Question about VHF ELT bandwidth and homing

[India Four Two]

I could have posted this question on the Tech Log, but I thought it might be of interest to the PPRuNe private pilots, given the low-tech homing method.

On Wednesday, I flew as a spotter on my second SAREX with CASARA (the Civil Air Search and Rescue Association). Luckily the minus thirties deep freeze from last week had been replaced by a Chinook (wind, not helicopter!), so the temperatures were above freezing. We did have to deal with 35 kt winds and moderate turbulence when in the Rocky Mountain foothills, but nothing too unpleasant.

The exercise involved looking for an imaginary missing aircraft, which was simulated by ground teams in unknown locations, somewhere along the flight planned track. We flew search patterns in three different areas, as requested by the CASARA office at Calgary International. The ground team laid out fluorescent panels at the "crash site" and operated a VHF ELT, using one of several specially licenced CASARA frequencies - 123.1 in this case.

While searching (at 1000' AGL), we usually picked up the ELT signal before we spotted the ground panels. Upon hearing the ELT, we would carry out an "aural null" procedure, which involved marking the Foreflight map with the location where the signal started and then where it stopped. We then returned to the midpoint and turned at right angles and carried out a similar procedure. This gave us a rough circle where the ELT should be. However it was typically about 10 miles across and therefore it was still difficult to pinpoint the location. However the navigator (in the right seat) was very experienced and he showed how you could check for the ELT signal on a frequency 50 kHz offset from the ELT's frequency - 123.15 in this case. When the signal showed up on this frequency, we were always within half a mile of the "crash site" so it was easy to pick it up visually.

For any radio experts out there, how does this work? My understanding is that the bandwidth for a VHF channel is only a few kilohertz either side of the centre frequency, so we shouldn't be receiving it 50 kHz off the tuned frequency.

PS The Chinook was generating excellent standing waves, and ridge lift in the foothills, which was an interesting experience for our pilot, who had no glider time and wasn't used to flying low in these conditions. I'm going to remedy that in the spring.

Looking northwest. High level lenticulars above 20,000' and low level wave bars and rotor at 8000-10000' (4000'-6000' AGL):

[First_Principal]

Originally Posted by India Four Two

However the navigator (in the right seat) was very experienced and he showed how you could check for the ELT signal on a frequency 50 kHz offset from the ELT's frequency - 123.15 in this case. When the signal showed up on this frequency, we were always within half a mile of the "crash site" so it was easy to pick it up visually.

For any radio experts out there, how does this work? My understanding is that the bandwidth for a VHF channel is only a few kilohertz either side of the centre frequency, so we shouldn't be receiving it 50 kHz off the tuned frequency.

Interesting information, thanks.

The phenomenon you experienced is due to the passband characteristics of the receiver, and the effective 'width'. of the transmitter signal.

While nominal channel spacing in the aviation band used to be 25kHz (and for many is now 8.33kHz) this doesn't mean that you'll hear nothing if you're say 5kHz away from a particular frequency - particularly as you get closer to a transmitter.

The reason for this is that receivers and transmitters aren't perfect. The width of the transmitting signal, and the receiver passband filter (ie. how 'wide' it will receive) is measured on a scale (in dB) and is declared to be xHz wide typically when it reaches the -3dB point, however below this point - and progressively further away from the centre point of the filter - it will continue to pass signals, albeit at a lesser level or 'signal', until the 0dB point is reached.

This diagram, courtesy analogictips.com, shows pretty much everything:



As you'll see in this case the bandwidth is declared at the -3dB level, but the filter is actually wider than this and signals will pass further from the centre frequency, albeit at reduced levels. Different filters will have different characteristics and therefore their slope can be steeper, or more shallow. I'd suggest that as you go back in age of equipment the passband filter slope will become shallower (wider) and so you could possibly receive signals further off frequency than a more modern receiver. In addition to this both transmitters and receivers, especially older units, may not be working perfectly on frequency and could be some way from where they're meant to be.

So, given that the transmitter will actually be transmitting wider in frequency, but at reduced signal, than its declared bandwidth, and the receiver will receive signals wider than its declared passband you can in fact be quite a few Hz off-frequency and still hear something. Obviously this will be down one side or other of the filter slope and so the signal is effectively attenuated, which means you need to be correspondingly closer to the transmitting source in order to hear it. Bear in mind that both the transmitter and receiver have similar characteristics (ie. the transmitter will transmit further from its centre frequency at reduced signal, and the receiver will receive further from its centre frequency at reduced sensitivity) and so the two will interact to give you what you'll hear.

Coming back to your experience - in other words going off-frequency on the receiver 'weakens' the signal as if you were quite a distance from it, meaning you can effectively get quite a lot closer to the transmitting source before it registers on your receiver, thus reducing the limits of the search area. Quite a useful technique, and a good thing for people to know - top marks to your navigator.

FP.

