reason of wrong results when you try to use AppCAD or similar program is in the quality of these programs. Simple programs usually use kind of simple numerical algorithms for calculations. Such algorithms are based on some assumptions, first of all on ratio of trace width to substrate height as well is for slot width. Such empirical approach works well only when ratios are within the suggested limits which definitely are not very wide. Many numerical approaches use simplified solution for elliptic integrals of the first and second kind. This works well for many practical scenarios, but not for very high slot width. If you want to check coplanar with very wide slot you need to use one of field solver simulators.

Best regards

 

[ADS_RFMW]

Use Agilent Momentum that gives you exact expected results

 

[flatulent]

One thing to watch out for in free or low cost transmission line programs is putting in dimensions outside the range of validity of the algebraic equations these programs use.

The advice above about using a field solver program is wise. These programs are more expensive but are more exact for any set of dimensions.