Mobile coverage on trains really is pants

You thought it was just *insert your provider here*, but now we have numbers


Mobile survey coverage company GWS, the company behind The Reg's awesome Monopoly test, has numbers to back up what we’ve always suspected: mobile coverage on trains is pants.

In what it claims is the most rigorous train connectivity study ever conducted, GWS has dubbed Three best for voice, Vodafone best for 3G and predictably EE best for 4G, but it’s nothing for any of the networks to crow about.

During testing, one in three mobile internet tasks and one in seven voice calls attempted on commuter train routes failed. The survey only looked at London commuter routes, so there were no rural or topographically taxing places like the Highlands and islands. These should all be places with good coverage.

Petek Ergul, head of networks at Vodafone, recently told El Reg that she was aiming for 96 per cent coverage. The London rail routes would need to be well covered to achieve that and they are not, indeed they found 82 per cent geographic coverage.

Because GWS uses equipment that can measure Layer 3 radio traffic, there is much more detail provided about what goes wrong than could be provided using crowd-sourced information and GWS determined that 23.2 per cent of 3G data packets 37.2 per cent of 4G data packets didn’t make it to their intended destinations.

Figure 1: Packet data reliability: the big four operators on 3G and 4G

Data Type

3

EE

O2

Vodafone

3G

77.4%

69.8%

78.0%

82.0%

4G

35.4%

74.1%

65.9%

75.1%

Figure 2: How long it is likely to take commuters with 3G data to upload a picture/download a song/download a video?

Task

3

EE

O2

Vodafone

Upload time for 4MB pic (secs):

33

36.2

45.4

48.4

Download time: 4MB song (secs):

23.3

21

26.8

16.1

Download time: 4 min of HD vid (seconds):

41.2

42.4

46.8

41.5

Figure 3: how long it is likely to take commuters with 4G data to upload a picture/download a song/download a video?

Task

3

EE

O2

VODAFONE

Upload time: 4MB pic (secs):

16.3

10.8

16.3

9.5

Download of 4MB song (seconds):

10.8

5.8

7.1

7.7

Download time for 4 min HD video (secs):

16.2

7.8

11.2

12.4

By measuring the Ec/Io, the ratio of a single user's intended signal versus the undesired interference and noise from all sources, GWS was able to determine voice quality. The test team reported that commuters are likely to experience poor call quality such as sound breaking up when mobile operators use older 2G networks to route their calls.

Paul Carter, CEO of GWS, commented: “Leaves on the track, the wrong kind of snow, having to stand up all the way to work and back – commuters have enough to contend with without the kind of mobile connectivity problems we’re revealing today. It’s hard to believe we’re in 2014 and in a situation whereby a trained wizard would have a tough time getting a signal on the Hogwarts Express while it’s sitting in St Pancras.”

The testing throws down the gauntlet to the mobile operators.

The next interesting survey will be from Ofcom. This was originally promised in October but when we spoke to Ofcom we were told that it will now be published “in the autumn”.

GWS has produced an interactive coverage map of its findings here. ®

Similar topics


Other stories you might like

Biting the hand that feeds IT © 1998–2021