[India Four Two]

FP,

Thank you very much. Your explanation is very useful. I'm very familiar with filters but in a different frequency range, from 10 to 100 Hz - I work with seismic data in the oil exploration.

Is there a specification for the filter slopes (dB/octave) of an aeronautical VHF channel?

[First_Principal]

Originally Posted by India Four Two

FP, Is there a specification for the filter slopes (dB/octave) of an aeronautical VHF channel?

Unfortunately I can't answer that question directly, my post was largely from first principles and it's been many years since I worked on this sort of thing. A quick search brought up a FAA AC which wasn't all that useful, I suspect the quickest way of getting some idea might be to look at the specs of modern transceiver - Icom for example says "Selectivity More than 7.5 kHz/-6 dB, Less than 25 kHz/-60 dB", but what - if anything - is mandated I'm not sure. Otherwise, with luck, an avionics person might pop up with the answer - although I doubt it's the sort of thing many will have come across let alone needed to know.

Otherwise good to hear you're familiar with filters, one is never sure of the audience when spouting off about things here so tend to over-explain or over-simplify sometimes...

Re your field of expertise, FWIW I developed a triggered XYZ analytical seismometer for assessing and displaying building responses during earthquakes, about 5 years ago. This used FFT + Hanning filter, converted the time domain data into the frequency domain, integrated as necessary then delivered the plotted results to those that wanted to know. I never quite got the displacement calcs right so if you're an Omega arithmetic guru I'd be pleased to hear from you!

FP.

[India Four Two]

FP, re your last paragraph, I’m intrigued. I’ll PM you.

[9Aplus]

That tuning off frequency act like sort of attenuator.
Device usually present in "real" dedicated homing receivers.

The method you are using is the standard one for "passing by" aircraft, or "better than nothing" method.
Now days it is not necessary to use expensive high end dedicated equipment 30-40 k USD value,
because even the modern "el chipo" SDR multi channel phase coherent receivers can do the job.

Here example on one of recent examples funded by Crowd Supply https://www.crowdsupply.com/krakenrf/krakensdr
Device can be fit within one pelican case, running on own battery power, antennas on top,
all strapped to the back seat bench - not necessary to do any major (to much paperwork) modification of aircraft.
Personally I tested (on purpose & for fun) 2 channel SDR called LimeSDR USB,
and even that is much better than standard Air Band transceiver use.
SDRs are digital toys so even direct decoding of 406 MHz ELTs data is possible. On the other hand the search for cell phones
out of range (of standard base station) is possible too. All with proper SAR licensing at local national level agency.
Search approach is to mimic real cell base station, that lost cell phone is happy to answer in case the power is still present.
The search pattern in that case must be for example within 30 km distance, because of timing limitations of that network which is theoretically 37,5 km.
Just my 0,02 $ & have a fun...

[megan]

Back in the days when we operated Vampires, Sea Venoms etc era aircraft I was issued a brand new Kiowa with the very latest digital tuning radios and we had great comms problems initially, checking by the techs found them to be spot on freq, no faults, then found if we tuned a step up or down we found a freq that worked, the stations we were trying to communicate with were old tech and their freq standard were rather more lax.

[Whopity]

Back in my RAF days we used to practice aural homings on beacons on 121.3 using the same technique as you describe. I recall winning the Lord Trophy a RAF Search an Rescue competition that was held annually. Our homer only worked on 243 MHz so anything on 121.5 had to be done manually. The diameter of your circle seems a bit low from what I recall, but the altitude was probably lower than we used. I doubt that the bandwidth of an ELT is particular low and could easily be 25 Khz, quoted bandwidths refer to the half power point and at say 60db down will be very much wider. By selecting the receiver to the next channel there is a fair chance that if the ELT signal is strong enough some of it will be picked up. It is equivalent to placing an attenuator in the antenna and working with a weaker signal so that close in it will be more sensitive resulting in a better overhead estimate. Sounds like a neat trick discovered by doing it quite a few times.

[India Four Two]

Thanks everyone for your input.

I participated in another SAREX on Saturday and the off-frequency technique worked perfectly yet again. On-frequency we defined a circle a few miles across, at 1000 feet AGL. We then started searching in the centre of the circle, using a 50 kHz offset and we picked up the signal within half a mile of the target ELT. It was fairly easy to find since we knew it would be on or close to a road. We were cheating a little bit!

9Aplus, thanks for the link to the SDRs. I’m going to look into that.

I am now a fully qualified Spotter and I’m on the list in case of a real call out. These typically happen about twice a year.

[India Four Two]

I was searching for an old post of mine and I came across this thread.

Following my last post in February 2022, I was a member of a ground party that went out east of Calgary to lay out the "wreckage panels" and operate an ELT on 123.1 MHz. We had a VHF portable for communicating with the searching aircraft.

I discovered that within 10 to 20 metres of the ELT, I could receive the very distinctive ELT signal on all VHF frequencies from 118 MHz to 137 MHz!

https://www.facebook.com/watch/?v=308409280982883