cpwg bend
and i just have another question in RF design
which will better 45 degree or circle
please the attchment picture
I recently ran across a discussion of a gcpw based rat-race that looked interesting in this reference:
"Microwave Circuit Modeling using Electromagnetic Field Simulators" Swanson, David G. et. al.
It is on my list of interesting things to examine one of these days. If you have an immediate problem to solve this may be useful. If you use it I would be interested in how it works out.
If all you want is a t-junction I do not see why that would be a problem. Pozar's book "Microwave Engineering" has a good discussion and points out what you can and can't get out of a t-junction.
I have recently used HFSS to work on a number of ms t-junctions. Other tools could be used as well. The biggest factor is probably what you have access to in the way of analysis software.
Regarding your second question about mitering. When the widths of rectangular miter are equal it is probably the same or very close to circular one. But in case when they are different circular should be better. I personally prefer circular.
About a year or two ago a designer I work with posed a question about miters and curved transmission lines relative to s11 performance. I created a number of of parametric models where I compared 90 deg sweeps in a microstrip and mitered bend following an "optimum" recommended in Ansoft's Designer package (bendo as I recall). There are also a series of tdr plots in Howe, "Stripline Circuit Design"' for various stripline packages that might be interesting.
We found that very tight radius bends did not perform as well as the miter for the particular microstrip we were modeling. It was not until the radius of the curve reached 3X to 4X trace width that the radius bend started outperforming the miter. I would be a little cautious about generalizing this to all PC transmission lines (microstrip, stripline, CPW, etc.). I would avoid simply bending a trace with a corner and no miter, at least above 1 GHz or so.
I would also be cautious about bends that are very close to each other. It takes a little distance for the fields to return to normal after a bend. Microwave Circuit Modeling using Electromagnetic Field Simulation by David Swanson et. al. discusses this.
I have seen slight differences with seemingly similar transmission lines and now just use an EM simulator for each situation. Via's and their placement and density add more variables to the stew. Where you happen to be in the spectrum also is a factor. My particular "experiment " was in the low microwave band, C and S bands.
The designer and I continue to have a lively discussion on this subject. We'll see if he notices this thread or choses to add his two cents on the matter.
I absolutely agree with Azulykit. Good and detailed explanation on the topic.