BSG@SCRC-STONY-BROOK.ARPA (Bernard S. Greenberg) (01/08/86)
Long message on NYCTA time indications.
[Someone please forward to MSB!!]
Date: 4 Jan 86 08:00:11 GMT
From: Mark Brader <msb@lsuc.uucp.arpa>
I enjoyed the item by Bernard S. Greenberg about the NYC signalling,
and I thought I'd describe the system in use here; it's similar, but
not *quite* the same. As in NYC the signals can be either one-head
or two-head,
Some ancient IRT signals are three-head, as I said, the number of heads
(in that scheme) corresponds to the possible number of diverging routes.
In the new scheme, there is (or was at the time I studied) one home signal
on the outbound (J1) track at East New York Station (Jamaica BMT) with
a -blue- light for the third diverging route!
and under a two-head signal there may be an additional
yellow light. But then, under any of the above, there may be a white
light; this is used for speed restrictions.
On the one-head, green, yellow, and red have the usual meanings.
On the two-head, they are rendered as green over green, yellow over
green, and red over red respectively. So far we are the same as NYC,
"new scheme"
except that (as far as I know) single and double red are the same here.
Do you in fact know it to be that there is no difference in indication
between R and R/R?
Under a double red there can be a FLASHING yellow if the signal is so
equipped. This is "calling-on" and means that there is a red ahead
and it is closer than usual to this signal.
Divergences are indicated by green over yellow or yellow over yellow.
Until today I would have said that these indicated divergence to the
right and to the left respectively, but I was just on a train that was
diverted to the extra track in Davisville station to have its last two
cars removed (as usual at end of rush hour), and this diversion to the
right was signalled with a double yellow. So now I'm confused. It is
NOT the case that double yellow is a divergence with caution, though;
any divergence is treated as a caution until the next signal is seen.
This is interesting.
Now the interesting part, speed restrictions. At the entrance to any
speed-restricted curve, junction, or downgrade, the first signal is
equipped with a white light, but no green in the upper (or only) head.
...
Over to you, Bernard. Does NYC use this method? Does anywhere else?
...
Rather than identify the substantial differences, I will describe all
the time aspects and indications in use in the NYCTA, as threatened.
"Creeping up to a signal" is a special case implemented by clever
spacing of trip, train, wheel, and electrical inter-section break (known
in the NYCTA as "IJ", or insulated joint. Every signal is at an IJ,
but NOT VICE VERSA.) I describe it further below,
so let us factor that out. It involves no special aspects or indications.
There are two forms of time control. GT (grade time) and ST (station
time). I will describe the former first. It is used to unconditionally
control speed. The basic idea is that some signal (although they appear
in strings of several) is designated a "GT" signal. That means that it
is normally "red" (which may be R/R if it is also coincidentally a home
signal), with whatever force (indication) "red" has for that type of
signal. The reason it is Red is because of a timer which must run its
course before ALLOWING the signal to act as if the timer were not there,
i.e., clear if blocks ahead (perhaps also interlocking conditions) are
suitable. Let us call this state of a signal, in which the signal
would be clear but for the timer not having timed out, the "GT State"
(invented adhoc terminology). Stop (tripper) management is not
affected by GT or ST; it is at a lower level of modularity.
There is a dedicated aspect associated with the GT State on both the
Old IRT and New signalling systems, it is the by-way-of-Toronto-mentioned
single pure white (-Fachsprache-, "Lunar white") under the signal.
Lunar white only shows under red or R/R, and with it indicates unambigously
the GT state.
I do not believe that newly-installed signalling uses the Lunar White
aspect. The "S" light aspect, which is not found on "Old IRT", is used
instead. This works as follows: When a signal is in the GT State, the
preceding signal, if clear, must show orange (proceed to stop at next
signal), because the GT State is by definition, Red. When this scheme
is used, a clear signal, ONE IN ADVANCE of a GT Signal that is in the GT
State, will show "Orange over a lit-up white letter "S". If the GT
signal is red for other reasons than the timer, it will not show this
"S".
