In article .com,
TheOneKEA writes
(1) The NR rules about transition between 3 and 4 aspects don't
apply. On NR, the two cases should not be mixed and there are
strict rules about how to handle the transition. On LU they are
mixed randomly.
I was already aware of this from a thread RPM had posted in previously.
What I'm more curious about is how the transition logic is handled -
what's the aspect sequence when a driver is confronted with
three-aspect pegs, then four-aspect pegs, then three-aspect pegs again?

On LU, I am not aware of any specific rules.

On NR, it goes something like this. A driver should have a clear idea of
when changing from one area to the other. Therefore there should be a
clear run of several signals of a given type. Acceptable exceptions:
- The signal starting from a terminus or bay is usually 3 aspect even
on 4 aspect routes.
- An extra signal may be inserted into a 3 aspect run just before a
station, and this will require the signal in rear to be 4 aspect.
- When changing from one route to another (diverging or converging
junction, or crossover between slow and fast lines) the driver can
be expected to realize that the system is changing.

The transition from 4 to 3 is simple:

|-O |-O |-O |-O |-O |-O
-------------------------------------------------------
Y R train
YY Y R train
G YY Y R train
G G YY Y R train
G G G G Y R train
G G G G G Y

The only apparent oddity is that 3 aspect signalling comes in *before*
the signal spacing widens out, but at each signal it is clear to the
driver what is meant and the driver always gets sufficient warning.

The transition from 3 to 4 is harder, because of a principle that the
same signal should not use both Y and YY for *start braking here*. So
the "natural" aspect sequences can't be used:

3 7 11 13 15
|-O |-O |-O |-O |-O
-------------------------------------------------------
Y R train
G Y R train
G YY Y R train
G G YY Y R train
G G G YY Y
G G G G YY

because at signal 7 the driver gets Y as a first warning *and* YY as a
first warning, the latter giving 1.5 times the necessary distance to
stop but no marker at the braking distance.

One solution is to make signal 11 approach-released when 13 is red: 7 is
then a 3-aspect signal. This solution is often used where the speed
restriction isn't a problem (e.g. where 11 is the junction signal); it
also means that 11 can be a 3-aspect signal as well (because even if 13
is single yellow a train will have sufficient braking distance after
starting from near-rest):


3 7 11 13 15
|-O |-O |-O |-O |-O
-------------------------------------------------------
Y R train
G Y R train
G Y RY R train
G G RG Y R train
OR G G YY Y R train

(RY means approach-release to yellow).

The alternative approach is to add an extra signal (9 in this case):

3 7 9 11 13 15
|-O |-O |-O |-O |-O |-O
-------------------------------------------------------
Y R train
? R ? train
G YY Y R train
G G YY Y R train
G G G YY Y R train
G G G G YY Y
G G G G G YY

Note that signal 7 can never show single yellow and signal 9 never needs
to show red (it will therefore either carry a triangle on the number
plate, or be numbered 11R). Its aspect may either ignore the train ahead
of it completely, simply tracking that of signal 11, or it may be
controlled to single yellow when there is a train between 9 and 11.

(2) LU has fog repeaters. In multiple-aspect areas, a fog repeater
shows
- yellow if the signal being repeated is red
- white if the signal being repeated is not red.
Does the yellow aspect still have the big black F burned into the lens?

I believe so.

--
Clive D.W. Feather | Home:
Tel: +44 20 8495 6138 (work) | Web: http://www.davros.org
Fax: +44 870 051 9937 | Work:
Please reply to the Reply-To address, which is:

Clive D. W. Feather wrote:

On LU, I am not aware of any specific rules.

So, following your rules as posted, a 165 approaching Moor Park NB with
a Met in front would see this:

3 3 4 4 4 4 4 4
A858 A860 A864 A866 A868 A870 JJ48 JP47
|-O |-O |-O |-O |-O |-O |-O |-O

R A60
Y R A60
G Y R A60
G G Y R A60
G G (!) YY Y R A60
G G G YY Y R A60
G G G G YY Y R A60
Moor Park

Note that a green 3-aspect signal (A860) can be followed by a double
yellow 4 aspect signal OR a single yellow 4 aspect signal (A864) - if
I've read your post correctly, on NR this is a Very Bad Thing.

