SuperDF Sensitivity Explained

About SuperDF:

SENSITIVITY EXPLAINED

Some people who see the SuperDF promotional material or specifications have difficulty believing our sensitivity specification of "20 dB below the noise level in the receiver used." We explain here how this is possible. (Note: The SuperDF is protected by US patent 4,475,106. The features described here (and others) are explicitly covered by the patent.)

Figure 1. System Waveforms.

SuperDF uses two simple but very effective techniques to increase sensitivity. (Refer to Figure 1 while reading this.) We now present the basic theory to explain this.

As can be seen in the Figure (1st and 2nd waveforms) the DF information coming out of the receiver is a series of alternating polarity pulses (+, - volts). Notice that each pulse first goes to (say +), then a very short time later it goes to the other polarity (say -). This second peak is smaller than the first peak.
The first thing SuperDF does with the audio from the receiver is to pass it through what is called a synchronous detector. (It is synchronous with the switching of the antenna.) This detector is an electronic SPDT switch which acts in such a way as to invert every other pulse. (The inverted pulse comes from a gain of 1 inverting operational amplifier.) Coming out of the switch will be a series of pulses of all the same polarity; either all + or all -, depending on which SuperDF antenna is closer to the transmitter. (See waveform 3.) This signal also contains all of the modulation and / or noise which is present at the output of the radio itself. (Modulation and noise are not shown in the figure.)
The width of the first excursion of the pulse (at above 50% of peak value) is about 40 microseconds for a typical scanner receiver.
During 100% of the time the "normal" audio output is also present. For DFing purposes this "normal" audio is noise.
SuperDF uses another electronic IC switch (SPST) to pass only a 40 microsecond wide slice of the synchronous detected audio signal through to be filtered. We call this the Window Gate. This happens once for every time the antenna switches (once for each pulse). There is a pot inside SuperDF which has to be adjusted so that this switch closes at just the right time to pass the first peak of the DF pulse through to be analyzed. This happens about every 1250 microseconds. (400 Hz switches the antenna twice every cycle, which is 800 times per second, hence 1250 microseconds between pulses.)
The percent of time the switch is closed is:
```
        (40 / 1250) x 100% = 3.2%
   
```
Thus, SuperDF throws away (does not analyze) about 97% of the audio coming out of the receiver. Most of that is just "noise" (not useful for DFing). It also throws out the second (opposite polarity smaller) peak of the DF information pulse (see Figure). This second half, if allowed to get to the following low pass filter, would act to cancel out part of the DF information from the first peak.. Not much useful DF information is lost. By using this Window Gate switch, we improve the signal to noise ratio of the information we want by a factor of about:
```
        100% / 3.2% = 31.25.
   
```
This is a dB improvement of:
```
        10log31.25 = 10(1.49) = 14.9 dB
   
```
Even before any bandwidth restriction (done in the low pass filter following this Window Gate) we have developed a 14.9 dB improvement.
After the Window Gate switch the signal is passed through an RC low pass filter. This filter has two capacitor values selected by the FAST / SLOW front panel switch. The larger capacitor produces a low pass filter of about 1/2 Hz.
Assuming a receiver which has an bandwidth (RF plus AF) of about +/- 3 KHz, the low pass filter will reduce the bandwidth of the information passed on to the LED control circuit of about:
```
        3000 / 0.5 = 6000    Bandwidth reduction
   
```
This is a dB improvement of:
```
        10log6000 = 10(3.78) = 37.8 dB
   
```

When we add up both improvements we get:

        14.9 + 37.8 = 46.15 dB    (Signal to Noise Ratio improvement)

The typical "bandwidth" of human ears (for picking a single audio tone out of noise) is about 25 Hz. This has been written up in scientific literature and is well know.
The SuperDF listens with "ears" which have a bandwidth of about 0.5 Hz, thus it can "hear" better than a human by a factor of:
```
        25 / 0.5 = 50 times
   
```
This is in dB:
```
        10log50 = 17 dB
   
```
Remember, this is the improvement over human ears, which have a bandwidth of 25 Hz.

The improvement of the SuperDF over the Radio is 46 dB. But the magnitude of the "Noise" (modulation + receiver noise) is much stronger than the DF information (even at maximum DFing pointing error) by probably more than 20 dB. Thus the absolute performance of SuperDF is degraded to be much less than 46 dB below the noise level. This is consistent with the field measurement described in Example 5, below, when DFing was still possible at 22 dB below the noise level seen by the radio S meter.

