London Banter - View Single Post

Clive D. W. Feather[_2_] · February 21st 13, 09:50 AM posted to uk.transport.london

In message ,
d wrote:
Bluetooth and WiFi in the 2.4 GHz band is limited to 10 mW and has other
restrictions on duty cycle and power density. Railway equipment in 2.446
to 2.454 is allowed 500 mW in a narrow channel ( 1.5 MHz). Given that
the application is a train passing straight over the transponder in the
four foot, I suspect that the WiFi signal won't even be noticed.

Thats all very well, but whats stopping someone shoving a bluetooth signal
through a linear amp to disrupt the comms? If you think thats a stupid thing
to do , well hackers tend to do stupid things.

It would have to be a big linear amp. These transponders are used with
the train over them, so if the idiot is standing 1m from the platform
edge he needs about a factor of 800 amplification to reach the strength
of the transponder signal.

In any case, the only purpose of these transponders is to confirm the
train's exact position. It works something like this. The train starts
at a known position, call it zero. It count wheel rotations to work out
how far it's gone. Now suppose that these have an accuracy of 2%. Then
after it's gone what it thinks is 200 metres, its position is somewhere
in the range 196 to 204 metres from zero. After 500 metres it is
somewhere between 490 and 510, and so on. It therefore makes all its
safety decisions based on whichever limit is the more conservative.
Whenever it passes over one of these boxes, that resets its knowledge of
the position. They're supposed to be no more than 200 metres apart, so
that means it is at most 4 metres wrong in its calculations. If an idiot
disrupts one box, that simply means that the maximum error increases to
8 metres before resyncing. But the signalling will be designed to take
this into account and provide wider margins around a train that hasn't
synced recently. (Also recall that there's a 25m static margin anyway
for safety.)

--
Clive D.W. Feather | Home:
Mobile: +44 7973 377646 | Web: http://www.davros.org
Please reply to the Reply-To address, which is: