What's the wrong with FR-4 at microwae frequencies?

Hi,

This is a very basic question.

What are the main reasons for using very expensive PCB substrates other FR-4 boards at microwave frequencies (specially above 4 GHz)?

I know most FR-4 boards don't have controlled dielectric constant.
However there are FR-4 boards with controlled dielectric constant.
Why they can not be used for microwave frequencies?

Could anyone tell few really good reasons for using expensive substrates for microwave circuits?

cheers,
per_lube

Loss at higher frequency will be more and also it reduces the Propagation speed of the signal.

Loss, loss, loss and loss! Oh, and sloppy permittivity control! (Feel free to reorder these issues depending on your application

The loss of FR-4 is an order of magnitude greater than your average "RF" substrates at only a few GHz, and the gap grows rapidly as you increase frequency. Whether this matters depends on your application - a PLL comprising a discrete VCO, phase comparator etc laid out on FR-4 wont care one iota, yet the loss will be devastating to a narrow-band (high-Q) coupled line resonator filter [making such structures impractical].

Even "controlled dielectic constant" FR-4 boards don't achieve the tolerances readily attained on RF substrates - whare values ~ +/-1% are commonplace. Such tight control over permittivity is essential for the accurate/repeatable/manufacturable design of resonant structures such as planar antennas, filters etc.

Re: sivamani's comment - yes, any substrate with a permittivity > 1 will "reduce the propagation speed of the signal" (over the free space value). This is not a problem, just a feature that is frequently exploited to achive miniaturisation.

It is the unpredictabiity that worries me. Circuit may work one time, next run of the board it may not work at all.

Thanks all for the good explanations...

In addition to the substrate losses, is there any problem if through holes (vias) are implemented on such a lossy substrate?

cheers,
per_lube

per_lube,

I believe not, I've seen several papers discussing EBG designs with via connection to the ground plane and working fine.

hi khaled.ayman,

if it's possible could you please provide me the info. of papers... so that I can find and read those papers

cheers,
per_lube

Here are some paper names to start with, good luck.

Design of Sievenpiper HIS for use in planar broadband antennas by means of effective medium theory --> S.K. Hampel, O. Schmitz, O. Klemp, and H. Eul

Radiation Characteristics of a Microstrip Patch Over an Electromagnetic Bandgap Surface --> Jing Liang, Student Member, IEEE, and Hung-Yu David Yang, Fellow, IEEE

Design of a Wideband Mushroom-like Electromagnetic Bandgap Structure with Magneto-Dielectric Substrate --> Tai Thanh Nguyen, Dong-Ju Kim, Seung-Han Kim, and Jae-Hyung Jang*, Member, IEEE

We use FR4 for our RF applications spanning from 100MHz up to (6-7)GHz (L,S, and C bands). Even if you were to get a consistent dielectric constant, the wavelength is so small, component edges and even the etching quality of the PCB traces begin to cause problems with the signal integrity.

So what options are available for a "good RF substrate"?
(What are they called and where do you get it?)

E:V:A said:

So what options are available for a "good RF substrate"?
(What are they called and where do you get it?)

Click to expand...

Generally a good one would be one with "low losses" tan delta should be the minimum possible, check the data sheet (or your simulators library) for this data.

You should also choose a dielectric with a specific "dielectric constant Er" according to your antenna design and parameters.

khaled.ayman said:

Generally a good one would be one with "low losses" tan delta should be the minimum possible, check the data sheet (or your simulators library) for this data.

Click to expand...

Sorry, that didn't really answer the question. I don't have any datasheets or anything. That is what I asked for. Could you provide some links?

E:V:A : Here's a list of substrates used by one of the PCB manufacturers I regularly use: **broken link removed** ... they've got datasheets against each of the listed substrates, and this list also serves as a good jump-off point into the substrate manufacturer's other offerings.
(And [free commercial plug] Lintek are brilliant too