[time-nuts] GPS antenna in silicon/RTV encapsulation

Jim Lux jimlux at earthlink.net
Tue Apr 15 07:43:10 EDT 2014

On 4/15/14, 1:53 AM, nuts wrote:
> I'd be inclined to look at radome construction.
>> http://www.mpdigest.com/issue/articles/2008/may/mfg/default.asp
> The E-3 AWACS is mostly S-2 glass, but they need the strength. For a
> radome sitting outside, you might be able to do better.

Radome design is considerably more complex than just putting something 
over an antenna.

Typically, they make them as two face sheets separated by a honeycomb, 
but the dimensions and materials are chosen to minimize the reflection 
losses (e.g. the spacing might not be constant in the radome, depending 
on the angle of incidence of the radiation).  A very thin face sheet is 
a tiny fraction of a wavelength, so the reflections from the two 
surfaces are almost in the same phase.

For folks like time-nuts interested in parts in 1E15, this kind of thing 
makes a difference; not so much because of the attenuation, but because 
the reflected waves can cause a small phase shift in the apparent 
carrier phase; e.g. a 20dB reflection at 90 degrees shifts the apparent 
phase by about arctan(0.9)= 25 degrees.

Typically in a radome, you shoot not only for low loss, but also low 
epsilon, so the reflection effects will be less.
> The advantage to S-2 glass is you can buy it easily, especially if you
> have a Tap Plastics handy. It works well with the Marine Resin they
> sell, so you know it will last a long time outdoors.

What are the dielectric properties of such a composite?  There's a 
reason why we don't build microwave circuits (in general) on G-10 or 
FR-4 material and use more exotic Rogers or Taconic substrates. (there's 
other reasons too..)

Typical fiberglass uses an acrylic resin and acrylic has the 
spectacularly high loss tangent of 0.02 at only 1 MHz. The glass would 
help bring the loss down (since glass has a somewhat lower tangent, and 
if you have a fairly "dry" mix with small resin fraction, that helps)

FR-4 has an epsilon of about 5 at low frequencies falling to a bit more 
than 4 at GHz frequencies, depending on the glass/epoxy ratio, and a 
loss tangent of around 0.01 at 1-2 GHz.  This is quite high compared to 
sheet plastic of one sort or another.

Take home message: I wouldn't fabricate a radome out of surfboard/boat 
building fiberglass unless you are very careful with the EM design.

More information about the time-nuts mailing list