"mr1956" ha scritto nel messaggio
...
I have posted on here before and found the member's comments very
helpful so here it goes again.

I am trying to stack two Yagi antennas to form a ground station to
receive GPS data from an experimental rocket. The rocket portion of
the system has already been tested but now I have two antennas to
couple together to make up the receiving end. The particulars are as
follows:

I have two 11 db 900 MHz Yagis (Pacific Wireless), both of which have
50 ohm pigtails. I am trying to hook these together in a cross
polarized fashion and need to insert two 75 ohm impedance matching
cables so that I end up with 50 ohms at the cable end attached to the
input of my receiver. Basically, there will be the two antennas,
matching sections, a tee connector, then the single coax going to the
receiver.

I figure that 75 ohm RG-11 coax should work for this purpose and am
trying to calculate the correct length of these impedance matching
sections. One formula I have found online is as follows:

Length (feet) = 246 * VF / Frequency (MHz)

The transmitter sends RF via Frequency Hopping Spread Spectrum from
910 to 918 MHz; consequently, I am using 914 MHz as a baseline.

When I plug in the numbers, I get a length of about 2.1" for the
length of the 75 ohm matching sections. Needless to say, it will be
difficult if not impossible to do a 2.1" length of RG-11 cable with
two connectors. I have considered basically fabricating a "tee"
section out of two short lengths of the 75 ohm coax and RG-8 going to
the receiver (soldering the whole thing together), then connecting the
two Yagi antennas to that.

I suppose my question is this: Is there a longer multiple of
wavelength I can use for the impedance matching 75 ohm sections to
develop a more practical design for what I need? Or, am I stuck with
the 2.1" length due to the frequency?

Thanks,

Curt Newport

The answer to your question is yes. You can use odd multiples of the calcuated
length (my calculation give 2.13" for PE-insulated cables like RG-11, or 2.25"
for teflon insulated cables. For foam-insulated cables, calculation must take
into account the actual cable velocity factor).

However, at 900 MHz, building the coax combiner you have described would be an
headache, due to the difficulty of precisely calibrating the coax pieces length
and to verify their performance.

For a receive-only system a solution you may consider is to use a small
preamplifiers connected to each antenna connector, either directly if feasible,
or using a very short piece of coax.

Doing so, the receive system signal-to-noise ratio would be set by the
preamplifiers and not influenced by any loss occurring after them. So you could
combine the two preamplifier output signals fully disregarding the impedance
issue. You must only make sure that the two coax lengths joining the
preamplifiers to the summation point are of identical length (whatever it is),
so that the two signals get summed in phase. You could even use small coaxial
cables (e.g. RG-58 or perhaps even RG-174 depending on length) that can be very
easily handled, as any attenuation after the preamplifiers will not influence
the receive system noise figure (within certain limits).

The above arrangement would be OK if using two antennas having the same
polarization. But reading your post it seems to me that this is not the case. As
a matter of fact, if I understood it correctly:

- you wish to mount one antenna on horizontal polarization and the other one on
vertical polarization
- your objective is to receive a linearly polarized signal with a randomly
slanted polarization plane.

If so, combining the two antennas using identical pieces of coax, you would
obtain a system polarized on a 45-deg slanted plane, that would not help in
receiving randomly polarized signals (signals that are orthogonal, or nearly so,
to the antenna polarization plane would be strongly attenuated). You should
instead aim at obtaining a circular polarization which causes a steady 3 dB loss
independently of the signal polarization plane.

To do that you must introduce an extra 90-degree phase shift by adding a
quarter-wavelength 50-ohm section to one of the two indentical-length 75-ohm
coax pieces.

All this in theory. In practice I doubt that, at 900-MHz, a system like that
would behave precisely as expected.

Regards.

Tony I0JX, Rome Italy

This is all good information and I can appreciate the difficulty of
creating the 75 ohm matchings sections due to the precision involved.
I have however found a power divider for this frequency that should do
what I need with respect to stacking the two yagis.

