The RTCA document quoted above (and hereafter) gives a good deal of contextual information and of explanation on how interferences can jeopardise radar altimeter operation.
However is does not give much numeric information and it is considered as discredited by some stakeholders.
https://www.rtca.org/wp-content/uplo...Altimeters.pdf
The following document from the ITU gives more numeric information and its publication date (02/2014) makes it more resistant to discredit.
https://www.itu.int/dms_pubrec/itu-r...2-I!!PDF-E.pdf
I'd like to highlight 3 pieces of information contained therein :
- the ITU recommendation for RA filters (ITU-R SM.337) : 24 dB per octave (or 80 dB per decade) with a maximum of 40 dB ; this is a 4th order filter (or four stages filter).
I checked the "paper" specifications of 2 radar altimeters ; they comply with the 80 dB per decade recommendation.
- the front-end overload level of tested actual radio-altimeters that ranged from -30 dBm (at best) to -56 dBm (at worst).
It means that these altimeters are overloaded with interference when their cumulated power is above 1 microWatt (1 10^-6 W) for the best device or above 2.5 nanoWatt (2.5 10^-9 W) for the worst, that power being measured after reduction by the input filter. (note: 0 dBm = 1 milliWatt, thus -30 dBm = 10^-3 x 1 milliWatt = 1 microWatt)
- the in-band sensitivity to interference noise is only 6 dB below thermal noise (probably caused by the large width of the band)
Let's put these numbers in perspective:
A signal from a 5G base station, at the frequency of 3.8 GHz, will overload the altimeter with -30 dBm overload characteristics if the plane antenna receives it with a power exceeding 2.2 microWatt. (I ignore cable losses).
For the worse device with -56 dBm overload characteristics, only 0.0056 microWatt will suffice to overload it.
If the same computation is made with a signal at 3.98 GHz, the best device will be overloaded with a power of 1.54 microWatt and the worse with a signal of 0.0038 microWatt.
So, allowing the frequency to step from 3.8 GHz to 3.98 GHz has the same effect as allowing the emitted power to increase by a factor 1.45 or it can be balanced by reducing the emitted power by that ratio 1.45.
We also see that the quality of the high frequency operational amplifiers used within or after the 4 stages filter is playing a much more important role than the cut-off frequency of the 5G band ; the factor between the best and the worst device is 400 !
[WRONG]
A typical 4G antenna emits a 40W signal which gives slightly under 1 mW/m2 (1 milliWatt) of radiated power at 100 meters (note: the signal is not emitted in all directions of the full 4pi solid angle).
If a 5G antenna emits with the same power, a signal at 3.8 GHz is 500 times too powerful for the -30dBm radio-altimeter placed at 100 meters and about 200,000 times too powerful for the other one.
However, at 2.5 kilometers, it becomes acceptable for the best device.
[/WRONG]
Apologies: the paragraph above is wrong : I mistakenly made calculations that assume a 1 square meter antenna for the radar altimeter. That's way too much. I found RA antennas with surfaces of 0.01 m2 and 0.025 m2. For calculation examples, I'll associate the larger antenna with the RA having an overload threshold of -30 dBm and the smaller one with the RA having -56 dBM,
If a 5G base station emits a 40W signal at 3.8 GHz that is received by a radar altimeter at 100 meters distance, the power transferred through the antenna is about 25 microWatt which is a bit more than 10 times too powerful for the first radar altimeter. Increasing the distance to 300 meters or more will reduce the signal below threshold. For the second altimeter, the power transiting through the antenna is around 2000 times too powerful. For that one, the distance needs to be increased to 5 km. So, the size of the antenna can explain a part of the overload characteristics, but not all of them.