Picture by ebemed
We've attempted the train torture test — seeing how well a 3G Internet connection works on a moving train — twice on a journey from Townsville to Brisbane, and once on the Eurostar from London to Paris. Having covered these relatively out-there locations, it seemed to make sense to run a similar test between Melbourne and Sydney, using the twice-daily XPT service that runs between the two state capitals.
My testing of this journey last Monday coincided with my switching from using Vodafone as my main source of mobile broadband to Telstra's Next G. I've always thought the Vodafone service offered a reasonable price/data combination, but the quality of the software has always been questionable, the network coverage seems to have slipped in the past year, and the roaming capacity — one of the reasons I chose Vodafone in the first place — turned out to be completely useless in practice. So I've switched to a prepaid Next G dongle, figuring that I'll get better regional coverage (and avoid paying for use when I'm overseas).
One of Telstra's big selling points for Next G is how good the coverage is across Australia, and that certainly seemed apparent on the first Townsville test I ran, which covered some seriously remote locations. I figured that on a journey from Melbourne to Sydney, there wouldn't be terribly many areas where signal wasn't available. However, this proved to be far from the case.
Within half an hour of leaving Melbourne, the signal had disappeared completely, and it remained absent for much of the cross-Victorian journey, which covers the first three hours of the trip. Victoria is Australia's most densely populated state, so it seems hard to reconcile Telstra's claim of covering 99% of the population with that lack of service. Unreliability continued throughout the NSW journey, with one period of more than an hour with no signal somewhere in the vicinity of Harden. Even within the greater Sydney region, Next G was absent as often as it was present.
I'm aware of all the caveats: a moving train isn't the optimum use case. Even assuming that's true, I was surprised to find myself on several occasions stopped on train platforms with no signal available either. You can add an extra antenna for low-signal areas. Sure — but why would that be necessary near Sydney? The network isn't necessarily optimised for data. Perhaps not, but there was so signal on a Next G-equipped BlackBerry either for making calls either.
The Next G software does handle drop-outs much more elegantly than Vodafone's software, and I ultimately got enough connectivity to get my work done. But it made me rather more sceptical of Next G's claimed coverage than I've been in the past, and I plan a few more journeys into rural (but populated) areas to see how well it holds up.
Some other side observations on the XPT journey itself. There's plenty of working room in a first-class seat, but the positioning of a raised ridge to hold a coffee cup on the fold-down table means that anyone with a moderately-sized laptop would have trouble balancing it. (Seats 1-4 in those carriages have a more substantial table, so would be worth hunting down if you want to do a lot of work.)
I'd really have liked a power outlet at the seat, but can't imagine that's going to happen any time soon. There's plenty of overhead luggage space, though the train itself was less than half-full in my carriage, which doubtless helped matters. Overall, while I like train trips more than the average person, I don't know that I'll be doing this particular 11-hour marathon again in a hurry — but I will be putting the Next G system through a few more tests.
Lifehacker Australia editor Angus Kidman notes that flying from Melbourne to Sydney would probably have cost much the same. His Road Worrier column, looking at technology and organising tips for travellers, appears each week on Lifehacker.