Hi

With things like the uBlox F9 now out on the market cheap …. I would go with
an antenna that will do L1 / L2 / L5 and work with everything that it up there.
You still are in the “under $100” range (delivered) for new product from China.
It’s a good bet that the guts of all of them are made there. It’s also a good bet
that they all are ceramic slab style designs.

Bob

Hi Denny and list,

Some thoughts from me...worth about what you paid for them considering
my Time-Nuts membership card was revoked the week before last for
heresy, hahaha!  ;)

I've got the GPS-only version of this antenna as my main antenna bought
off the e-place - example:  https://ebay.us/pbDgU2  (no connection to
the seller, not even sure if this is the place where I got mine)

It's a very nice antenna and has a nice mounting kit with it.  The
filters on the GPS-only version I've got are tight enough to block out
GLONASS.  IMHO, this is nice, as it shows they paid attention to the
design.  On the other hand, now that I do want to receive GLONASS...I've
got one of the eBay shipped-from-China "CORS RTK GNSS Survey Antenna
high gain measurement GNSS GPS GLONASS BDS" antennas on the way that
folks were discussing in the uBlox F9 thread here.

I have no experience with commercial active splitters.  I've used either
commercial passive units, homebrew active units, or modified commeriical
passive units.

Your proposed setup is pretty close to mine.  From my PCTEL to the
Polyphaser lighting arrester is about 30 feet, and from the arrester to
my splitter is about 10 feet.  The 30 feet is Cable Experts CXP-1318,
the 10 feet is RG-58.  Depending on what receiver is on the splitter,
add a couple of feet of RG-316.  (some are farther, some are closer, but
it's about two feet from the splitter to the Z3801 or the TruePosition.)

My splitter is a modified Mini-Circuits ZC6PD-1900W.  This is a six way
splitter, theoretical loss 7.8 dB.  I've modified it by adding a DC
block capacitor to 5 of the output ports so that the PCTEL is fed +5VDC
from the Z3801, and none of the other receivers see this DC voltage.
This was simple - cut the traces at the five output ports I wanted to
block, soldered on a SMD capacitor of value I cannot remember.

The VNA at work says I've got from 9 to 11 dB loss from 1.0 GHz to 2.5
GHz.  It's actually a little less, but those are easy numbers to use.

Why the passive splitter?  I was able to get these off of the e-place
for like $10 each...while an active 6-way splitter was at least $100 used.

This setup, while sub-optimal in so many ways, works pretty well!
Here's a screen dump from Lady Heather:

http://www.kd5byb.net/heather/trueposn.gif

Lowest signal is 40 dBc...and it's pretty much always has the full 8
satellites it can simultaneously track.

The Z3801 shows the lowest signal strength right now as SS = 118 for PRN
20, and it's not tracking PRN 14 if you want to compare it's signal
strengths versus the TruePosition.

Doing some quick loss calcs:  (warning:  I've not had my coffee yet, so
very prone to error!)

30 feet of CXP-1318 = 2 dB loss (had to assume similar to Belden 9913 as
Cable Experts doesn't have loss above 400 MHz)

arrester = 1 dB loss (a guess)

10 feet of RG-58 = 2.5 dB loss

splitter = 10 dB loss

2 feet of RG-316 = 1 dB loss

Sum of the loss = 16.5 dB.  Probably should toss a few more dB in there
because of connectors, etc...

So even with my crappy, far-less-than-optimal setup, I get pretty good
results.

Like I said earlier, I haven't had my coffee yet so all of the above may
be totally worthless.  ;)

thanks much and 73,
ben, KD5BYB

Hi again,
The very best GNSS antennas tend to be based on suspended patch antenna
(air-dielectric) structures because they give the best
bandwidth/radiation efficiency (and hence, noise temperature)
performance.  The very best include choke-rings for multipath
suppression (Dorne-Margolin & variants), but these are costly items. 
Miniature ceramic pucks can have pretty horrible radiation efficiency,
which degrades noise performance, so be careful when buying Chinese
cheapies.    I have actually seen commercial antennas where resistors
were added (before LNA) to improve antenna return loss!

On 27.01.19 14:28, Bob kb8tq wrote:

Hi

True indeed and for about $10,000 you can get pretty close to “the best”. If you
want to slum it, you can get “pretty good” for about $2,000.  If you are truly
after high end performance new is the way to go. Getting the full  modern suite
of signals on a used one …. not so much. Even getting a used one that isn’t a bit
beat up may be a challenge. I have also seen used antennas listed for more
than they cost new …..

Since all that “best” stuff applies mainly to survey work, it’s not clear that
one needs to spend that much for a timing antenna. In the context of
what’s coming, you do very much want all the bands / all the signals /
all the services. That will matter, even for timing. At the very least the European
system should be a great timing source (as GPS already is). The Russian
system keeps getting better. I have not seen a lot of papers showing the
Chinese doing quite as well. I’d bet they will catch up if they have not
already.