Unless A860 is held to yellow if A864 is single yellow? Or have I
managed to misunderstand?

snip/

Does the yellow aspect still have the big black F burned into the
lens?

I believe so.

Thanks.

In article .com,
TheOneKEA writes
So, following your rules as posted, a 165 approaching Moor Park NB with
a Met in front would see this:
[...]
Note that a green 3-aspect signal (A860) can be followed by a double
yellow 4 aspect signal OR a single yellow 4 aspect signal (A864) - if
I've read your post correctly, on NR this is a Very Bad Thing.

Correct; NR rules don't allow this (on continuously signalled routes; it
is allowed on an intermittently signalled route).

Unless A860 is held to yellow if A864 is single yellow? Or have I
managed to misunderstand?

No, it sounds like you've got it exactly.

But, as I said, LU does it differently.

--
Clive D.W. Feather | Home:
Tel: +44 20 8495 6138 (work) | Web: http://www.davros.org
Fax: +44 870 051 9937 | Work:
Please reply to the Reply-To address, which is:

Clive D. W. Feather wrote:

Unless A860 is held to yellow if A864 is single yellow? Or have I
managed to misunderstand?

No, it sounds like you've got it exactly.

But, as I said, LU does it differently.


True. I'm just curious as to what the circuit that links the HR/HHR
relays at A864 with the GR relay at A860 looks like.

On 4 Feb 2005, TheOneKEA wrote:

Clive D. W. Feather wrote:

Unless A860 is held to yellow if A864 is single yellow? Or have I
managed to misunderstand?

No, it sounds like you've got it exactly.

But, as I said, LU does it differently.

True. I'm just curious as to what the circuit that links the HR/HHR
relays at A864 with the GR relay at A860 looks like.

I'm curious as to what that sentence even means!

There's something strangely reassuring about reading dense technical
conversations that you don't even remotely understand. It tells me that
there is someone, somewhere, who knows what they're talking about ...

tom

--
POTATO POWER IS UNTRACEABLE POWER

Tom Anderson wrote:

I'm curious as to what that sentence even means!

In a 4-aspect signal, there are three relays:
- the GR relay, which controls the switch from green to double yellow
and back
- the HHR relay, which controls the switch from double yellow to single
yellow and back
- the HR relay, which controls the switch to single yellow to red and
back again

Under normal circumstances (i.e. an auto signal on a TCB line), the GR
relay will be energised, thus making the signal head show a green
aspect. When a train crosses the overlap into the track circuit
connected to the signal, the circuit is shunted, causing the GR relay
(and additionally, the HHR relay) to drop out and the HR relay to be
energised, thus switching the signal from green to red. After the train
has crossed the overlap of the next signal and moved off of the track
circuit, the HR relay drops out, causing a circuit connected to the
next track circuit to energise; this is what lights the single yellow
aspect. After the train moves off of that track circuit onto the next
one, the process is repeated with the HHR relay, causing the second
yellow aspect to be lit. Finally, once three track circuits are clear
between this signal and the train, the HR and HHR relays drop out and
the GR relay is energised, which finally lights the green aspect.

(hopefully this isn't completely and totally wrong...)


There's something strangely reassuring about reading dense technical
conversations that you don't even remotely understand. It tells me
that
there is someone, somewhere, who knows what they're talking about ...

Clive does - some would say that Clive knows everything about
signalling.

In article . com,
TheOneKEA writes
In a 4-aspect signal, there are three relays:
- the GR relay, which controls the switch from green to double yellow
and back
- the HHR relay, which controls the switch from double yellow to single
yellow and back
- the HR relay, which controls the switch to single yellow to red and
back again

Not quite. The HR controls the switch from red to non-red. Once yellow,
the HHR allows it to switch to double yellow, and then the GR to green.

