Bob kb8tq writes:

The only antenna that u-Blox themselves advertise for use with the
(eval) F9P module is an unassuming 28dB gain puck (ANN-MB, not seen it
anywhere for sale so far).  So I don't think that the gain matters very
much for this module unless you have to compensate for distribution
losses, at which point you'll probably end up with the more typical
35…41db gain figures for fiyed installation antennas.

Interestingly enough the antenna is mentioned in a comparison with two
other multi-band patch antennas here:

https://www.ardusimple.com/ardusimples-oem-antenna-vs-u-blox-tallysman/

You can apparently buy one from their store, with or without an
accompanying F9P RTK-module.

Regards,
Achim.

+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf microQ V2.22R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Hi

Often the “puck” that comes with a demo board is set up to only work when
directly connected to the board. No long cable runs, no splitters, no lightning protection
to all add a bit of loss here and there. It also is a good bet that whatever is supplied
“free” with the demo board came from the low bidder.

About all we have to go by is a line in the spec that talks about 40 db max and 30 db
typical as being the “recommended” gain characteristics. The more common numbers
would be 20 db typical and 30 db max for sort of telecom devices we all are used to working with.

The lack of a “minimum” spec generally is because the problem is actually far more complex
that a simple gain number. Things like antenna element gain matter. Gain delta between
overhead and horizon matters. LNA noise figure matters. Since these are multi band
devices, gain delta between bands matters.

All of that assumes the antenna is sitting on a pole with a perfect view of the sky in all
directions. Does this gizmo get used under tree cover? Do birds land on it? Does snow
and ice pile up on it? How well should it work in those conditions? ( Looking out at the
multiband antennas here with ice and snow rapidly piling up on them yet again …..)

All I’m suggesting is that if you have a choice between antennas up around 40 db and
ones down below 30 db …. the higher gain is a lot more likely to do the job. Since you
can get the higher gain devices for not much money, there’s not much downside.

Bob

Hi Attila,

The "good" patch antennas thus

Mini-Circuits has such a 4-port phase splitter, part number SCQ-4-1650+.
It's $25, which is not super cheap, but it's unclear one could implement
such a network with a planar circuit for less, and the planar circuit would
be more bulky.  I've been thinking of picking up one of these to give an
untuned 4-filar helix a try (not a fancy resonant backfire quadrifilar
helix, which are hard to make broadband, but just a plain set of 4 helices
and the combiner).

broadband:

https://www.minicircuits.com/pdfs/SCQ-4-1650+.pdf

Cheers,
Peter

Hi

If you take a hammer to one of these antennas, they seem to mostly play games
with transmission lines on pc boards. The number of components involved is pretty
small.

If hammering one apart does not sound like a great thing to do, there are pictures
here and there. I have yet to see a picture that shows enough to actually make sense
out of. I’d bet that is intentional.

Looking at network analyzer sweeps, it becomes pretty apparent that whatever
combination of things are involved, they only work inside each of the target sub-bands.
Once you get to the edge, it all falls apart. It comes back together once you get to
the edge of the next sub-band. Some of that is intentional filtering so sorting it out
that way … not so much.

Bob

On Wed, 30 Jan 2019 20:33:23 -0800
Peter Monta pmonta@gmail.com wrote:

That one looks interesting indeed. Especially as the datasheet mentiones
quadrifilar. Now I wonder whether that means the internal structure is
a quadirfilar transformer or it is meant for quadiriflar antennas.
While the latter would be a weird way to describe it, the former sounds
more plausible. But then, I have no idea how one could use a quadrifilar
transformer for a polyphase splitter. Does someone have an idea?

		Attila Kinali

--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neal Stephenson

On Thu, 31 Jan 2019 08:51:46 -0500
Bob kb8tq kb8tq@n1k.org wrote:

Transmission line phase shifters are pretty narrow band.
Getting them to 5% bandwidth with reasonable phase and
amplitude stability is already not straight forward. But
for a combined GPS/Galileo/Glonass L1/L2/L5 antenna,
a bandwidth of over 30% would be needed. Not an easy task
at all. Making them work for a few select bands is a lot
easier, if still a bit non-standard. At least EM simulation
tools help a lot with that.

		Attila Kinali

--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neal Stephenson

Hi

I get the impression that they do indeed run things up a band at a time and
pretty much ignore what is in-between. Again, we’re talking about the “affordable”
end of the new antenna range here and not the price is no object part of the
market.

Bob

On 1/31/19 5:51 AM, Bob kb8tq wrote:

These things are designed by someone sitting down with HFSS or similar,
starting with a cookbook layout and iterating the design manually until
it works.  It's faster than xacto-knife, copper foil tape, etc.
You can literally click and drag and watch the S11 change in real time
if you've got enough CPU horsepower.

Someone who is reasonably facile with the tools (and has a design to
start with) can probably knock one out in a day or two of work.

Then a week to get the prototype back from fab and you test it, and call
it done.

I'll bet they don't agonize too much about axial ratio off boresight or
perfect phase center vs look angle.. It's more about "is the degradation
tolerable for the desired application".

On 1/31/19 9:41 AM, Attila Kinali wrote:

More likely the outputs are 0,90,180,270.  Basically some stripline
couplers cleverly arranged.  A standard Lange coupler generates a 0 and
90, and you can probably arrange the input circuit to make the flipped
phase versions.

Doing it for wideband is the challenge.

https://www.microwaves101.com/encyclopedias/lange-couplers

On 1/31/19 9:45 AM, Attila Kinali wrote:

A straight up branch line coupler or two parallel microstrips would have
that problem.

That's the beauty of the Lange coupler.. it holds the 90 degrees over a
pretty wide bandwidth (octaves are doable)