[SOLVED] FR-4 permittivity during high frequency

FR-4 permittivity during high frequency (RF circuit)

Hi, I know some theory, but no practice. So I don't understand some statements.
I know that people discard the FR-4 for frequencies higher than the GHz. They say that the reason is the increase of the dielectric loss tangent (I agree) AND the change of the dielectric permittivity (I disagree).
Usually, lower permittivity (Er) of the PCB dielectric is equal to a higher speed of the electromagnetic signal through the traces. FR-4 decreases it's permittivity in frequencies over 1GHz. Isn't this good?


In practice, I would like to make a RF circuit with FR-4, which works on 1.5 GHz. If the "only" problem is the dielectric loss tangent, my conclusion may be simply that the high frequency lines need to be shorter than the others. This will be better than paying 3-4 times the price for another material, right?

The main problems related to the usage of FR-4 in high frequency board are (at least in my experience):

high loss tangent
relatively high permittivity
poor temperature stability
poor repeatability from lot to lot (different batches can have quite different permittivity)

this last, in general, doesn't allow to build reliable distribuited constant components (f.i. microstrip filters).

The change of permittivity over the frequency can have a not negligible effect sending a high speed digital signal over a microstrip (or stripline). In this case in fact, each of the various harmonics will see a different track impedance, then the received signal will be distorted with respect to the original one.

However I personally used many times FR-4 in board up to roughly 2 GHz (pure RF, not digital signals were involved) with no particular problems. Take into account that also no RF power was involved (they were receivers), as well as no distribuited constant components were used.

Second above, and also have done numerous RF low gig boards on FR4.

https://www.edaboard.com/threads/41342/