Dividing the output of a VCO or....

[rdawe]

TL;DR Summary

To drive 3 identical loads with the same Local Oscillator frequency, is it better to divide the signal of a single VCO by using a Wilkinson Power Divider, or a buffer amplifier with multiple outputs or just 3 individual VCO's?

Hello, I am looking to build a small RF circuit to drive a pan/tilt antenna tracker. The tracker will have 3 identical directional antennas arranged in a triangular pattern and will use RSSI to follow the signal (obviously) horizontally and vertically.

My idea is to design a single PCB containing 3 individual RSSI receiver IC's (LT5504), each connected to one of the 3 antennas.
The RF circuit will output analog voltage signals to a small microcontroller (arduino) which will then drive 2 stepper motors.

The LT5504 requires a Local Oscillator input to set the appropriate received frequency. In my case, the LO frequency required is approximately 1140mhz for a 915mhz receiver frequency.
I have purchased a Crystek CVCO55BE-1000-2000 VCO for the LO.

My question is, how to properly drive the 3 LT5504 receivers with the identical LO frequency, while maintaining the appropriate RF input power and impedance (of the LO output) while keeping any noise/interference to a minimum?

My ideas are:
1. Purchase 2 more VCO's and drive each receiver with it's own LO. (Seems like a waste of money and battery power.)
2. Divide the output signal of the single LO using a PCB microstrip Wilkinson Power Divider. (LO power will be divided by 3 and reduced below that required by receiver IC's. Also difficult to design for 3 outputs.)
3. Use some type of LO Buffer Amplifier with multiple outputs and baluns at each receiver LO input. (MAX2472/2473 look like possible options, not sure how to arrange correctly for 3 outputs. Do I need baluns to match impedance of LO to receiver after dividing the signal?)
4. Build 3 individual PCB's using option 1 (more cost, more space required and more battery power consumed.)

Although I have no formal RF training, I do have a pretty good understanding of basic electrical theory and the many issues surrounding frequency generation, noise, ground plane effects etc.
Also, this is just a fun/learning project I am doing as a hobby.

Thanks in advance for any advice and/or opinions.
Rob
https://www.analog.com/media/en/technical-documentation/data-sheets/5504f.pdf
https://www.digikey.ca/htmldatasheets/production/59230/0/0/1/cvco55be-1000-2000.html
https://datasheets.maximintegrated.com/en/ds/MAX2472-MAX2473.pdf

I

 

[Baluncore]

Welcome to PF.

To drive 3 identical loads with the same Local Oscillator frequency, is it better to divide the signal of a single VCO by using a Wilkinson Power Divider, or a buffer amplifier with multiple outputs or just 3 individual VCO's?

My first thought would be to have one LO and use it to drive three identical buffers amplifiers. That will best drive the lines and isolate the separate distribution systems.

The tracker will have 3 identical directional antennas arranged in a triangular pattern and will use RSSI to follow the signal (obviously) horizontally and vertically.

Pan and tilt = azimuth, elevation in antenna terms.
But then I wonder how you can detect off-axis direction of arrival.

If the dishes are parallel, all pointing the same way, they will all produce the same RSSI.

If you built a monopulse feed, with the three antennas and receivers, all will be near the focus, and you will only require one dish.

Maybe you are using a sum and difference hybrid to make an interferometer? Then you need to measure the phase differences between the antennas, so separate LOs are not possible.

If you are converting direct to baseband, using an image reject filter, then using the same LO frequency as the RF input, you will need a high pass filter to remove DC, and the RSSI will need to cover the bandwidth being received.

So, how will you make the decision, up or down, left or right?

 

[Baluncore]

The normal way to track a signal source is with a monopulse system, where the separate antenna elements surround the focus of the one dish. If the source moves left, the left signal increases as it approaches the focus, while the right signal falls away out of the main lobe, the left-right balance tells the servo controller which way to drive the azimuth axis. The same for elevation.
https://en.wikipedia.org/wiki/Monopulse_radar#Implementation_for_reflector_antennas

You can eliminate the errors between the receivers and the RSSI systems by using only one receiver and detector. You switch the antennas sequentially into the receiver, say every 10 msec, the RSSI shows a DC step pattern that gives the relative strength of all the signals at the antenna. You can subtract or digitise those to control the servo motors.

 

[tech99]

I think a three way Wilkinson Divide is possible and looks simple, and in any case, you have to obtain at least three times the power whatever you do (unless switching outputs).
If converting direct to baseband, the control system will require a response down to dc I believe.

 

[Baluncore]

The device @rdawe linked to, the LT5504, is specified as having an internal IF band from 50 MHz to 450 MHz. DC will not be a problem. The front-end RF mixer appears to be symmetrical, and so is not an image rejection mixer. Both the LSB and USB will fall in the RSSI detector. All selectivity must be done before the RF input, by the antenna and coupling network.

The position and arrangement of the antennas will determine the LO distribution requirements, which will have implications to the RSSI. That has not been made clear yet.

