PA8W Amateur Radio              

Back
The RDF43 Color pseudo-doppler/amplitude  radio direction finder

The RDF43color:

Full color TFT display 480x320 pixels.
All settings controlled by a Rotary Encoder.
Gated sampling for best performance with any particular radio.
Antenna drive control for best possible soft commutation.
Real time array fault detection.



All Features:

Professional changeable soft switching for highly reduced noise floor.
Gated sampling for best performance with your radio.
Works as a Pseudo Doppler as well as an Amplitude direction finder.
High sensitivity: Suitable for weak signals.
Accuracy in good conditions, using a UHF doppler array: 2,2 degrees averaged, or 1,5 degrees using a UHF Amplitude array.
Wide frequency range, 27MHz-2GHz depending on antenna array.
Several antenna array designs available.
Dual mode: Adaptive Averaging mode for mobile applications, Burst mode for catching very short transmissions down to 200mSec.
Quality weighing of measurements, using the best measurements to generate a stable long term 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. (mapping program available)
480x320 pixel Color Display.
On screen symmetry indicator, gives instant insight of multipath distortions.
Bearing pelorus shows the last 2 measurements plus the long term average.
Digital display of long term average bearing and quality factor.
Automatic display freeze below set squelch point.
Elevation indicator for airborne sources. (not suitable for short signal bursts)
Antenna test mode and real time array fault detection.
Built-in audio amplifier with volume control.
Built-in speaker.
Runs on 12V power supply or car battery, consuming less than 250mA.
(11V up to 15V dc, Minus pole connected to mass)
Reverse polarity protected.
32 preset memories including 30 character text lables per preset, to store and name your settings.

For working principles please check the RDF40 and RDF41 page.

We'll zoom in now on the adjustable parameters of the RDF43.

Screenshots:


The Menu page gives direct access to the hunting mode and all settings.

Simply use the Rotary Encoder to navigate and Push to enter your choice.

Then you can change the corresponding settings and you get back to normal operation by pushing the encoder.

Below we will go through all menu options.

Note that a changed setting will be active immediately UNTIL the RDF is restarted.

So, if you want the changed setting to stay active after power-up or reset you will first have to save the new settings in one of the 32 presets.






This screenshot shows the standard operational (Hunt) screen: 
Right after the model name there's the battery voltage: 10.7V
The next line shows the Average number, currently 5,
followed by the Preset number 2.

Just below this block a red colored CHECK ANTENNA warning can be displayed.
In hunting mode, the RDF43 does a real time check of the antenna control lines.
In case of passive arrays, it can only detect control wire short circuit to ground.
In case of a preamped array it can also detect a broken wire or a defect preamp.
Additionally, an OVERLOAD! warning can be displayed here, in case the audio input is too high.


In the center of the screen in bold the Angle of Arrival of the hunted signal is displayed. (135 degrees)
Right below the AOA
the word "Export" appears as soon as a good bearing is being exported to the USB serial bus.

In the right bottom corner there's the elevation indicator.
(Note that the elevation indicator only works well on good, constant signals.)


A little to the left: Quality:8, followed by the symmetry indicator.
The better the symmetry of the two horizontal lines, the better the credibility of the measurement.

The pelorus shows the direction corresponding to the numerical AOA of 135 degrees.
The center green ball in the pelorus indicates a good, accepted measurement.
If the measurement is rejected because of poor quality, the center ball will be red.

Note that in Burst mod, both the symmetry indicator and eleveation indicator are empty.

Burst mode is only intended for short signal bursts. In this mode, the RDF collects data until the doppler tone stops growing in amplitude.
This way, the very best part of the incoming short signal pulse is used to calculate an angle of arrival.

In the Hunt screen, rotating the encoder will force a bearing export to USB serial, pushing it will return to Menu.


Backlight intensity:
Backlight runs from 1 (dim) to 15 (bright)
Simply set a value by rotating the encoder.
Doing that the number shfts its horizontal position on the screen so you get a feel for the number of options.
Backlight intensity 1 would be on the left side and 15 would be on the right.
Push the encoder to activate the chosen intensity and automatically return to the menu.

Digital Filter:
Digital filter settling time runs from 1 (fast) to 8 (slow)
Lower settings are great for fast response but the RDF43 may act nervous especially in mobile applications.
Also modulation on the received signal wil have a larger impact.
Higher settings do a much better job stabilizing the averaged bearing, ideal for mobile applications and on signals with high modulation levels.

