CSA: Mobile Phones and Fringe Coverage
.
I have this habit with mobile phones and cell coverage that when I see something that interests me, even if I am holiday, I have to see what I can find out about it there and then. Whilst on holiday in Cornwall I noticed the area I was staying had fringe coverage. I thought this was strange as I would have expected to find the Cornish village of Mousehole to have at least a microcell, given the popularity of this tourist attraction to see the Mousehole Lights at Christmas. I decided to conduct an experiment to see how various mobile phones would react under fringe coverage radio conditions. I used no special equipment, nor did I switch ON an network engineering software. The mobile phones were as any ordinary user would have them and the radio conditions with which they would be faced. Yes, I know I know, I can be a bit of an anorak at times.
1. The place I was staying was Duck Street, Mousehole, Cornwall. Accessed at one end by a no through road for cars to use a car park and Duck Street narrows to an alley for pedestrians and no access for cars. 
2. The place in Duck Street where I conducted tests is in a car park that has been marked with a black cross (X) in the photo above. The close proximity of clutter (housing and a warehouse) falls within the clutter range of 10m to 30m in line with propagation models for dense to urban areas (ITU-R P.1546-2)
3. The above image displays how far Mousehole extends and the terrain clutter, along with natural phenomenon.
4. This last image provides an approximate indication of how far the main town of Penzance is from Mousehole and the general area where the Masts were located..
Okay, so I have now laid out the background to the tests. For the tests I used three mobile phones, all with built-in antennas, which were a Motorola Pebble U6 (which I nicked off my wife, much to her annoyance), Alcatel BH4 735 and a Nokia 3210. I ran tests at different times of the day (morning, afternoon and evening) and with three battery charge levels (charge in battery nearly empty, charge in battery half full and fully charged battery). The GSM networks were Orange, T-Mobile, O2 and Vodafone. The test area as has been shown is at para 2, above.
.
An interesting factor I noted was that all network coverage there was fringe all day, so that was the first matter. The second, and of far more interest, was how the mobile phones reacted to the radio conditions profiling the phones when switched ON.
.
Motorola Pebble U6:
Displayed 'No Service Available'. No calls or texts could be sent or received.
.
Nokia 3210:
Displayed 'Emergency Calls Only'. No calls or texts could be sent or received.
.
Alcatel BH4 735:
Displayed 'One bar of coverage and, intermittently, no coverage'. Attempted made and received calls either rang and when answered no voice communication and/or call dropping. However, surprise, surprise I could send and receive text messages.
.
Now, bearing in mind all the tests were being conducted in the same area, yet varying results obtained indicated alot about the various sensitivity of the antennas for these mobile phones. Technically, the BER of 2% for the measurement where the received signal strength is at a standard -100dBm and that a c-value would not be obtained, theoretically, below the reference BTS1 (-105dBm). I hear, though, that because of the loose wording in the standard -112dBm has been noted in some cases. Surprising, yes, but not in the realms of fantasy as GSM defines a mask lower level received signal strength, during testing, of -120dBm.
.
These very basic practical tests I conducted opens the door, though, to considering what could be extrapolated from the results when dealing with mobile telephone call records and Mast usage. For instance, if I were to conduct radio testing at this location where a person said they were for a particular call or text and I used a Motorola test handset, I may not get a positive result and may report back that not even a text could be sent from that location. Where then might the mobile phone be suggested to be located? What if a mobile phone is put nearer to the scene of a crime than it really was? What might be an inference drawn from that?
.
What may seem an ugly, awkward problem being raised here, is indeed not as bad as it seems. It really requires taking pragmatic steps before going to site to find out what mobile phone the person was using at the material time and then formulate from there how the tests should be conducted. There are other considerations, of course, to be taken into account, as well.
.
Acknowledgement: the screen images were obtained using Google Maps:
http://maps.google.co.uk/maps?hl=en&tab=wl