Regardless of the lunar white or S, the procedure for clearing a GT
Signal in the GT State (the only interesting case) consists of
spending enough time in the BLOCK or BLOCKS before it. There are
two cases of this:
1) a single "timing section", going from the GT signal backtrack to
one specific IJ. The timer starts running when the train crosses
the IJ into the timing section, regardless of GT signal state.
2) Multiple "timing sections", one in front of the other. By some
logic I have never understood, they all have their own timers, and
spending the required time in ANY clears the signal (iff was in the
GT State).
Note that in the multiple case, multiple timing levels for strings of GT
signals will overlap, and thereby, the motorperson can easily clear two
at once by going the right speed early enough, giving the impression to
the inquisitive little boy standing in the front of train that the
motorperson is "clearing the S's into Greens", which is not the case,
but a side-effect as the GT's 2 down clear!
If the GT signal is a home signal that can clear to a diverging route,
its GT-ness would imply "I have a secret about which way I'm going to
clear, and I'm not telling until it's too late." In that case, the
signal in advance of the GT may have a white "D", which means
"the same as a white `S', but the next signal, which must be
a home signal in the GT State, will clear to a diverging route when
the timer expires."
GT is the standard speed control mechanism. An orange, non-electrical sign
of either of the forms
T
35
or
GT
40
indicates the entry to the first timing block of a series of GT's, and the
speed required in MPH.
ST is a bit harder to explain. It does not exist on "old IRT". In
older BMT and IND, it often exists without dedicated aspect.
The idea of ST is to allow trains to close in on each other at stations.
Define the "control length" of a signal (ANY signal) as that distance,
measured forwardtrack from the IJ at the signal, to some IJ downtrack,
the presence of a train in any contained block (inter-IJ extent) forces
the signal RED (force of that red depends, independently, on signal type).
Control lengths in NYCTA are ALWAYS overlapped. IJ's are lovingly
distributed such that it is quite common for their to be IJ's without
signals for terminating control lengths, GT timing sections, the features
to be described, etc.
The idea of ST is to divide the control length into a "required" and
"optional" portion. The "optional" portion is the distant end, which
is usually in a station. Again, these guys (ST signals, i.e., signals
whose control lengths have this feature), tend to come in strings.
It is not at all uncommon for ST, GT, and home signals to be combined
orthogonally in the same signal (although the GT speed had better be
higher than the ST speed!).
Define analogously an "ST state" of an ST signal. It is red; it
would be clear BUT FOR THE PRESENCE of a train in the optional
(distant, don't confuse with home/distant) section of the control length.
Implied that GT time, if any, has expired, interlocking lever, if any,
is clear, and "required" control length is clear.
When a signal enters the ST state, one or multiple timers, associated
with leading-up-to-the-signal, usually overlapped, need not be same as
GT timing sections, are enabled by train entry to the section at issue.
A train going through those sections slowly enough (i.e., spend enough
time in it before entering the next), will effectively cut the
"optional" portion of the control length out of the consideration of the
ST signal, clearing it (and maybe its overlappees) iff it was in the ST
state. It follows at once how this allows trains to creep up
on each other approaching a station.
In newer (60's +) signalling, all divisions, a lit-up white number
(miles per hour of ST timer) under a red (or R/R) means "I am red
because I am in the ST state, and if you approach me slowly enough,
I'll clear." (For purposes of counter-interlocking, BTW, home or
approach signals in the GT or ST states are considered "clear for
all negative purposes.")
A big yellow "T" sign with no number (older), or an explicit
ST
20
indicates the entry to a string of ST's, or sometimes nothing at all.
(ALL aspects of NYCTA are a historical pastiche of styles and implementations).
London goes the other way; most of their route mileage has 2-aspect
signalling, just red and green. They have a form of the "creeping past
red" that Bernard described.
