The Atchison, Topeka & Santa Fe single track Automatic Block Signaling system that was installed in the early 1900s was explained in the document Automatic Block Signaling, AT&SF Single Track Overlap (original version). The system featured a number of unique functions of which some were relying on the special ATSF signal rules in force at the time. When the AT&SF started to move towards a more standard set of signal aspects and indications the original system became known as "System 1" signaling whereas the new system became known as "System 2".
Around 1960 "System 1" was removed from the rulebook and instead "System 2" became the only set of signal asppects and indications on the AT&SF. This necessitated that the single track ABS was modified to remove the reliance of the "System 1" signal rules. The modifications were quite simple and for many sections of ABS this was the only modification. Probably around the same time ATSF started a redesign of the ABS that converted it to "Flat Pair" ABS based on the use of Absolute Permissive Block (APB) signaling circuits. The "Flat Pair" ABS supposedly was a preparation for a conversion to Centralized Traffic Control and is found on most of the Glorieta and Raton subdivisions. This document, however, explains the modifications to the original "Single Track Overlap ABS" that were carried out, probably in the late 1950s.
The relevant AT&SF signal aspects and indications from around 1960 and forward were:
|9.56||Approach||Proceed prepared to stop at next signal, trains exceeding 40 MPH immediately reduce to that speed.|
|9.61||Stop and Proceed||Stop, then proceed at restricted speed.|
The colored lights for night visibility were approach lit, i.e. the signal would normally be dark but light up when a train was nearby. The signal would light up at least when a train was within a mile of the signal on straight track and at least within sighting distance on curved tracks. In most cases the signal would light up when the previous signal was passed but in some cases where visibility was reduced the signal would only light up when the approach track circuit was occupied. Due to lack of reliable documentation on the approach lighting scheme the illustrations in this document show all signals as constantly lit.
The most common layout of ATSF Overlap ABS placed the block signals in a staggered fashion, typically some quarter to half a mile apart. The controls for a signal (i.e. the track circuits that would place the signal at "Stop (and Proceed)") of course extended to the far end of the block but also included the track circuit(s) up until the following opposing signal. In the example below the controls for signal 682 extended all the way to opposing signal 661 and vice versa. In addition, at least one of the block signals' controls further overlapped into its following block. In this case signal 661's controls extended past signal 682:
If two opposing train closed in on each other (which only happened at speed in case of a mistake by either the dispatcher, a train order office or a train crew) both signals protecting the block would be yellow. But the overlap meant than when the trains closed in on signals 661/682 at least one of them would be replaced to red:
The arrangement meant that the westbound train would be required to stop and then proceed at restricted speed. But the eastbound would be allowed to proceed, slowing first to medium speed and then towards a stop at signal 662. This implies that, in order to be safe from a collision, the stopping distance of the westbound train had to be contained within the stagger distance between signal 661 and 662. With the signals staggered only a quarter to a half mile (or less), this scheme was unfit for higher speeds. As discussed in Automatic Block Signaling, AT&SF Single Track Overlap (original version), the original ATSF signal aspects and indications ("System 1") allowed speeds to be higher, as long as the two trains' combined stopping distances were less than the distance between signals 661 and 682. However, with "System 2" signaling in force, only the distance between 661 and 662 was available.The solution required that, in a situation like above, the train speeds were low enough to contain the stopping distance between a pair of staggered signals (i.e. approximately 40 mph), while at the same time allowing a high line speed. The solution was to, quite simply, add another "Approach" signal leading up to a red signal. This function will here be referred to as "two yellows", since it is not quite identical to the "double yellow" function for approach signals to sidings (see later). With the extra "Approach" in place, both trains would see an "Approach" indication and be prepared to stop at the signals between them, as illustrated below where the trains will be able to stop at signal 661 resp. 662:
The downside of two "Approach" signals was, obviously, that two trains in the same direction would have to keep one further block apart in order for the 2nd train to obtain a "Clear" signal.