Why is best different for survey vs timing? For a good survey, you want
sat’s spread out all over the sky. That quickly gets you to sats that are
close to the horizon. Multipath it going to be an issue when you do that.
For timing (with a well known static location of course) your best sat’s
are straight overhead. (the path through the atmosphere is shorter, the
path is easier to estimate / correct ). You set things like an elevation mask
to toss out the sat's likely to give multipath.

Lots of variables ….

Bob

On Sun, 27 Jan 2019 14:24:52 +0000
Bill Slade slade_bill@hotmail.com wrote:

While it is true that air-patch antennas give good bandwidth
and radiation efficiency, they are not really ideal antennas
for GNSS applications. The main drawback is their quite high Q,
which results in high group delay variation. Another annoyance
is that patch antennas form wave guides between the radiating
element at the top and the ground plane at the bottom. This leads
to radiation out at the sides and thus increased side and backlobes,
which in turn limit both multipath surpression and the axial ratio
(ie the ratio between RHCP and LHCP radiation)

The cheap patch antennas also employ only a single feed-point, which,
due its asymmetry, leads to non-uniform radiation and phase patterns
(aka movements of the phase center). The "good" patch antennas thus
employ a four point feed, but this makes the whole antenna quite a
bit more expensive, as a 0°/90°/180°/270° phase spliter/hybrid is
needed. Of course, any such circuit is rather difficult to make
wide-band and thus becomes the bandwidth limiting element.

Of the common geodetic class antennas I am aware of, only the
Trimble Zephyr is a patch antenna[1], but with an rather complex
n-point feed. The Novatel PinWheel[2,3] is a variation of the
archimedian spiral antenna. Most of the single frequency geodetic
GPS antennas are, AFAIK, crossed dipoles, often combined with
a electromagnetic bandgap (e.g. choke rings) to increase impedance
towards low and negative elevations. I have seen bow-tie antennas
as well, but I do not know how common these are.

The jury is still out on what the best antenna structure for geodetic
work is, but it seems that it is more an issue of how much money
someone spends on optimizing the non-ideal behaviour than a fundamental
issue of the structure itself.

For more information on GNSS antennas, I recommend the book
by Rao et al. [4]. It's quite expensive if you buy it normally,
but Artech has some sale every half year. If you wait for the
right one you can get the book for 30-50% less.

		Attila Kinali

[1] "The Design and Performance of the Zephyr Geodetic Antenna",
by Krantz, Riley, Large, 2001
https://kb.unavco.org/kb/file.php?id=135

[2] "A Novel GPS Survey Antenna", by Waldemar Kunysz, 2000
http://meridware.com.tw/NOVATEL/Documents/Papers/gps600antenna.pdf

[3] "High Performance GPS Pinwheel Antenna", by Waldemar Kunysz, 2000
http://webone.novatel.ca/assets/Documents/Papers/gps_pinwheel_ant.pdf

[4] "GPS/GNSS Antennas", by Rao, Kunysz, Fante, McDonald, 2013

--
<JaberWorky> The bad part of Zurich is where the degenerates
throw DARK chocolate at you.

On 1/27/19 6:24 AM, Bill Slade wrote:

In the choke ring world, the typical elements I've seen recently tend to
be crossed drooping dipoles

There's also the ever popular helibowl (a helix wound on a plastic cup
(red Solo beer cup is a bit too big) in a metal bowl - think the things
under a burner in a stove) - non-critical, wideband, etc.

https://www.febo.com/pipermail/time-nuts/2012-May/067138.html
One reference says Spitzmesser apparently based it on the "helicone"
Carver, K, "The Helicone - A Circularly Polarized Antenna with Low Side
lobe Level" Proc IEEE, vol AP-55, #4, Apr 67, p559.

Maybe, maybe not - I get the impression it's a "lets try this and see if
it works" antenna.

So I'm going to guess it's not the bit "IEEE Proceedings", but the Trans
on Ant and Prop.

Helibowls don't have horizon to horizon coverage.

Like these?

https://www.pctel.com/antenna-product/multi-gnss-l1-l2-l5-precision-antenna-gnss-l125-tnc/
https://www.pctel.com/antenna-product/multi-gnss-l1-l2-l5-precision-antenna-high-gain-gnss-l125-40tnc/

Denny

Hi

The first one is designed to mount on a ground plane (which is fine if it is going
on a truck or airplane). It also is a bit low in gain for what I think the F9 is looking
for.

The second one has the right sort of gain, but still is targeted at a ground plane
(rolling stock) application.

Bob