The best Digital Filter setting also depends on other settings: If you have set a Gated Sampling number higher than 0, this means that the sampling time is only a small portion of the cycle time.
Therefore, you should take the Digital Filter number down about 2 or 3 steps to get comparable behavior.
So, Digital Filter 4 up to 8 is a good starting point when Gated Sampling is set to zero,
and
Digital Filter 1 up to 5 would be fine when Gated Sampling is set to any higher number than 0.

Be sure to pick a very low Digital Filter setting in Burst mode (Averaging=0  setting) for fastest response.










New: Gated Sampling, please read carefully!

Gated Sampling is set to zero by default.
That means that the sampling takes place during the entire cycle of every antenna, similar to how it's done in the RDF40/41/42.
However, depending on the used receiver, the phase information in the audio has a certain delay and is maximum only for a short portion of the antenna on-time.
So it is beneficial to concentrate sampling to exactly that portion that has the best phase information.
Because it gets most of the important information out of your audio but it cuts out a lot of disturbing factors such as modulation.
Result: Increased sensitivity and accuracy,  better suppression of noise, multipath clutter, modulation etc.








Method:
Simply because the group delay in receivers vary massively, we can't use a fixed sampling time frame, since it would be great for one specific radio, and garbage for most others...
Therefore I never implemented that idea in my RDF's.
But for the RDF43 I designed a rather simple method to set the best time window for any connected receiver:

First consider the array to be a circle like a pie with 4 sweet cherries on top.
Every rotation cycle exists of 4 antenna quarters, so we neatly cut our pie in 4 parts, with a cherry on each part.
Normally, we sample during the entire time of a antenna quarter so the entire quarter of the pie is eaten to judge the cherry...
But now we cut each quarter of the pie in 8 slices. And find out which slice has the cherry.
That's the slice we are going to use.
And in case the cherry is cut in half we can even pick two adjacent slices to make sure we got the whole cherry!




That's exactly how this Gated Sampling works.
For those who don't love cherry pie I've added a timing graph with the sampling moments in red.

Operating Procedure:
Tune to a stable test signal that shows good symmetry (clean, no-multipath reception, preferrably without modulation).
Then step through the 15 gate options to see which one gives the highest output number.
When the best setting is found, push ENTER to activate.

Automatic:
The last option is the auto-function:
Again, using a good stable test signal, select the Auto-function.
The RDF43 will take a few seconds stepping through all possible gating options and pick the best one,
and then return to Menu.

In case you find it difficult to pick the right sampling moment, or simply don't want to bother, leave this setting on zero, since this is the full sampling time like used in the RDF41 and RDF42.

After changing this parameter you need to re-calibrate Azimuth and Elevation, since a lot of parameters will be shifted.

Export Q:
Runs from 0 to 8.
Sets a threshold for measurements to be exported over USB.
Value 6 prevents measurements with Quality 6 or lower to be exported on the map.
Only higher Q measurements are exported.

Note that every automatic export is indicated by the word "Export" in the bottom right corner of the normal operating (Hunting) screen.

You can always force a bearing export in hunting mode simply by rotating the encoder.

Export Window:
Runs from 0 to 8.
Sets the maximum deviation from the long term Average for measurements to be exported over USB.
Value 1 means that the latest measurement may only differ 1 degree from the Long Term Average to be exported to the mapping program.

Q-Squelch:
Runs from 0 to 8.
Sets the Quality thresholde for measurements to be displayed and to have any impact on the Long Term Average.
Value 3 means that only measurement higher than Quality 3 will be used.

Averaging Mode:
Runs from 0 to 255 in increments of 5.
Sets the MAXIMUM amount of averaging of the Long Term Average.
Higher numbers will stabilize the reading and give better suppression of modulation but will also make it react slower to direction changes.
Note that -due to the Adaptive Averaging- actual Averaging will be automatically changed within the range
1 up to the set Maximum, depending on circumstances.

Averaging Mode set to 0 will switch the RDF43 to burst mode, with no averaging at all and using a special algorithm to catch very short transmissions, down to 1/5th of a second.
In Burst mode, the Elevation Indicator is not available.
Also, Export of a bearing will occur much more often.

Azimuth Calibration:
Runs from 0 to 359.
Sets the amount of degrees added to a measurement to make the outcome fit reality.
So, with any array orientation and with eny radio, you can have the RDF43 point into the right direction simply by adding the right amount of calibration degrees.