An Overlap ABS line was essentially a sequence of blocks like explained above. In the example below there are 3 blocks between sidings but it could be any number. The overlaps in signal controls were usually arranged so that signals for the direction of train superiority were not overlapped, while the signals for the inferior direction did include overlaps. In this ficticious example eastbound trains are superior to westbounds and thus the controls for signal 661 contains an overlap while the controls for signal 682 do not. This in theory allowed trains in the superior direction to run a little closer but in reality other factors prevented this (one such factor being that for practical reasons the overlaps for blocks bordering a siding needed to have their overlaps facing the open line):
To illustrate the signal controls consider an example where the track circuit leading up to signal 691 is occupied. The effect is that signals 681 and 692 both drop to the most restrictive aspect, i.e. "Stop and Proceed". The signals on approach to 681 and 692, i.e. 661, 651, 702 and 722 respectively, change to "Approach" accordingly:
If a block contained switches, for example for a spur leading off to an industry, the ABS supervised that the switches were in their normal position. An open switch basically had the same effect on signals as if its track circuit was occupied. Switches leading from the main track to a siding generally were be more complicated, see the section on sidings below. Special rules applied as to when mainline switches in ABS territory might be opened.
The original AT&SF Overlap ABS featured a somewhat sophisticated siding arrangement compared to most other ABS installations of the time. The transition to "System 2" signaling necessitated the removal of some features, but still the siding arrangements were more flexible than the standard ABS solutions.
The siding switches were located between staggered pairs of signals. The signal controls were the same as explained above except they not only tested for vacant track but obviously also checked that switches were in their normal position. In most other ABS installations the sidings, including their switches, were contained between two signals facing each other (signals 692 and 701 below) but staggering the signals around the switches allowed for some mechanisms to optimize meets - more about that below.
Many siding switches were spring switches to expedite trains leaving the siding after meets. Due to the rules of eastbound trains being superior to westbounds of the same class, most meets would see the westbound take siding. Originally it seems to have been common to equip the siding with a spring switch at the west end while the east switch was rigid but sidings with two spring switches or 2 rigid switches were also found. Spring vs. rigid switches only bears a significance for trains leaving a siding, where spring switches allows for a faster exit. Over the years it seems more and more of those sidings that were regularly used for meets were changed to spring switches in both ends. Sidings mainly used for local switch jobs were left with rigid switches. For the examples below the more common dual spring switch configuration was chosen.
The original AT&SF Overlap ABS featured a "double yellow" function for signals on approach to a siding, i.e. if for example signal 702 is at "Approach" ("yellow"), then signal 722 would also at most be able to show "Approach". This function, standard for single track ABS systems, is to ensure that both trains are approaching a meet at lower speed. At meets it is not uncommon that the siding signals drop to "Stop and Proceed" right in front of an approaching train, due to the other train closing in from the other side of the siding:
The "double yellow" on approach to a siding was basically the same situation as the "two yellows" mentioned earlier for two trains accidentially approaching each other between sidings. However, the "double yellow" was implemented in a simplified form that could not be used generically. With "double yellow", the approach signal (681/722 on the illustrations) could never show "Clear" if the following signal (691/702) was at "Approach". But with "two yellows" the approach signal would have shown "Clear" if both signals 691 and 701 (resp. 702 and 692) were showing "Approach" (or better). This difference in implementation, which the ATSF apparently did not find worthwhile to change, meant that the sequence of yellows became irregular, as illustrated in the below sequence where, at some point, there are 3 "Approach" signals behind the westbound train:
Placing the siding switches between the staggered signal pairs allowed for a special function to optimize meets. The function allowed a train to reverse its siding switch and pull into the siding while still upholding an "Approach" signal for the opposing train to enter the siding area. In the example below the meet is set up for the westbound train to take siding and if it arrives earlier than the eastbound the eastbound will be able to arrive to the siding area on "Approach" (signal 702) even while the westbound occupies the track between signals 691 and 692. To further expedite meets the ABS allowed signal 702 to stay up after the switch was restored - provided that only the switch track circuit was occupied when the switch was restored. The meet may therefore be executed quicker than if the eastbound had to stop and proceed at restricted speed from signal 702.
