Hi everyone,
I've simulated a microstrip device using HFSS with RO4350.The ROGERS data sheet says the Er(relative dielectric constant)is 3.48.But the suggest Er in the software is 3.66.CST also says 3.66 is recommended but 3.48 is for specification.
After fabrication of my design,it turns out that the simulating with Er=3.48 results in a lower centre frequency(40MHz below 2.4GHz).Back to simulation,I change the substrate Er to 3.66,the result agrees well with the measurement.
One of the devices ,produced when I operate the a engraving machine incorrectly,suffered a greater erase depth.So the substrate is thinner. I assume that the effective Er is lower than 3.48.This one is right on my required centre frequency.I begin to find out that the value 3.66 is correct. I made a mistake.
Usually,this difference can be ignored.However in my work here a different Er comes to a different centre frequency.
So I get confused.I know that I should simulate using the recommended value 3.66.But I just want to know why.Why I should use 3.66 rather than 3.48.Does anybody know?please help me.
Hi,
I had a look into the Rogers datasheet: a bit difficult to read - you have to interpret...
What I understand:
er = 3.48 is what they tried to design (the word 'process' is hard to understand in this context, but I am quite sure)
er = 3.66 is an average value - the outcome from the laminates actually produced. (See the footnote. Here it is quite clear.)
SungWon is right regarding the frequency dependency, but both characterizations in the Rogers datasheet are roughly within the same band. This will not make for the difference. And yes, on request Rogers is able to provide more data.
The main reason is that the substrates are made up of composite materials which means they are not isotropic, so you cannot describe their relative permeability in an accurate way by a single scalar number.
The standard test method Rogers and other substrate manufacturers use for characterization of their products is “X-band clamped stripline resonator” test. The field distribution in this test is different from field distribution in microstrip circuits, so if you are designing a microstrip circuit using the relative permeability obtained from clamped stripline resonator (process Dk) you will end up getting the wrong results.
To solve the problem, Rogers uses another test method for measuring relative permeability (microstrip differential phase length method) and publishes the result as design Dk. Using this number as the relative permeability in designing microstrip circuits results in more accurate results.
As SungWon Chung said, the permittivity is frequency dependent. Make sure the quoted permittivity from the manufacturer is at the frequency you are measuring at. One way you could check is obtain a sample of the substrate and use a network analyzer to measure the S-parameters of the substrate. From these parameters and the thickness of the material you can determine the permittivity in your frequency range.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence