Best Chamfer for microstrip discontinuities

[g86]

chamfered microstrip bend

How to get best chamfering in microstrip lines (in terms of chamfering factor or angle) for:

1. 45, 90 and 135 degree bends
2. single step and double step
3. T-junctions

What are the merits and de-merits of these methods?
How to calculate the exact path length in such chamfers?


:!: :idea: :?:

 

[flatulent]

45 degree microstrip bend

It falls in two categories.

1. Routing requirement. The line is restricted on where it can run and it has to turn a corner to get to the destination. 90 degree bends with the outer corner cut off can totally eliminate reflections. The 135 degree bend comes from digital design where the layout programs do not cut the corners of bends off. This 135 bend has a low enough reflection for digital signals that it is acceptable. The 45 internal angle bend is totally unacceptable and should be avoided.

2. Electrical requirements. The step in impedance is intentional for filtering purposes or impedance transforming purposes. The step can be modeled as a shunt reactance. Your simulator program will take this into account. T junctions also fall into the intentional discontinuity area. Adding extra triangles of metal will only make things worse because it add extra unknowns to your design.

 

[g86]

45 degree microstrip bend chanfer

Oops, I am expecting exact numerical values . These theoretical basics I know

:!: :idea: :?:

It falls in two categories.

1. Routing requirement. The line is restricted on where it can run and it has to turn a corner to get to the destination. 90 degree bends with the outer corner cut off can totally eliminate reflections. The 135 degree bend comes from digital design where the layout programs do not cut the corners of bends off. This 135 bend has a low enough reflection for digital signals that it is acceptable. The 45 internal angle bend is totally unacceptable and should be avoided.

2. Electrical requirements. The step in impedance is intentional for filtering purposes or impedance transforming purposes. The step can be modeled as a shunt reactance. Your simulator program will take this into account. T junctions also fall into the intentional discontinuity area. Adding extra triangles of metal will only make things worse because it add extra unknowns to your design.

 

[g86]

microstrip bend

Yesterday I was trying to optimize 90 degree chamfer and found 0.55 chamfering is much better than 0.5. I have used IE3D as simulator.

:!: :idea: :?:

 

[flatulent]

microstrip chamfer

Be careful with the results of simulators. There are many variables that come into play such as the thickness of the metal and the angle of the edges. There is also the operating frequency. Simulators are not accurate enough to take these items into effect. I have seen many comparisons of simulator results vs measurements for patch antennas. They were always off by a few percent in frequency and a few dB in magnitude.

 

[g86]

microstrip 45 bend

Why? I have taken care of all the parameters requred and the value nou I obtained in 0.56 which is very closer to that suggested by KC Gupta or T Edwards.
:!: :idea: :?:

Be careful with the results of simulators. There are many variables that come into play such as the thickness of the metal and the angle of the edges. There is also the operating frequency. Simulators are not accurate enough to take these items into effect. I have seen many comparisons of simulator results vs measurements for patch antennas. They were always off by a few percent in frequency and a few dB in magnitude.

 

[mwpro]

chamfer microstrip line

Curious to know how much is much better? My experience is that they doesn't differ by much.

mwpro

 

[g86]

45 bend microstrip

You are correct today I chacked with some more materials as your suggestion and found the factor depends upon the material too. Now it is too complecated. Means I could not end up with a solid answer. And for some cases 0.5 is "BEST Choice" for 90 degree which includes air. But I could not get the answer of exact path length. That also varies...nothing is stable.

Later I will try with other chamfers. Thanks all of you.

:!: :idea: :?:

Be careful with the results of simulators. There are many variables that come into play such as the thickness of the metal and the angle of the edges. There is also the operating frequency. Simulators are not accurate enough to take these items into effect. I have seen many comparisons of simulator results vs measurements for patch antennas. They were always off by a few percent in frequency and a few dB in magnitude.