That is probably the easiest solution in this case.

C. Newport


On Jan 3, 10:13*am, "Antonio Vernucci" wrote:
"mr1956" ha scritto nel ...





I have posted on here before and found the member's comments very
helpful so here it goes again.

I am trying to stack two Yagi antennas to form a ground station to
receive GPS data from an experimental rocket. The rocket portion of
the system has already been tested but now I have two antennas to
couple together to make up the receiving end. The particulars are as
follows:

I have two 11 db 900 MHz Yagis (Pacific Wireless), both of which have
50 ohm pigtails. I am trying to hook these together in a cross
polarized fashion and need to insert two 75 ohm impedance matching
cables so that I end up with 50 ohms at the cable end attached to the
input of my receiver. Basically, there will be the two antennas,
matching sections, a tee connector, then the single coax going to the
receiver.

I figure that 75 ohm RG-11 coax should work for this purpose and am
trying to calculate the correct length of these impedance matching
sections. One formula I have found online is as follows:

Length (feet) = 246 * VF / Frequency (MHz)

The transmitter sends RF via Frequency Hopping Spread Spectrum from
910 to 918 MHz; consequently, I am using 914 MHz as a baseline.

When I plug in the numbers, I get a length of about 2.1" for the
length of the 75 ohm matching sections. Needless to say, it will be
difficult if not impossible to do a 2.1" length of RG-11 cable with
two connectors. I have considered basically fabricating a "tee"
section out of two short lengths of the 75 ohm coax and RG-8 going to
the receiver (soldering the whole thing together), then connecting the
two Yagi antennas to that.

I suppose my question is this: Is there a longer multiple of
wavelength I can use for the impedance matching 75 ohm sections to
develop a more practical design for what I need? Or, am I stuck with
the 2.1" length due to the frequency?

Thanks,

Curt Newport

The answer to your question is yes. You can use odd multiples of the calcuated
length (my calculation give 2.13" for PE-insulated cables like RG-11, or 2.25"
for teflon insulated cables. For foam-insulated cables, calculation must take
into account the actual cable velocity factor).

However, at 900 MHz, building the coax combiner you have described would be an
headache, due to the difficulty of precisely calibrating the coax pieces length
and to verify their performance.

For a receive-only system a solution you may consider is to use a small
preamplifiers connected to each antenna connector, either directly if feasible,
or using a very short piece of coax.

Doing so, the receive system signal-to-noise ratio would be set by the
preamplifiers and not influenced by any loss occurring after them. So you could
combine the two preamplifier output signals fully disregarding the impedance
issue. You must only make sure that the two coax lengths joining the
preamplifiers to the summation point are of identical length (whatever it is),
so that the two signals get summed in phase. You could even use small coaxial
cables (e.g. RG-58 or perhaps even RG-174 depending on length) that can be very
easily handled, as any attenuation after the preamplifiers will not influence
the receive system noise figure (within certain limits).

The above arrangement would be OK if using two antennas having the same
polarization. But reading your post it seems to me that this is not the case. As
a matter of fact, if I understood it correctly:

- you wish to mount one antenna on horizontal polarization and the other one on
vertical polarization
- your objective is to receive a linearly polarized signal with a randomly
slanted polarization plane.

If so, combining the two antennas using identical pieces of coax, you would
obtain a system polarized on a 45-deg slanted plane, that would not help in
receiving randomly polarized signals (signals that are orthogonal, or nearly so,
to the antenna polarization plane would be strongly attenuated). You should
instead aim at obtaining a circular polarization which causes a steady 3 dB loss
independently of the signal polarization plane.

To do that you must introduce an extra 90-degree phase shift by adding a
quarter-wavelength 50-ohm section to one of the two indentical-length 75-ohm
coax pieces.

All this in theory. In practice I doubt that, at 900-MHz, a system like that
would behave precisely as expected.

Regards.

