[time-nuts] Re: Testing a GPS mag mount antenna

[Dana Whitlow]

Skipp,

Several responders have correctly referred to the "noise floor", but I
believe that
none have said what level that is.  And it's not trivial to say so, either,
because it
is dependent on the extent to which the antenna under test sees the sky and
ignores blackbody radiation from the ground.  An ideal GPS antenna would see
only the sky, but not the ground at all.  But none are ideal in this regard
unless
placed on a large level metallic "ground plane".

It may happen that in this era of higher-powered L1 transmitters aboard some
of the satellites, their total noise power spectral density might exceed
cold sky
noise by enough to be obvious on a spectrum analyzer looking at the output
of
an active GPS antenna.

But I've saved the good news for last.  The noise PSD of the ground as
sensed
by a passive antenna is about -114 dBm in 1 MHz BW, or -174 dBm in 1 Hz BW.

So here is what I suggest for your antenna functional tests:
Mount the antenna under test on the end of an insulating horizontal rod,
which
can be rotated to make the antenna face either the ground or the sky.

Examine the spectrum on a span setting of several MHz, and see if the noise
level over a BW of a few MHz is higher with the antenna pointed down than
when pointed up.  Oh, I forgot to mention: for tests like this the antenna
should
be supported a few feet above the ground level, so that input noise of the
LNA
does not get reflected off the ground and back into the antenna to an
appreciable
extent.  It also helps if the ground is very rough, like maybe tall grass.

What you may see with the antenna looking at the ground is blackbody
radiation
from the ground being filtered by the input filter of the LNA- and that
filter hump
is indicative (but not yet proof) of the antenna's working.  When the
antenna faces
the sky, the ground noise hump should disappear and be replaced by the
spectrum
of the combined signals from all the GPS satellites above you.

This is all provided the cascaded combination of the GPS antenna's LNA  and
the
SA's noise floor is is very quiet.  With practical antennas and SA's, this
is unlikely
to ever yield a rise of more than a very few dB if that much.  So, to see
this, you
should setup the SA so that you get an average power spectrum, meaning using
lots of video filtering on an old-fashioned swept analyzer, or trace
averaging of many
traces on an FFT-type SA.  The required averaging time to get good
smoothing of
the noise levels will be minimized by using as wide a resolution bandwidth
as you
can, but always less than the BW of the spread-spectrum GPS signals or the
antenna's
inpuy filter.  But you don't want to go too wide, either, for you will lose
the frequency
resolution that may  let you distinguish between the antenna's filter hump
shape and
that spectrum comprising the GPS signals in view at the time.  I'd suggest
starting with
a BW of around 100 kHz as a fair compromise.  For more on this aspect of
things,
look up "radiometer equation" on the web- we'll make a radio astronomer of
you yet :-)

BTW, it will likely be needed to add some additional gain between the
antenna and the
SA's input.  Many of the older SA's have really rotten NF, like 30 dB or
worse.

Dana    K8YUM


On Fri, Aug 20, 2021 at 4:33 PM Gerhard Hoffmann <
ghf at hoffmann-hochfrequenz.de> wrote:

>
> Am 20.08.21 um 22:14 schrieb skipp isaham via time-nuts:
> >
> > Hello to the Group,
> >
> > I picked a box of used (removed from commercial radio APRS type
> > service) mobile/vehicle GPS Antennas. They are mostly the classic
> > square molded, black plastic magnetic mount type, about the size of a
> > bar of soap when cut to square (2/3 the size of a large bar of soap).
> > The coax length terminates to an SMA connector.
> >
> > I'd like to use some of these unmarked (obviously also unbranded)
> > antennas for a few projects. The initial goal is to first set up a
> > system to test (good/bad) the antennas, then determine their operation
> > voltage, I suspect them to be 3.x to 5.x Volts. They are probably not
> > "new enough" to be the type to operate of 3 or 5 Volts DC.
> >
> > For testing... I purchased a nice NOS Mini-Circuits bias-Tee.
> >
> > The intent is to now operate the antenna through the bias-tee, in to
> > an analyzer. I would initially start the bias supply off at 3 Vdc,
> > while also monitoring current.  If I don't receive an adequate/valid
> > GPS signal off air, I could increase the bias up to 5 Vdc (rinse/repeat).
> >
> > Should I be able to "see something" on or around the GPS frequency
> > other than what I suspect will be something visual looking like a
> > noise/pulse source/signal?
>
> You won't see anything interesting on the spectrum analyzer. The
> signal(s) look like noise, and they are buried in the real noise.
>
> Deeply!  In a real receiver, there are probably just 1 bit ADCs, aka
> comparators, and the receiver  needs to know the pseudo random
>
> polynomial that was used to blow up the bandwidth of the 50 baud message
> to some MHz in order to reverse that effect.
>
> And you have to know which part of the polynomial is currently used.
> This is done by search & correlation tries. Sloppy wording, I know.
>
> Only when that reversal is done you have a positive signal/noise ratio.
> GPS receivers are 95% math, the rest is electronics
>
> and packaging.
>
> You may see a noise molehill at the nominal frequencies, but that means
> only that the pre-amplifier and maybe some filtering works.
>
> And there is a hefty preamplifier to make up for 5 meters of El Cheapo
> RG-174  coax. The GPS pseudo-noise is only
>
> 1 promille of that what you see.
>
> Disclaimer: Last time I was involved in this was with the Plessey
> 1010/2010 chip set in a previous life, for GPS/Glonass combined,
>
> which was new then.
>
>
> Gerhard, DK4XP
>
>
>
>
>
> > I don't yet have a GPS receiver with a signal strength indicator, else
> > I could probably not have to send this post.  But, I do have access to
> > an analyzer, I bought the bias-tee (was reasonable in price) and I'd
> > like to test these 30 antennas to see if they work and determine if
> > 3.x volts is enough... or 5 volts is required.
> >
> > Thank you in advance for any replies and comments.
> >
> > cheers,
> >
> > skipp
> >
> > skipp025 at yahoo.com
> > _______________________________________________
> > time-nuts mailing list -- time-nuts at lists.febo.com -- To unsubscribe
> > send an email to time-nuts-leave at lists.febo.com
> > To unsubscribe, go to and follow the instructions there.
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com -- To unsubscribe send
> an email to time-nuts-leave at lists.febo.com
> To unsubscribe, go to and follow the instructions there.