|PA8W Amateur Radio|
The PA8W RDF41
pseudo doppler radio direction finder kit,
|This microcontroller based Radio
Direction Finder is the result
of 2,5 years of testing and experimenting with my RDF40
In this period, I found ways to remarkably improve the stability and reliability of bearing estimates produced by the RDF.
Although the working principle of the RDF41 does not really differ from the simpler discrete LED-pelorus versions I designed,
the available computing power of a simple microcontroller can enhance performance substantially.
By weighing the quality of measurements this RDF41 manages to calculate a long time average bearing even in very poor conditions.
This can only be done "on the move", since multipath effects -which normally corrupt the measurements- rapidly change randomly during the ride.
Therefore these changes can be considered and treated as noise.
The RDF can still dig out a good bearing estimate simply because it can distinguish good from bad measurements.
It displays 4 real time measurements every half a second, checks their credibility, and displays a long time average using the best measurements.
The necessary algorithms are tested and tweaked over and over again over the last years.
So, now all posibilities are sorted out I decided to develop an almost ready to use kit for those radio amateurs -and other operators- that show interest in this concept.
Professional soft switching for highly reduced noise floor.
High sensitivity: Suitable for weak signals.
Accuracy in good conditions, using a UHF doppler array: 2,2 degrees averaged.
Wide frequency range, 30MHz-1GHz depending on antenna array.
Easy calibration over 360 degrees using the potentiometer.
Several antenna array designs available.
New: Dual mode: Normal mode for mobile applications, Burst mode for catching very short transmissions.
Quality weighing of measurements, using the best measurements to generate a long time average.
Automatic Adaptive Averaging continuously optimizes the amount of averaging in the current situation.
Best measurements are automatically sent over USB to computer in order to plot bearing lines on map. (free mapping program available)
Clear 128x64 pixel display, perfectly readable even in bright daylight.
On screen quality indicator, gives instant insight of multipath distortions.
Bearing pelorus showing the four last measurements plus long time average.
Digital display of long time average and quality factor.
Automatic display freeze below squelch point.
Antenna testing mode available.
Pre-assembled, programmed and tested kit.
Runs on 12V power supply or car battery, consuming less than 80mA.
(8V up to 14V dc, Minus pole connected to mass)
Reverse polarity protected.
The RDF41 does 500 cycles per second over 4 antennas, so it collects 2000 measurements per second in four sample buffers of a digital filter.
Here, initial averaging is done and modulation and noise are suppressed.
The 4 buffers are then sampled by the microcontroller, 4 times every half second.
So twice every second the 4 latest samples are displayed, including the newly calculated long time average and signal Quality.
How does the RDF41 know good from bad measurements?
The following oscilloscope screendump shows a few signals in the RDF41, where the blue signal is the output of the connected FM receiver.
In a pseudo doppler, the jump to the next antenna results in a phase jump in the receiver and therefore a pulse in the receivers audio.
The audio pulses 1,2,3 and 4 belong to antennas 1,2,3 and 4 obviously.
Consider the 4 antennas as two pairs of opposing antennas, say a 1 - 3 pair and a 2 - 4 pair.
In theory, a clean RF field will produce a similar reaction in both antennas of a pair but with opposite polarity.
If that is not the case the RF wavefront must be distorted by multipath (Or the receiver is not tuned to the signal properly)
So, in a clean RF field: If we add up audio pulse 1 and 3, the outcome should be close to zero.
The same is true for pulse 2 and 4.
If not? Then we have a distorted RF field.
The RDF41 compares the pulses of both pairs and calculates a Quality figure depending on pulse amplitude and symmetry in both pairs.
This "Q" figure is a very reliable indicator for the accuracy of a bearing measurement.
The RDF41 uses this Q to decide how much this measurement is allowed to have an impact on the long time average,
thus resulting in a much more stable bearing indication, especially in difficult multipath conditions.
Soft Antenna Switching:
The above picture also shows one of four antenna control signals (yellow).
Like all professional pseudo doppler designs the RDF41 uses soft commutation, massively reducing the receiver's noise floor.
Simple hard switching RDF's (almost all amateur designs!) perform very poor on weak signals and are very susceptable to strong signals near the hunting frequency!
Normal mode using Automatic Adaptive Averaging:
The long time average obviously is an averaged bearing and therefore lags the current measurements a bit.
