Movilla wrote:
wrote in message
ups.com...
It's always been like this since the dot matrix indicators were
installed in the late-80s.

Fair enough. But how do LT work out their minutes?

Basically they were the estimated time for off peak trains (senior operators
did not like the idea of a train arriving "early").

The original signs (on the Northern Line) were fed from the central control
computer (at Coburg Street) data for each platform. This computer was only
reading states - not actually performing control (political reasons)
functions. I wrote the interface software from the computer to the signs.
When we commissioned a specific sign the guy who wrote the software
interfacing the train running data to my output process calculated the
approach time for a reasonable off peak timing and would then go (or send
someone else) to the appropriate platform (checking that the service was
running fairly well before he left the office) and time trains entering the
platform with those of us at the computer end making hand adjustments to
each sign. This allowed the local run in to be fairly accurate as long as
the approaching train was not held on the approach by a tardy departure by
the previous train. The computer was monitoring the position of trains by
track circuit occupancy. The times generated were usually off peak -
bearing in mind that the average interstation time on LUL tube sections is
less than 2 minutes, any time longer than 2 minutes was subject to the
vagaries of platform dwell time at intermediate stations as well as the
difference in timing due to loading and traction voltage variations which
were than common. The system was designed to blank the time for any train
which overdwelled and only restart the timing when the train moved again.
The system was only passive and was subject coping with signallers changing
train details or destinations to try and get service back on time.
Junctions were only monitored and it was almost coincidental that the run
from Camden Town to Euston was about 2 minutes. If the system was
projecting a train down a particular route it would count down on the Euston
signs until it got to Camden (2 mins shown). If it was held by the
signaller the time would blank and if he let the one from the other platform
go first this would be detected as the train left the platform and displayed
shortly thereafter (usually still as 2 mins). If the destination of the new
train had changed you would get the "CORRECTION" display. I thought (and
still do) that most regular travellers would much prefer to be told the
actual position of the train (e.g. "at TCR") because they would soon get the
hang of working out how long a train would take to arrive - but simple
displays were preferred by the operators.

The stand-alone system developed for most other lines was basically
independant of the signalling system using three trigger points (taken from
the signalling track circuits) which counted down from each timing point
with each timing point recalibrating the count (or holding/blanking, if
required). The train description was taken from whatever system was in use
on the line. This was still subject to the vagaies of train running and
platform dwells.

I can't remember what we did to the Victoria Line Identra System!

As the centralised computers became more of a real control system
(signallers using them for control functions as well as monitoring & info -
Met & Jubilee plus Bakerloo Lines at Baker Street) the signs controls were
updated (Met signs again came from the computer control points). The
Central Line displays were, I think, integrated within the computer system -
eventually full ato operation would be able to give much more accurate
arrival data.

PTI (Positive Train Identification) also had a role to play - this is the
transponder bit you may have been thinking about. This is a system which
identifies each train, its destination, number and crew at transponders
sited at critical points around the lines - it provides positive
identification at fixed points only which the other systems can use as
appropriate.

Fully integrated systems now being intalled should allow a much better (more
accurate) display.
--
Peter & Elizabeth Corser
Leighton Buzzard, UK



----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

On Apr 21, 4:34 pm, "Peter Corser"
wrote:
Movilla wrote:
wrote in message
oups.com...
It's always been like this since the dot matrix indicators were
installed in the late-80s.

Fair enough. But how do LT work out their minutes?

Basically they were the estimated time for off peak trains (senior operators
did not like the idea of a train arriving "early").

The original signs (on the Northern Line) were fed from the central control
computer (at Coburg Street) data for each platform. This computer was only
reading states - not actually performing control (political reasons)
functions. I wrote the interface software from the computer to the signs.
When we commissioned a specific sign the guy who wrote the software
interfacing the train running data to my output process calculated the
approach time for a reasonable off peak timing and would then go (or send
someone else) to the appropriate platform (checking that the service was
running fairly well before he left the office) and time trains entering the
platform with those of us at the computer end making hand adjustments to
each sign. This allowed the local run in to be fairly accurate as long as
the approaching train was not held on the approach by a tardy departure by
the previous train. The computer was monitoring the position of trains by
track circuit occupancy. The times generated were usually off peak -
bearing in mind that the average interstation time on LUL tube sections is
less than 2 minutes, any time longer than 2 minutes was subject to the
vagaries of platform dwell time at intermediate stations as well as the
difference in timing due to loading and traction voltage variations which
were than common. The system was designed to blank the time for any train
which overdwelled and only restart the timing when the train moved again.
The system was only passive and was subject coping with signallers changing
train details or destinations to try and get service back on time.
Junctions were only monitored and it was almost coincidental that the run
from Camden Town to Euston was about 2 minutes. If the system was
projecting a train down a particular route it would count down on the Euston
signs until it got to Camden (2 mins shown). If it was held by the
signaller the time would blank and if he let the one from the other platform
go first this would be detected as the train left the platform and displayed
shortly thereafter (usually still as 2 mins). If the destination of the new
train had changed you would get the "CORRECTION" display. I thought (and
still do) that most regular travellers would much prefer to be told the
actual position of the train (e.g. "at TCR") because they would soon get the
hang of working out how long a train would take to arrive - but simple
displays were preferred by the operators.

The stand-alone system developed for most other lines was basically
independant of the signalling system using three trigger points (taken from
the signalling track circuits) which counted down from each timing point
with each timing point recalibrating the count (or holding/blanking, if
required). The train description was taken from whatever system was in use
on the line. This was still subject to the vagaies of train running and
platform dwells.

I can't remember what we did to the Victoria Line Identra System!

As the centralised computers became more of a real control system
(signallers using them for control functions as well as monitoring & info -
Met & Jubilee plus Bakerloo Lines at Baker Street) the signs controls were
updated (Met signs again came from the computer control points). The
Central Line displays were, I think, integrated within the computer system -
eventually full ato operation would be able to give much more accurate
arrival data.

PTI (Positive Train Identification) also had a role to play - this is the
transponder bit you may have been thinking about. This is a system which
identifies each train, its destination, number and crew at transponders
sited at critical points around the lines - it provides positive
identification at fixed points only which the other systems can use as
appropriate.

Fully integrated systems now being intalled should allow a much better (more
accurate) display.
--
Peter & Elizabeth Corser
Leighton Buzzard, UK



I am not sure what the original question was based on. Regardless of
accuracy, the system can only give whole minutes. So if it shows 1
minute when the the train is exactly (based on distance) 1 minute
away, how long should it display 1 minute for? Just for an instant?
Till the train arrives? Or should it display 1 minute for the time
that 1 is the nearest number of minutes, ie from 1 minute 30 seconds
till 30 seconds, after which it displays 0?

In that case it would always display 1 minute from 1 minute 30
seconds. But it should display 2 minutes from 2 minutes 30 seconds to
1 minute 30 seconds. Did the original post imply that 2 minutes is
displayed for longer than this?

"MIG" wrote in message
ps.com...
On Apr 21, 4:34 pm, "Peter Corser"
wrote:
In that case it would always display 1 minute from 1 minute 30
seconds. But it should display 2 minutes from 2 minutes 30 seconds to
1 minute 30 seconds. Did the original post imply that 2 minutes is
displayed for longer than this?

That's what I meant. For example, the 2 minutes time is displayed for 1 min,
20 seconds. The worse example I've seen of stuck minutes has been on the
Circle line where a minute can be anywhere up to 2 minutes in real time.

Peter's reply explained a lot.