London Banter - View Single Post - Train-numbers on the LU

[email protected] · February 25th 04, 10:30 AM posted to uk.transport.london

In article , (Richard
Griffin) wrote:

writes:

(Adrian) wrote:
In normal Binary, 16 is 10000, but in BCD each decimal digit is coded
separately in four-bit Binary, so 16 would be 0001 0110 - which is
why
it's called Binary Coded Decimal.

Sorry, my mistake. I forgot about that. I'll have to look at the
programme machine roll and check what is used - I think it just uses
plain binary.

I have the answer, Roger: it is Binary-Coded Decimal.

A few years ago I went on an arranged visit to Cobourg Street Control
Centre and East Finchley Signal Cabin [1] and, being a keen sort of
chap,
was invited to write an essay on the visit for the members' quarterly
journal CHTbulletin.

[1] - A group tour run by Cravens Heritage Trains for some of its
members.

I have pasted into this article the relevant section of that essay [with
some minor clarifications -- Ed]. With reference to the question under
discussion, the answer is in the 5th line from the bottom:

---------- begin quote ------------------------

A word about programme machines. Each of these consists of a roll of
"Melinex" plastic with holes punched in it; as the day progresses, the
roll is spooled from one drum to the other past the reader. Saturday
and
Sunday information is located beyond the end of the weekday information
on the roll, so rewinding is done at 0200 _only_ on Sunday to Thursday
nights. Each row of holes relates to a given movement for one train
(the
machine we were shown was for departures from Edgware).

Around thirty "tracks" are arranged across the width of the roll, each
of
which may or may not have a hole punched in it on any given row; the
presence of a hole in any given track is detected electrically by one of
a row of contacts in the reader, and the Working Timetable information
thus encoded is transferred to the signalling system.

The first holes or tracks impart timing information in binary format
(with the most significant "hours" digit being 8, the machines working
in
a sort of crude double-twelve-hour format), down to half-minutes at
track
eleven. At the other edge of the roll is Train Number information,
whose
available digits are 100, 100, 40, 20, 10, 4, 2 and 1: this explains the
lack of 8s and 9s in Northern Line train numbers! In the central part
of
the roll is a destination code, and the all-important signalling tracks
(e.g., if a hole is punched in track 12 then the signal for "plat 3 to
SB" will clear).

---------- end quote -------------------------

We see that Train Numbers are encoded using bits 200, 100; 40, 20, 10;
and 4, 2, 1.

This allows Train Numbers to be encoded that lie within the range 000 to
377, so long as the digit in the units and tens column is neither an 8
nor a 9. There appears to be no way for the Programme Machine to store
and transmit a Train Number in the 4xx or 7xx series; I'm pretty sure I
asked about this at the time, and was told that as these are special
workings they would be entered manually from Cobourg Street anyway. (I
have a trio of rings of Train Numbers off a scrap 1959ts train, which
are
still set to "736" which was Golders Green to Ruislip depots.)

Yes - anything that is in addition to the normal timetable would be
manually entered by the signalman. This includes the additional paths such
as the 4XX which are shown for stock transfers (and don't run on a regular
basis anyway), the 310 and 330 which are used for the training train, plus
all the other sundry odd paths such as the AVI train, track recording
train etc. In theory, any number from 200 - 777 could be picked for a
path, although I suppose there is a logic to it. The training train
started off as 310, then I added 330 for the Saturday version.

As long as nothing else is running with the same number, then that
shouldn't be a problem. With the additional paths the signalman will
normally fit the trains in as and when to suit the service. E.g. the
training train has a specific timetable with a set time for each location,
but as far as I'm concerned the only relevant times on it are those when
it comes out of Morden depot in the morning and starts from the siding at
Golders Green after meal relief.

With "." being plain Melinex and "O" being a hole in it, we can
translate
the following examples thus:

. . . . . . . . == 0 0 0
O O O O O O O O == 3 7 7
. O . O O O . . == 1 3 4
O . O . . . O O == 2 4 3

(Only the Train Number section of the roll is shown here, with four rows
shown; the roll would scroll upwards or downwards relative to this
article. ISTR the Train Number holes are actually laid out with the
*least*-significant digit at the left.)

The only thing that I may have mis-interpreted is whether a *hole* or a
"non-hole" represents a binary "1"; seeing as I had to inspect my
photograph (currently buried somewhere in clutter-heap!) of the
programme-machine very carefully in order to determine what each track
was for, when writing the essay, I hope that it is unlikely.

When demonstrating the programme roll to trainees, I define the "1" as
being the hole. This matches up with everything. If the "1" is defined as
the space, then that will give all sorts of weird answers! The said roll
at Cobourg St. used to have the ID for each group of holes shown, but that
has now disappeared. I think the roll is from the S1 machine at Edgware
which controls the SB departures.

Roger

HTH,

--
Richard Griffin
http://www.squarewheels.org.uk/
"It seems that, nowadays, there is no NOUN that cannot be VERBED!"
-- Professor Griffin, University of Cambridge