+- not GR -- yellow --+
| |
+- HHR -+--- GR ----- green --+
| |
+- not HHR -+ |
| | |
+- HR ---+- not GR --+------- yellow --+
| |
+ve -+- not HR --------------------- red ---+-- -ve

Under normal circumstances (i.e. an auto signal on a TCB line), the GR
relay will be energised, thus making the signal head show a green
aspect. When a train crosses the overlap into the track circuit
connected to the signal, the circuit is shunted, causing the GR relay
(and additionally, the HHR relay) to drop out and the HR relay to be
energised, thus switching the signal from green to red.
[...]
(hopefully this isn't completely and totally wrong...)

Um, I'm afraid there are errors in there.

Let's consider a simple bit of track:

1 3 5 7 9
|-O |-O |-O |-O |-O
-------+-----A-----+-----B-----+-----C-----+-----D-----+----

Track circuit A connects to 1's HR relay (1HR).
Track circuit B connects to 3HR and 1HHR.
Track circuit C connects to 5HR, 3HHR, and 1DR.
Track circuit D connects to 7HR, 5HHR, and 3DR.

With all of A to D clear, signal 1 will be green. When the train moves
on to track A, signal 1 will drop to red. Moving on to B drops 3 to red
but, when it clears A, signal 1 can raise to single yellow. When it is
on C, 1DR is down but 1HHR and 1HR are up, so it will show double
yellow. Finally, when track circuit C is cleared, 1DR picks up and the
signal goes green.

If, say, there's a set of trailing points between 7 and 9, 7HR can only
pick up if:
- D is clear
- the signaller has set the route from 7 to 9
- the points are proven to be locked in the correct position.

With facing points, the HHR has to include logic to "look ahead" in the
correct direction.

There's something strangely reassuring about reading dense technical
conversations that you don't even remotely understand. It tells me
that
there is someone, somewhere, who knows what they're talking about ...
Clive does - some would say that Clive knows everything about
signalling.

Far from it.

--
Clive D.W. Feather | Home:
Tel: +44 20 8495 6138 (work) | Web: http://www.davros.org
Fax: +44 870 051 9937 | Work:
Please reply to the Reply-To address, which is:

Clive D. W. Feather wrote:

Not quite. The HR controls the switch from red to non-red. Once
yellow, the HHR allows it to switch to double yellow, and then
the GR to green.

+- not GR -- yellow --+
| |
+- HHR -+--- GR ----- green --+
| |
+- not HHR -+ |
| | |
+- HR ---+- not GR --+------- yellow --+
| |
+ve -+- not HR --------------------- red ---+-- -ve

This is the sort of thing I hoped to see - thanks :-)


Um, I'm afraid there are errors in there.

As I expected :-((


Let's consider a simple bit of track:

1 3 5 7 9
|-O |-O |-O |-O |-O
-------+-----A-----+-----B-----+-----C-----+-----D-----+----

Track circuit A connects to 1's HR relay (1HR).
Track circuit B connects to 3HR and 1HHR.
Track circuit C connects to 5HR, 3HHR, and 1DR.
Track circuit D connects to 7HR, 5HHR, and 3DR.

With all of A to D clear, signal 1 will be green. When the train
moves on to track A, signal 1 will drop to red. Moving on to B
drops 3 to red but, when it clears A, signal 1 can raise to single
yellow. When it is on C, 1DR is down but 1HHR and 1HR are up, so it
will show double yellow. Finally, when track circuit C is cleared,
1DR picks up and the signal goes green.

Thanks; I get it now. You should write another article for your site
that explains all of this.


If, say, there's a set of trailing points between 7 and 9, 7HR can
only pick up if:
- D is clear
- the signaller has set the route from 7 to 9
- the points are proven to be locked in the correct position.

With facing points, the HHR has to include logic to "look ahead" in
the correct direction.

So what does the circuit fiagram for this logic look like? Presumably
it involves the detector boxes on the point blades as well as linkages
to the HHR relays in the signals in advance of the facing points.


Far from it.


Maybe not everything, but certainly enough to seem like it...

In article .com,
TheOneKEA writes
Let's consider a simple bit of track:
[...]
Thanks; I get it now. You should write another article for your site
that explains all of this.