This "creeping" (automatic key-by, or "AK", as some old IRT signs label it)
is implemented by having to get the train's leading wheel in the new block,
but NOT SO FAR INTO IT that the stop trips the train. The delicate spacing
and careful maneuver implements the effective timeing. It takes seconds
to accomplish, and is hardwarily in effect on all new-scheme (not unresignalled
IRT) single-head signals, and no old-scheme signals except single-head ones with an
explicit yellow "AK" sign.
(Implementation note: this is implemented (in the simple case)
by "anybody in the block directly forwardtrack from the signal holds the stop down", which
also prevents a train from tripping itself!)
It is called the "stop and proceed rule".
If a train has been standing at a red for 1 minute, the driver is allowed
to creep the train past the red. This trips the tripper, or automatic
train stop as they say, and the driver has to get out and reset it, which
is presumed to take another minute. After this the driver is allowed to
creep the train forward until the next train is in sight! Naturally this
means that there is NO protection from the driver's misjudging the speed
and colliding with the train ahead, and in fact this has happened a number
of times, sometimes with serious loss of life.
As I said, since 1970+x, it has been impermissible in NYC to use
the AK feature, due to increasing incompetence of motorpersons, combined
with the observation above. Although the blocks are generally short
enough that the train cannot gather vast speed before ramming the end
of the previous one.
Mark Brader, Toronto... ihnp4!utzoo!lsuc!msb
From ARPA, append @seismo.ARPA (I think)
Have I bored everybody yet, or do you want to hear about approach
signals, interlockings, and relays?
Bernie Greenberg, Symbolics, Inc., Cambridge Mass. (BSG@SCRC.ARPA)
(I don't read railroad.rochester, I get these indirectly and via the
DEC-distributed list, please cc me directly to anything relevant to this.)msb@lsuc.UUCP (Mark Brader) (01/30/86)
Well, I did enjoy Bernard S. Greenberg's second NYC signalling article, and I want to comment on a few points, making more remarks about Toronto. > [Someone please forward to MSB!!] If it gets to the ARPA railroad list, it gets to net.railroad on Usenet, and I see it. > In the new scheme, there is (or was at the time I studied) one home signal > on the outbound (J1) track at East New York Station (Jamaica BMT) with > a -blue- light for the third diverging route! In Toronto, a blue light marks an emergency power shutoff switch. There's one at each end of each station and one every 500 ft or so between stations -- usually one on each side of the tracks, in fact. > Do you in fact know it to be that there is no difference in indication > between [Toronto] R and R/R? Only by observed behavior of the trains. Certainly both are equipped with trippers. (In fact, I forgot to mention before that at one time some of the terminal stations were equipped with extra trippers along the platform, to make sure trains didn't speed up on entering the station. There were no signals at these trippers because the trains were going slowly enough that the drivers could see them easily anyway. The lines have now been extended beyond the stations that had these.) > Let us call this state of a signal, in which the signal > would be clear but for the timer not having timed out, the "GT State" > (invented adhoc terminology). I wish I'd thought of inventing terminology; it'd've made my description of the Toronto system more comprehensible. > Stop (tripper) management is not > affected by GT or ST; it is at a lower level of modularity. In other words, if you see a stop signal, there's a tripper, even if it's really a timing signal? Same here. > GT is the standard speed control mechanism. An orange, non-electrical sign > of either of the forms > T > 35 > or > GT > 40 > indicates the entry to the first timing block of a series of GT's, and the > speed required in MPH. The same sign is used here, except it's white on black, and may be electrical if the restriction is for a junction. There are also fixed signs "O", "S", and "P", for power off, slow (or series), and power on (or parallel), which the drivers take as hints. About half a dozen steep downgrades also have instructions to apply certain amounts of braking at certain points, written in full, e.g. "All trains 1/3 brake", then "All trains brakes off". As for the ST timing, I don't believe we have any such thing here. > Have I bored everybody yet, or do you want to hear about approach > signals, interlockings, and relays? I for one would love to. Don't expect any more from me, though! Mark Brader, Toronto... ihnp4!utzoo!lsuc!msb From ARPA, append @seismo.ARPA (I think)