Tony I0JX, Rome Italy

Antonio wrote:

Combining the two antennas using identical pieces of coax, you would
obtain a system polarized on a 45-deg slanted plane, that would not help
in receiving randomly polarized signals (signals that are orthogonal, or
nearly so, to the antenna polarization plane would be strongly
attenuated). You should instead aim at obtaining a circular polarization
which causes a steady 3 dB loss independently of the signal polarization
plane.

To do that you must introduce an extra 90-degree phase shift by adding a
quarter-wavelength 50-ohm section to one of the two indentical-length
75-ohm coax pieces.


Instead of this, why not use identical feedline lengths into a Tee and
position one cross polarised antenna 0.25WL in front of the other, to obtain
the 90-deg phase shift?

David, VK3BDK

Instead of this, why not use identical feedline lengths into a Tee and
position one cross polarised antenna 0.25WL in front of the other, to obtain
the 90-deg phase shift?

David, VK3BDK

Doing as you propose is possible, and would produce the same circular
polarization on the antenna boresight.

But moving away from the boresight, the circularity of polarization would
degrade (into elliptical polarization) faster than if instead using a
quarter-wavelength delay on one of the two feedlines.

73

Tony I0JX

Antonio Vernucci wrote:

Instead of this, why not use identical feedline lengths into a Tee and
position one cross polarised antenna 0.25WL in front of the other, to
obtain
the 90-deg phase shift?

David, VK3BDK

Doing as you propose is possible, and would produce the same circular
polarization on the antenna boresight.

But moving away from the boresight, the circularity of polarization
would degrade (into elliptical polarization) faster than if instead
using a quarter-wavelength delay on one of the two feedlines.

73

Tony I0JX

Interesting..
I just ran some NEC models with a pair of crossed dipoles. Two cases:
1) where the two are a quarter wave apart and fed in phase
2) where the two are in the same place and fed 1/4 wave apart

here's the axial ratios
angle spaced phased
0 0.99895 0.99373
5 0.99259 0.99173
10 0.9686 0.9774
15 0.92991 0.95122
20 0.87836 0.91519
25 0.81619 0.87053
30 0.74579 0.8185
35 0.66954 0.7604
40 0.58964 0.69752
45 0.50804 0.6311
50 0.4265 0.56226
55 0.34658 0.49196
60 0.26988 0.42098
65 0.19811 0.34989
70 0.13339 0.27906
75 0.07827 0.20868
80 0.03576 0.13878
85 0.00889 0.06928
90 0 0

Not a huge difference.. you have to be 45 degrees away before it's 10%
different. If you were using directional antennas, the gain would be
way down by the time you were that far off boresight, so it would be a
non-issue.


I suspect that other factors would be bigger in importance.

Not a huge difference.. you have to be 45 degrees away before it's 10%
different. If you were using directional antennas, the gain would be way down
by the time you were that far off boresight, so it would be a non-issue.


I suspect that other factors would be bigger in importance.

I agree that, with a high gain antennas, it makes almost no difference at angles
within the antenna main lobe. I could make more difference with regard to the
susceptibility to strong interfering signals coming far away from the boresight,
but this may be quite a theoretical case.

It could be more important in the case of two crossed dipoles (turnstile)
typically used with low-orbiting satellites, but even in that case differences
may be hardly noticeable.

In any case it is good to keep that in mind.

73

Tony I0JX

I suspect that other factors would be bigger in importance.

BTW, if you work at JPL, please give my 73 to Eric Archer N6CV who was at my QTH
at least ten times!

73

Tony I0JX

Antonio Vernucci wrote:
I suspect that other factors would be bigger in importance.

BTW, if you work at JPL, please give my 73 to Eric Archer N6CV who was
at my QTH at least ten times!


N6CV is just down the hall, and speaks well of you.

N6CV is just down the hall, and speaks well of you.

Good to know that, hi.... Please tell him that we are waiting for him to come to
Rome again soon.

73

Tony I0JX