FR4 specificatins at 5.9GHz

Hi,

I would like to know the dielectric constant and the loss tangent values for FR4 at 5.9GHz.

Do these parameters vary with the material thickness?

Thanks,

FR4 is a Fire Retardant version 4 Specification for PWB Dielectrics that often go thru wave soldering with flux and make huge fires in a factory.

You are concerned about is estimating the Dielectric Constant for Epoxy and Fiberglass at a microwave frequency that depends on the mix ratio of glass to resin and the and the loss tangents are especially bad > 1GHz. You can get guidelines from a Vendor or Board Fab shop but generally it is an "uncontrolled " parameter. You must include a test coupon and specify an impedance and pay the shop to control that impedance in order to guarantee 5 or 10% tolerance. Beyond that you take your chances. There is no constant form vendor to vendor but from my experience over 10yrs ago, we used GE's GETEK material since it had lower loss tangent than "std FR4" and the epsilon value drops with increasing frequency. I suspect it may from 4 to 2 relative permittivity going from 1 to 6GHz but your research may prove otherwise. Most people will choose ceramic substrate for this range or use Polyamide or more expensive , Teflon. My preference is Ceramic and many shops make these custom hybrid MMIC's.

Check out current state of art materials in low cost. but expect trade-offs between cost and stability, loss tangent, temp and mechanical stability.

Yes Thicker Dielectric means more LOSS but constant Permitvity. However this tolerance is 10% at best and shops will shrink artwork or expand artwork track sizes to get the controlled impedance you desire in your spec. for a fee.$ which depends on volume.. If you cant afford then experiment and expect to trim your antenna or whatever.

I tried FR4 at 5.8 ghz recently, and it was a total failure. I would recommend against it! RO4350 or Megtron6 would be much better.

Attenuation = 2.3 ⋅ f ⋅ tan ( δ ) √εr [dB/ inch]

Where:
(1) f is the frequency in GHz
(2) tan(δ) is the dimensionless loss tangent
(3) εr is the relative dielectric constant of the materialariables)

Unless you know the the Loss Tangent when working well above test Frequency e.g. 1GHz Do not try until you know :shock:

N.B. Do not guess, find it or ask supplier tech support for test datasheet or ask PWB shop for info on;

tan(δ) vs f vs thickness

N.B. Please note :shock: this next chart for cost vs . δ

See future changes in RF dielectric costs. I am not surprised PU & resin based materials going down but very surprised at teflon or PTFE materials going up. You need to shop around. I aso have used Ceramic substrate and is very good as hybrid laminate as thin ceramic is brittle.

I would look further into Panasonic & Hitachi products. I have used GETEK @1GHz only. but not GETEK II :| GETEK was known to be 30% more than std FR4 in 1995.

You have many options today.

This design is around 4 years old.. and is patented now. I suggest you do more research.

As I know, the FR4 is an anisotropic material, which means the permittivity is different in different direction. So does the wave propagation mode.

FR4 is difficult to be used in microwave area. 5.9 GHz is already out of the range.

I know FR4 is difficult to be used in microwave area.

However in the board I am going to develop for WLAN 802.11p communications, all rf components (a transceiver MAX2828, a Power amplifier, a LNA and a RF Switch.) are 50ohms I/O matched, then I won’t have any matching network with, I only need some transmission lines, as short as possible (low insertion loss is expected, even with FR4), to connect them.

I think FR4 is a good option for this type of application. Am I right?

If you have to do so, Coplanar Waveguide (CPW) probably would be better instead of microstrip structure.

If you are running narrowband and don't mind a lossy dielectric your transceiver design can be made to work. But if sensitivity and group delay distortion in broadband is important then you will want to avoid FR4 and use something more compatible at 6GHz. If your dielectric constant changes with frequency you may get phase noise from a non-flat group delay. If it absorbs energy from loss tangent then you suffer on transmission loss and return loss and hence sensitivity.



Ideal may be quartz over alumina or ceramic substrate but you did not specify budget constraints. There are many poly plastic substrates being developed now. See HItachi, Panasonic etc

If you are running narrowband and don't mind a lossy dielectric your transceiver design can be made to work. But if sensitivity and group delay distortion in broadband is important then you will want to avoid FR4 and use something more compatible at 6GHz. If your dielectric constant changes with frequency you may get phase noise from a non-flat group delay. If it absorbs energy from loss tangent then you suffer on transmission loss and return loss and hence sensitivity.

View attachment 75187

Ideal may be quartz over alumina or ceramic substrate in non PTFE materials which are predicted to triple in next few years, but you did not specify budget constraints. There are many poly plastic substrates being developed now. See HItachi, Panasonic etc

I have worked with ceramic LNA's & duplexers up to 6GHz with very low loss and high return loss using MLM filters with Motorola in Phoenix and I provided the automated test using Anritsu SA accurate to 0.1dB after calibration.