Mode-S Beast:Performance

From Beast Wiki
Jump to: navigation, search

I am quite proud that the Mode-S Beast is around the best-in-class Mode-S units that is available, if not even that it is the best. I do not compare the Mode-S Beast against any of the PIC solutions or such without an ADC. So there is only one unit that I can compare it to, and during the last few month I had the chance to verify my development three times against another commercial unit.

During the first test in January 2011 I had both units connected to my G7RGQ omni antenna through a 6dB splitter.

Fpga-shot15-framerate-fpga-competitor-big.gif

Result: The Mode-S Beast (red) is providing around 15% more frames with the same signal.

Fpga-shot17-competitor-splitter-big.gif

I then let Planeplotter run for 30sec using the output data of the other commercial unit, and made a screenshot. The integration time in Planeplotters IO setting is 1min for all pictures, and please disregard the speed meters.

Fpga-shot17-fpga-splitter-big.gif

And same 30sec run immediately afterwards using the Mode-S Beast.
It is very impressive if you download both pictures and toggle it with a viewer like Irfanview.
Result: With the Mode-S Beast, there are much more planes displayed in the same time intervall. (Also read below)

A word to the two Planeplotter screenshots: I have to admit that I do not completly understand why there is such a big difference between the number of planes beeing displayed within the same time interval. I unfortunately did not compare the frames that both units have provided to Planeplotter, so I may explain that the 15% more frames of the Mode-S Beast are just from those planes displayed additionally. This plus with planes also happens in situations when both units show nearly same frame rate. It might also be that the software of the other unit does not forward all planes to the Planeplotter interface. Anyway, I do not have time to check this further, and finally I am happy that my Mode-S Beast is the better one.

The second test was the most intensive, because one of the users of the evaluation units was not able to reproduce my frame rate. It figured out that he used an magnet mounted antenna on the window sill, such antennas are quite often used. Luckily I got his complete equipment and so I was able to reproduce his issue, and finally solve it: At that time the receiver was not in a box and so there was feedback from the receiver to the antenna. I solved this by putting the receiver into a screening box, and also the unit shown above behaves as above. So be aware that if you are using close-by antennas, your receiver must be screened in order to give maximum performance.

Fpga-competitor-v10sn1-windowsill-big.gif

I swapped the antenna (mag mount on windows) between the two receivers while they were connected to ModeSCompare.
Result: Both units showing same number of frames.
(There are not more frames displayed by the Mode-S Beast, because there just were no more planes it could see from that position).

30sec-competitor-big.gif

Again, I did the 30sec run of Planeplotter connected to the other unit first and immediately afterwards to the Mode-S Beast. For this I used my G7RGQ omni on the rooftop. Low altitude traffic is masked here, and notice that total number of planes received was 140.

30sec-beast-v10-big.gif

And here I can show again more planes on the screen when running Planeplotter connected to the Mode-S Beast.
Also note that here I got a total of 167 planes.

Fpga-shot15-competitor-2-big.gif

The effect of the 2nd antenna can be shown here, when I connected the 2nd receiver with a yagi towards Munich airport to the 2nd port for a short moment during the second test.
The maximum DF-11#/DF-17 frame number ever seen with two antennas was 400 per second.

Mode-S-Beast-screened-big.jpg

The Mode-S Beast in a screened box that was used to solve the issue of the second test.
The small screening box above the receiver part did not solve the issue, but it is an example what will be needed for Bluetooth interface.
The box that is delivered with the box kit also does a sufficient screening.

Competitor-beast-v10sn1-omni crc-driver-big.gif

Here I swapped my omni anenna between the two devices under test.
The plus for the Mode-S Beast is less than in all tests, maybe there were no planes in the area that the Mode-S Beast is able to receive additionally during the test.
At the end you can see the effect of the CRC Bit error correction driver which improves the frame rate by about 40%.

The third test was done somewhere in northern europe, and I just swapped the antenna between the Mode-S Beast and the other unit under test. I had two Mode-S beasts with me, one development version V0.1 and one V1.0 engineering sample, both showed the same data.

Framerate-competitor-beast-northeu-big.gif

Again, there is a 15-20% higher frame rate shown for the Mode-S Beast.