Thus, we can see that the SuperDF has very good "ears" to hear the DF information, and we can understand how it is able to DF a signal which is so weak that we humans cannot hear any modulation, quieting, or DF tone.

This extraordinary sensitivity is supported by actual (2 meter band) field experiences.

Example 1

One hunter reported that he checked out our claims by attempting to DF mobiles which were 120 miles away (from Seal Beach to Santa Barbara here in Southern California). Both ends of the path where just a few tens of feet above sea level, and much of the path is over the ocean. This is what he did.

First, on one receiver, he tuned in a Santa Barbara repeater (2 meters) which is at a site considerably above sea level. With a second receiver connected to SuperDF he tuned in the repeater input frequency. He could actually get bearings consistently towards Santa Barbara on the input frequency whenever one of the mobiles there accessed the machine. When the repeater was not in use, there were no bearings.

Example 2

One hunting team went on one of our "All Day T-Hunt." (All day, as in 24 hours! It has to be inside the continental USA! One hunt had four transmitters located in 4 states!). They had a 4 element cubical quad beam mounted on one side of their car and the SuperDF mounted on the other side. They reported that when they could barely detect the signal with the beam that they could get a bearing with the SuperDF, even though they could not hear it on that receiver. Also the SuperDF bearings were very consistent with those from the beam. They did not try to take SuperDF bearings when they did not hear it with the beam. (They probably would have been able to.)

Example 3

One hunting team stated out from Orange County (Southern Calif.) looking for a transmitter. They got bearings which drew them more and more South. They ended up in San Diego, some 100 miles later, where two fellows had been having a long conversation on the (simplex) T-Hunt frequency. (Apparently the real hidden transmitter had problems.) The 6000 foot Santa Ana Mountains, in the Cleveland National Forest) are between where they started and where then ended up! (Note: I suspect that their initial bearings were either from knife-edge refraction from the mountain tops, or scatter from the high spots along the coast, and coming around the mountains.)

Example 4

The author started a T-Hunt not being able to hear the signal on the repeater input (a start anywhere hunt). Using the repeater output on a second receiver to verify signals on the input (a transmit on request hunt), SuperDF said to go South. 40 miles later the signal began to show a bit of receiver quieting, popping in and out . Another 10 miles before audio could be poorly made out. The transmitter was found near the coast line, just South of the Orange County / San Diego County border. Much of the foothills of the Santa Ana Mountains (Cleveland National Forest, 6000 feet high), and two ranges of major hills were between the transmitter and the start site! The initial (rough) bearing would indicate that the signal had traveled up the coast line, and scattered inland from somewhere around Newport Beach. There are hills in that area and further South.

One interesting aspect of this hunt was as follows. The start site was just a couple miles from the San Gabriel Mountains (5,000 to 6,000 feet high; Angeles National Forest), but shielded from it by a low hill about 200 feet to the North. After getting the initial bearing to the South, we went on the 210 freeway. Very quickly we were away from the hill, and the bearing swung towards the nearby mountains. We correctly analyzed this as signal scattering down upon us from the mountains, and continued South. We won the hunt, first to arrive! (Hidden Transmitter power was 25 Watts.)

Example 5

The author did an experiment using a calibrated attenuator between SuperDF and the antenna input of a mobile receiver with an S meter. A moderately strong repeater output signal was tuned in. Attenuation was inserted until the S meter stopped moving downward. It was not Zero and was randomly bouncing ever so slightly indicating it was still reporting signal plus mostly noise. DFing was still possible and quite precise. The amount of attenuation was increased another 22 dB, and the signal (then inaudible) could still be DFed to give a rough bearing (say +/- 30 degrees).

George R. Andrews
President

Contact

George R. Andrews (Russ, K6BMG)
BMG Engineering, Inc.
9935 Garibaldi Avenue
Temple City, CA
91780, USA

Voice 1(626)285-6963
Fax 1(626)285-1684 (24 hour automatic)
America OnLine: Grandrews
Web: http://members.aol.com/bmgenginc

(19 Sep 1996)

Send E-mail to grandrews@aol.com. (A message window will open.)
Return to Top Page
Return to About SuperDF.