Elevation Calibration:
Runs from 0 to 32.

Elevation indicator depends on the level of the doppler tone and therefore it depends on the audio output level of your receiver and the audio gain setting on the RDF43 interface board.
Elevation Calibration enables you to make the Elevation indicator point to 0 degrees (horizon) for earthbound signals.
Try to achieve about 1 or 2 degrees elevation on an earthbound signal to prevent over-compensation.

Rotation Frequency:
Runs from 128 to 1016Hz in increments of 4.

Value 500 sets antenna rotation to 500 Hz, that's 500 cycles per second.
This is a good value for most NBFM communication receivers.

After changing this parameter you need to re-calibrate Azimuth and Elevation, since a lot of parameters will be shifted.

In case of a 145MHz pseudo doppler array, 500 cycles per second will equal a rotational speed of more that 1000 meters per second. That's almost Mach 3, much faster than an F16 fighter jet can fly....



Antenna Drive:

Sets the antenna drive saturation.
A higher number means more on-time per antenna and therefore more overlap between antennas.
Listen for a nice round and smooth "doppler" tone out of your radio.
That's the soft spot for your array.
Generally, passive arrays perform fine with Antenna Drive 100.
Preamped arrays turn on and off at a significantly higher voltage so Antenna Drive will be best at 200 or higher.









Due to the fact that the RDF43 is a soft switching device, a high Antenna Drive number will result in substantial overlap between antennas, as illustrated in the picture, which is absolutely fine.

Underlap however (= gaps between antenna on-states) massively degrades performance.
So it is safer to go not too low with this setting.



After changing this parameter you need to re-calibrate Azimuth and Elevation, since a lot of parameters will be shifted.

Manual Antenna Rotation:

Stops antenna rotation.
Rotating the encoder will let you step through all 4 antennas, so you can check proper antenna performance.
Also listening to a station will be easier because the whining doppler tone will be absent.

Unlike the RDF41 and 42, the RDF43 display will show the currently active antenna.

Reflections may cause the 4 antennas to show quite severe differences in field strength.
This is normal.


Load Preset:

Runs from 1 to 32.
Loads all settings from EEPROM memory.
32 presets are available.

At power-up the RDF loads the last used Preset.

A 30-character text line is added to each Preset, so you can give a name or an explanatory text to each Preset.



Save as Preset:

Runs from 1 to 32.
Stores all settings in EEPROM memory, so they are not lost after reset or power-up.
32 presets are available.

A 30-character text line is added to each Preset, so you can give a name or an explanatory text to each Preset.

When you have selected the right preset number using the encoder, push it to save.

Having saved the preset, you will see a < cursor appear left below the text line.

If you push the encoder again you will skip the text editing mode and return to Menu.

Instead, turning it clockwise below a character and pushing it will enter editing mode for the character right above the cursor.

You will see the cursor change into a character and using the encoder you can scroll through all available characters.
Once you see the right character for that particular position, push the encoder to save the chosen character to that text position.

Available characters are:
! " # $ % & ' ( ) + , / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
and
SPACE

Turn the encoder to the far left position and push it 
to get back to the Menu.
Housing considerations:

A nice enclosure for the RDF is Conrad # 523232, measuring 103x56x168mm, offering enough room for all parts:





If you want to send bearings to a computer you should  keep the on-board USB connector close to the right hand side of the housing,
so the USB is accessable through an opening in the side panel. I simply drilled a 20mm hole in the side for that purpose.

The volume control only controls the loudspeaker volume.
In case you want to change the audio level into the RDF43 you can adjust the small blue trimmer on the interface board.
(At the component side)

The 12V dc input is protected against reverse polarity.
The voltage should be somewhere between 11V and 15V
The current draw will be between 180mA and 240mA depending on backlight level.
So a 300mA 12V dc power supply  will do fine. Or the 12V of your car battery.
Or a 12V gel battery used in emergency lighting units. (They are quite cheap second hand)
If you want to use a tiny lightweight battery you could pick a 3-cell Lithium-Polymer pack. A 1200mAH version would run the RDF42 for 5 hours or so.

Internally the RDF runs on 12V and 5V.
(Don't use the on-board dc socket of the Arduino!)

The antenna control outputs speak for themselves.
You need a cable and connector system of at least 5 conductors: 4 antenna control signals plus ground.
General current will be around 10mA so thin signal cable (cat-5) will do fine for cable runs up to 30m.
4 wires for the antennas, and the rest of the wires for ground.
The antennas themselves have to turn/run clockwise looking down on them.
If you discover that you hooked them up the other way around you just have to swap antenna 1 and 3  (or 2 and 4) to get it right.