Starting March 2018 the RDF41 uses a special algorithm to automatically set the amount of averaging, depending on the circumstances:
You are driving on an country road, with lots of reflections from trees and bushes:
The Averaging will go up (256 max) to enhance accuracy and bearing stability.
Also the immunity to reflections will be served by this.
Now you are taking a turn.
Normally very high Averaging will make the RDF respond very slowly to that turn.
The RDF41 will recognise your change of heading so it lowers averaging a lot to rapidly adapt to your new heading.
The same happens if the station you are hunting goes off air and another station starts to transmit;
The RDF "sees" this sudden change and reduces the amount of averaging to be able to focus on the new signal more rapidly.
Note that this response to changing circumstances may not be exaggerated, since it would make the RDF respond to strong reflections as well...
Therefore this algorithm is tweaked "on the road" extensively to get the best possible performance in real life conditions.
New: Burst mode for very short transmissions
Starting August 2018 the RDF41 has an improved second mode specialized for hunting very short transmissions down to 1/10th of a second.
It doesn't build a long time average: simply because there's no time for that.
Instead, it gathers and displays real time measurements, until the signal starts to weaken.
That means that the last bearing is calculated using the strongest part of the signal burst, not the much weaker part just before the squelch closes.
Depending on modulation content of the tracked signal, the RDF41 manages to get good accuracy (within 2-3 degrees) on bursts of only 1/5th of a second.
Even bursts as short as 1/10th of a second yield pretty good results.
In theory this response to extremely short signal burst can be further improved by a small modification in the hardware,
bringing bursts of 1/40th of a second within reach.
Just contact me if that's importanf for your type of use.
Note that this burst mode is meant for fixed or semi-fixed hunting; On the move you will greatly miss the stabilizing effect of the long time average!
Doppler plus Amplitude mode!
This RDF41 is capable of using the doppler working principle as well the amplitude principle, depending on the attached antenna array.
For the doppler principle, an external FM receiver is needed.
A -pseudo- doppler RDF needs some kind of carrier, so it can track FM, AM, FSK signals very well, as long as the signal fits within the receivers passband.
Generally, a 430MHz doppler antenna array will work properly from 350MHz up to 500MHz.
If need be, a UHF array can be used down to 140MHz with reduced sensitivity and accuracy.
For the amplitude working principle, an external AM receiver is needed.
An Amplitude RDF can be used to find ALL kinds of signals, including sparking electric connections, noise sources, etc.
In this website I published a design of a UHF amplitude antenna array.
It works great from 390MHz up to 470MHz.
kit will be delivered as in below picture, including calibration
without connectors and switch, so you can pick your own types.
On the bottom of this page you can find a simple wiring diagram.
And this website shows how to build a simple but high grade antenna array.
So any radio amateur can do the job and attend the next foxhunt with semi professional equipment!
This screenshot shows a reading in a fast curve; the long time average arrow lags the 4 current measurements.
The length of the current measurement arrows show their Quality.
Overall Quality is very good: Q=8.
This is also clear due to the nice symmetry showed by the symmetry indicator.
Averaged bearing is 127 degrees,
Battery voltage is 13.1V,
Calibration is set to 352.
Averaging is automatic, and will in this case be automatically reduced to speed up the long time average indicator.
The following settings are factory set:
Rotation frequency: 500Hz,
The center dot in the pelorus indicates that a measurement was good enough to be accepted,
although Quality is only 4.
The reason for this poor Quality is multipath reception.
Reflections add up to the direct signal and distort the received wavefront.
This is clearly illustrated by the symmetry indicator.
Both horizontal lines should be vertically aligned on the vertical line, which is clearly not the case.
A sure sign that this bearing of 245 degrees may be off by a fair amount.
If Quality drops below 1, the center dot in the pelorus will disappear and the reading will freeze until a good signal is received again.
The mode switch can put the RDF 41 in antenna test mode.
In this mode the RDF slowly steps through all 4 antennas, enabling the user to check if the performance of all four is similar.
In most situations there will be substantial difference in signal strength due to multipath reception.
A defect antenna though will clearly drop out compared to the rest.
Note that the numbering of the antennas is not absolute: The antenna called number 1 may physically be number 2 or 3 or 4, it depends on the moment you switch to antenna test mode.
The numbers are only there for you to recognize that one and the same antenna drops out in performance.
Which antenna that physically is can only be determined by measurement.