how to know the dielectric constant of a board

[khouly]

i have a PCB board claded with copper
and i need to know the εr of this material
is there any procedure to determine it ?
thanks
khouly

 

[vfone]

Here you can find the methodology how to measure the dielectric constant.
**broken link removed**
Is a relative measurement between two capacitors. One using air dielectric (Er=1) and one using your unknown dielectric.

 

[sergio mariotti]

Many articles has been wrote abouy this topic, but the question is really necessary to accurately know the εr of the laminate?
A big accuracy is needed only if you design High Q, sharp, narrowband circuits; but in this case is also important to know how the εr change vs. temperature, vs. R.H.% etc.

Only a moderate accuracy is required for microstrip and stripline design. So the info ( the answer) can be found on the documentation rather than into the measurement.

You may do a rough estimation just looking the laminate.
Does the dielectric look like PTFE? (if yes probably εr is around 2.2...2.6)
Does the dielectric weight more than Rogers 5880 or similar (more weight means higher εr)
Does cutting the material destroy the cutter blade? if yes the laminate contains ceramics, εr>6 , probably 10.

Be confident that exist few class of εr : the 2.2...2.6, the 4.5 (G10, FR4), the 6 and the 10 (Alumina compatible).
It's rare the case that your laminate is outside the 4 classes.

 

[molloy]

A difference in εr (2.4 instead of 2.3) may cause the differences of ten degrees in the phase of S21 of an amplifier module and if you have the necessity of put them in parallel this can cause a great deal of problems.
So often you need to measure precisely the value of εr and yes of course the εr change vs temperature.

stefano

 

[Johnson]

I am in aggrement with all of you, but the application is not known, khouly: why do you need this parameter? Is it a high speed PCB design(which you can ask you PCB Manuf.) or a sensitive Microwave circuit?

 

[flyhigh]

A difference in εr (2.4 instead of 2.3) may cause the differences of ten degrees in the phase of S21 of an amplifier module and if you have the necessity of put them in parallel this can cause a great deal of problems.
So often you need to measure precisely the value of εr and yes of course the εr change vs temperature.

stefano

Hi,

I also agree that a precise knowledge of Er of a substrate is very important for some applications. The higher the frequency the more important it is.

For example, in manufacturing of branch coupler at 20 GHz it is very important to have all, VSWR minima, isolation minima and insertion loss maxima at the same frequency or you will fail in your design! For teflon based boards on these frequencies the tolerance for Er might not be critical. However for high Er boards, like alumina, even second decimal can play a roll!

But Sergio Mariotti is right from another point. All closed-form models for transmission lines in comercial simulators have around 0.5% error for Eeff and 1% error for Zc, the accuracy for modeling discontinuities are even larger, so after the circuit is put together even knowing the Er right, you can miss your target without good models.

flyhigh

 

[khouly]

i want to know the εr of the board to learn the design of RF and MW circuits and measure it
it is not nessecary to determine it preciselly .
and also cannot work in very high frequency coz the test equipmnet is limilted
and i donot know the board manufcaturer
here it is very hard to find RF componnet and find information about
thanks for information

khouly

 

[abdoeng]

A difference in εr (2.4 instead of 2.3) may cause the differences of ten degrees in the phase of S21 of an amplifier module and if you have the necessity of put them in parallel this can cause a great deal of problems.
So often you need to measure precisely the value of εr and yes of course the εr change vs temperature.

stefano

i agree wih u,it may shift resonant frequency of oscillator design rom freq to another,bu i think that there is atolerance+ or - .1 in epsr ,there is no accurate epsr
regards