Designing Microstrip on a PCB with Unknown Substrate - RF Cafe
Post subject: Designing Microstrip on a PCB with unknown
Unread postPosted: Fri Mar 25, 2005 12:24 pm
I have bought some double sided PCBoard for Microwave
work from Ebay, it was cheap, looks of good quality and is in bulk
but unfortunately I don't know what the substrate material used
in it is, or how thick the copper is. How can I find out, is there
some test that I can perform? I will be etching 50ohm microstrip
lines on it, how can I determine the right thickness for the lines
without any knowledge of the board material or Er (relative substrate
electric permitivity)? Any ideas?
Sat Mar 26, 2005 9:07 am
First check if there are any labels
or markings attatched to any pieces. Post this info up, as it may
contain the information you are looking for.
Arlon and Taconic websites ( I know there are more manufacturers,
but these are the most common)
The physical dimensions are
easily measured with a micrometer or vernier caliper.
a first guess, use a line width that is the same as the material
thickness, although absolute value of impedance will be unknown,
it will be reasonable to work with.
To measure Er, I would
suggest that you etch some test resonators of known dimensions using
this line width onto a spare piece, and measure the response on
a VNA. If you have no VNA available, any means of measuring VSWR
with frequency will do. I would recommend a ring resonator, coupled
lightly as possible. Using ring resonator eliminates the capacitive
end effects of a open line. The S11 (VSWR) response will dip at
the point where the ring is electricaly one wavelength around (use
centre line of strip). Be aware that this will repeat at integer
You can now apply the microstrip design
equations (or goal seek in excel) to give a reasonably accurate
value for Er.
Unread postPosted: Mon Mar 28, 2005 3:40 pm
There are no
markings or stickers on the board itself but I'll try your procedure,
thank you for the tips!
Unread postPosted: Tue May
17, 2005 9:29 am
The guest who suggested the resonator is
correct, the method suggested is the standard procedure used to
determine er ( dielectric constant).. The equation he was refering
to is: Lambda=C/(sqrt(er)*F); where Lambda is the wavelength , c
is the speed of light, f=the frequency.