It would help if we knew what the frequency and bandwidth of the tracked signal would be. The modulation envelope may have a significant effect on the ability to accurately estimate RSSI.

 

[rdawe]

Hello all and thank you for the replies.
Sorry for the delayed response, long day yesterday.

To answer some of the previous questions...
Frequency of tracked signal is 915mhz band (telemetry radio with approximately 250 to 500mW transmitting power) not sure about channel bandwidth. These radios are FHSS and I am not sure how that will affect the system. If it is a problem, I can always use the signal from a 915mhz video transmitter.

With this system, the LT5504 IC's output a voltage proportional to the RSSI. This voltage output would be sent to the microcontroller and a comparison made by the software to then drive the array in the direction of antenna (by way of stepper motor controllers) with the higher RSSI.

The idea of switching antennas to a single receiver seems a bit complicated, as I have no idea how the switching mechanism would function or how to implement it.

The actual layout of the antenna array will be an experiment in itself, with varying spacing and perhaps introducing some splay (although I think splay may cause the array to hunt erratically and be far less accurate).
I may actually have a 4th antenna centered in the triangle that is not connected to the system, but to the actual video or telemetry receiver I want to get the feed from.
I am thinking on a spacing of about 18" to 20" between antennas. Even if they are parallel, I would think there should be some difference in the RSSI due to the offset angle.
This will not be for extremely long distances, maybe 5 to 10km max.

The idea of the Monopulse system is interesting but beyond my understanding to experiment with at this time (although I did read the entire Wiki you linked to... thank you).

I think the suggestion of 3 individual buffer amplifiers may be the easiest to implement.
The MAX2473 looks like a good candidate for this purpose.
Ensuring the feed from the VCO to the 3 buffer amplifiers remains free from stray interference will be an important consideration.

Thanks again.
Rob

 

[Baluncore]

Frequency of tracked signal is 915mhz band (telemetry radio with approximately 250 to 500mW transmitting power) not sure about channel bandwidth. These radios are FHSS and I am not sure how that will affect the system. If it is a problem, I can always use the signal from a 915mhz video transmitter.

The Frequency Hopping Spread Spectrum (FHSS), with a frequency range 910 to 918 MHz, is going to be jumping around everywhere in that narrow 8 MHz wide band. Any other FHSS in that band will bury your signal, as will your vision carrier. I believe you will have to use your vision carrier as the tracking reference.

I am thinking on a spacing of about 18" to 20" between antennas.

The actual layout of the antenna array will be an experiment in itself, with varying spacing and perhaps introducing some splay (although I think splay may cause the array to hunt erratically and be far less accurate).
I may actually have a 4th antenna centered in the triangle that is not connected to the system, but to the actual video or telemetry receiver I want to get the feed from.

You will find it much easier to use a single reflector dish. If you consider the focus of a parabola, it is in the middle of an image plane. Signals from slightly different directions will map onto that plane. So by placing all your antenna elements near the focus on one dish, you can greatly increase the physical stability of the system. What is the diameter of the dish you will use? Your suggested 20” between centres is about 500 mm, so your array would be about 1 metre across. Let's replace that with a single 1 metre dish = 1000 mm. At 915 MHz the wavelength is 300 / 915 = 328 mm.
With a 1000 mm dish, the beam width between the -3dB points will be 57° * 328 / 1000 = 18.7°.
So the -3dB point (where the offset elements could be placed) will be 9.3° off the prime focus.
Let's assume a 750 mm focal length, we get antenna element separation of 750 * Tan( 9.3° ) = 123 mm. That is about the size of the problem. Adjust what you will to make it fit.
You must also design the elements. Will the signal be linear or circular polarised?

There are mobile phone bands around there so there is a question of interference from other channels or operators. Your RSSI is 400 MHz wide, so is highly likely to be receiving other signals, including from aircraft.

 

[rdawe]

I can definitely use the 915 video transmitter as the source. I will then use 433 for telemetry so not to add too much interference from a similar frequency transmitter (although harmonic from 433 comes close to 915 band).

As far as a parabolic dish goes, I do have an old radar dish off one of our old HS748 aircraft. It is 18" diameter with a thru- hole in the center, so easily adaptable for mounting the new elements.
Circular polarization would be my preference as it has better resistance to reflectivity to avoid multipath issues.
The elements would most likely need to be custom fabricated, as I am not likely to find off the shelf solutions.
Thanks again for being generous with your knowledge, I am very appreciative.

 

[Baluncore]

Circular polarization would be my preference as it has better resistance to reflectivity to avoid multipath issues.

Circular polarisation also has the advantage of being orientation independent.
A quadrifilar antenna under the aircraft would transmit efficiently towards the horizon, but there should be no problem transmitting vertical linear from the aircraft, with circular polarised tracking elements.
It is important to realize that the sense of circular polarisation will be reversed by reflection in the parabolic reflector. That would need to be handled correctly if the aircraft also transmitted circular polarisation, but would not matter if the aircraft transmitted linear vertical.