If you really manage to mix them up you will see erratic behavior of the RDF...

A completely assembled,  programmed and tested kit comes with audio jack, speaker, potentiometer and encoder/switch including knobs.


The TFT color display:

Allthough the used color display is an absolute beauty, it has some serious drawbacks:

First of all, and most important, it absolutely can't compete with the monochrome display I normally use in terms of brightness:
Brightness is sufficiŽnt for indoor use in normal light conditions but application in a car in bright sunlight will be a challenge.
Plus the fact that it has a reduced viewing angle for best brightness doesn't help here either.
So in my opinion its use is limited to indoor applications like your shack, where its beauty will really shine.

For me as a designer it had its challenges as well since the standard model has no brightness control,
and it consumes around 150mA from the 3,3V power supply of the Arduino, overheating the Arduino's voltage regulators.

I had to take one apart to find solutions for that, and I found an elegant way to feed the background LED's using a separate adjustable power regulatur.
So that solved both issues: Brightness control plus the overheating of the Arduino.
However, experiments with the sacrificed display learned that the background LED's shouldn't be driven beyond approx. 160mA.
beyond 200mA they don't produce any more light than at 160mA and they slowly start to look blue-ish, indication over-temperature.
So 150mA is the safe maximum I use in the RDF43color.

Additionally the higher resolution of the display plus the fact that you need to un-draw a text/drawing in the background color to erase it and then
draw the new content in the foreground color make it a bit of a sluggish process to refresh a page.
(In the monochrome display you simply write the complete content to it and it automatically erases the screen lightning fast before drawing the new content.)

Another drama was the poor performance of the resistive touchpads on these displays.
They suck badly so I decided to completely remove them and return to the proven rortary encoder concept.

Mechanically, the display is quite vulnerable and it has no frame.
So I had to call in a friend who designed and 3d-printed a nice bevel/frame for it.  Thanks Danny!

Overall, I managed to work around above handicaps and I am quite happy with the end result, which looks far more beautiful than I can catch in a photograph.


Connectors:

I use the following connector standards when I build a RDF43 in a housing:
For DC power I use a 5,5/2,1 mm bus, center pin is +12V.
Audio input: 6,3mm bus, tip is hot.
Antenna control: 5-pin DIN bus, antennas wired as in below picture. (antennas are always numbered clockwise, looking down on the array)
A Remote Encoder bus is optional.





The following table gives a good starting point for your personal settings:

Mobile work, constant signal Fixed site, constant signal Hunting intermittent signals very short bursts
Backlight Depends on light conditions Depends on light conditions Depends on light conditions Depends on light conditions
Digital Filter 3-7, depends on sampling method  3-7,depends on sampling method 1-3, depends on sampling method 1
Gated Sampling Depends on receiver audio output Depends on receiver audio output Depends on receiver audio output Depends on receiver audio output
Export Q 5 7 3 3
Export Window 1 1 1 1
Q-Squelch 2 2 2 2
Averaging Mode max 255 max 255 max 255 0 = Burst Mode
Azimuth Calibration Depends on array orientation,
0 degrees should be the front of the car.
Depends on array orientation,
0 degrees should be north.
Depends on array orientation Depends on array orientation
Elevation Calibration Approx. 12, depends on receiver output Approx. 12, depends on receiver output N/A N/A
Rotation Frequency 500Hz 500Hz 500Hz 500Hz
Antenna Drive 255 for preamped array, 100 for passive arrays, 50-80 for passively switched Amplitude arrays.
Antenna Test N/A N/A N/A N/A
Save as Preset N/A N/A N/A N/A
Load Preset N/A N/A N/A N/A

System testing:

For testing purpuses you may take the antenna driver outputs as testing signals.
Use a 220k series resistor to feed an antenna driver signal into the audio input.
If you calibrate to 0 degrees on one driver output, the others will show 90, 180 and 270 degrees plus or minus 2 degrees averaged.
If this is the case you know the RDF43 itself works like a charm.
Even if one of the antenna drivers should have a 4 degrees error that would pose no problem:
When hunting a normal signal, the error would be smeered out over the full circle and therefore have a 1 degree impact on the bearing at max.
That's also true for a comparable deviation in one of 4 antennas.


73, Wil.

Back

/>