One day.

If, say, there's a set of trailing points between 7 and 9, 7HR can
only pick up if:
- D is clear
- the signaller has set the route from 7 to 9
- the points are proven to be locked in the correct position.

With facing points, the HHR has to include logic to "look ahead" in
the correct direction.

So what does the circuit fiagram for this logic look like? Presumably
it involves the detector boxes on the point blades as well as linkages
to the HHR relays in the signals in advance of the facing points.

No.

As soon as you put points or other issues into the system, it changes
from simple automatic signals to far more complicated controlled
signals. [The following applies to NR, not LU, though LU is similar in
many respects.]

There is a list of something like 35 items that have to be satisfied
before a controlled signal may clear from red (not all apply to every
signal, of course). The relay logic to implement this is completely
different to automatics.

The logic will be defined in terms of a "route", which normally runs
from one signal to the next over a specific bit of track. Conventionally
the routes from signal 5 are called 5A, 5B, 5C, etc. with 5A being the
leftmost one. [You also get the distinction between 5A(M) 5A(W) 5A(C)
and 5A(S) - (M) is a normal running signal, (W) is a "warner" route with
a reduced overlap, so the signal can only clear to yellow once the train
is stopped at it or almost so, (C) is calling-on into an occupied
platform, and (S) is shunting).]

The signalman attempts to select a route. Each route has a set of
requirements, such as certain points being either in the right position
or available to move to that position. It may also be incompatible with
other routes. If the conditions are all met, however, a route relay
(e.g. 5ARR) picks up. This triggers the movements of points and locks
out conflicting routes.

Consider something like this:

|-O\ 5 101 |-O 7
----+------A-----+-B--*---------+------C-------+-----E--I---
\ |-O 9
\-------+------D-------+-----F--I---

Once the points are all in the right position and all the track circuits
in the route are clear, a second relay - 5GR - can pick up. This is
driven by a circuit something like:

+ve - 5ARR --- E TR --- C TR --- 101 NKR -*- B TR --- A TR --[5GR]-- -ve
|
+ve - 5BRR --- F TR --- D TR --- 101 RKR -+

X TR is up when track circuit X is clear. 101 NKR picks up if points 101
are detected in the locked position - that is, both blades and the bolt
are all correct. The NKR and RKR relays are driven directly from
microswitches within the points machine or related mechanism, checking
the exact position of the blade.

So 5GR indicates that the route is clear. This in turn drives 5HR, which
we discussed in a previous message, through various other tests such as
the junction indicator being correctly lit as appropriate. [Again, I
omit various details.] Finally, to answer your question, 5HHR (which
allows 5 to move to double yellow or green) will be driven as something
like
5HHR = (5ARR AND 7HR) OR (5BRR AND 9HR)
That is, the route relay will be used to decide which signal to look
ahead to, while the "signal can clear to yellow" logic is used to check
the points are set correctly.

====

Having written this, I've just remembered something. Look into Google
last year for posts in uk.railway concerning platform 4c (I think) at
Birmingham New Street.

If I recall correctly, in some installations using what's called
"geographic logic", the HHR logic looks more like:
5HHR = (101NKR AND 7HR) OR (101RKR AND 9HR)

The problem at Birmingham was that an extra platform was added via an
extra set of points, as if we modified the above:

|-O|\ 5 101 |-O 7
----+------A-----+-B--*---------+------C-------+-----E--I---
\ |-O 9
\--*----+------D-------+-----F--I---
102\
\--+------Z-------]

The extra logic for 5CRR was all done correctly, but nobody remembered
to include 102NKR in the 5HHR logic. As a result, 5 would show double
yellow (or green, I forget which) if the route was set into the bay
while 9 was not red. Oops.

--
Clive D.W. Feather | Home:
Tel: +44 20 8495 6138 (work) | Web: http://www.davros.org
Fax: +44 870 051 9937 | Work:
Please reply to the Reply-To address, which is:

snip

blink

You really need to write more articles on this sort of stuff, so that
people like me will stop asking silly questions!

Thanks for the info.