*) In the first situation above, signal 691 in the original AT&SF Overlap ABS would have shown yellow ("Caution" in System 1 signaling), for the westbound to enter the siding. With System 2 signaling, and the different meaning of a yellow ("Approach"), this could no longer be allowed and the function was removed. Red, however, now meant "Stop and Proceed", instead of "Stop" in "System 1", and the train therefore could pull into the siding at restricted speed.
A remnant of this funtion remained. The previous intermediate signal (in this case signal 681) would show "Stop and Proceed" if the siding switch was reversed, thus requiring the westbound train to proceed at restricted speed towards the siding. A train already holding the main track for a meet therefore cannot line the switch to the siding for the meeting train until that train has entered the last block:
Had instead the eastbound train been the one to take the siding signal 691 would have dropped to "Stop and Proceed" on the westbound train until the eastbound reached the siding switch and reversed it. Then signal 691 would go back to "Approach". This was due to the asymmetry in the overlap between signals 691 and 702. The westbound train would only see signal 691 at "Stop and Proceed" if it arrived more or less simultaneously with the eastbound, in which case it would have to wait for the eastbound to pull into the siding anyway. Rules of train superiority required the inferior eastbound train to clear the main track at least 5 minutes before the superior westbound arrived, minimizing the chance of the westbound encountering "Stop and Proceed" in signal 691.
Leaving a siding in Automatic Block Signaling territory required the train to have authority to occupy the main track. Apart from that, in ABS territory there were also rules to directly assure that a train was not heading out into another train (whether either of them were there in error). After all a train in the siding was invisible to the ABS and the type of ABS discussed here did not providing specific protection against a train pulling out.
One method to assure a clear main track was to observe the block signals. If they were clear in both directions nothing was approaching from either side. After a meet - when an opposing train just went by - there was obviously nothing coming from behind. And observing the signal on the main track (signal 701) would indicate a free track ahead:
With a spring switch the train would simply pull out of the siding. Occupying the switch track circuit made the signals drop accordingly and after the train was clear of the siding switch - and the switch had restored - the signal behind the train would clear again. In the case shown only to "Approach" due to the eastbound train.
If instead the train was to leave the siding over a rigid switch the train crew would have to reverse the switch manually. The same rules applied regarding ensuring clear track but when leaving over a rigid switch the train would have to move forward and foul the switch track circuit first before reversing the siding switch. The rule of fouling the track circuit was added at a later date, possibly due to some incidents, but the exact reason is not known to the author. It is likely that the rule was intended to allow the train crews to observe a track circuit failure - at least in other situations the crew was required to observe the signals dropping to "Stop and Proceed" as the train passed them:
Rounding up the description of siding functions an illustration of the difference between a reversed siding switch and an open siding switch (as would usually be the result of an equipment failure). Obviously a siding switch will be open a brief moment when operated between normal and reverse but the slowacting semaphore signal motors will bridge this timing gap:
In order to illustrate the features of the Overlap ABS the following example shows a meet between a westbound and an eastbound train. The eastbound train is superior and will hold the main track while the westbound takes siding. First the westbound makes its appearance at the previous siding proceeding west. While in the siding block the train is protected by opposing signal 652 at "Stop and Proceed" while signal 662 and 682 are at "Approach":
When the train passes signal 651 it enters the controls of signal 662 which drops to "Stop and proceed" and signal 692 consequently drops to "Approach". Once past both signals 651 and 652 signal, 652 clears behind the train:
Similarly when the train passes signal 661/66,2 signals 682 drop to "Stop and Proceed". Signals 702 and 722 drops to "Approach" (722 due to the "double yellow" function) and signal 662 clears behind the train:
Signal 651 goes back to "Approach" once the train clears its overlap:
The train slows as it approaches the siding and stops before the siding switch. The eastbound train is not approaching yet so signal 691 is still "Clear":
The train crew then reverses the switch to the siding. Signal 691 drops to "Stop and Proceed" when the switch is opened. The westbound can now pull into the siding. Even though the switch is within the controls of signal 702 the special circuits at sidings allow signal 702 to stay at "Approach" with the switch reversed. The signal even stays up when the trains pulls into the siding and occupies the track circuit between signals 691 and 692:
Had the train just passed signal 691 and occupied the switch track circuit without the switch being reversed first signal 702 (and 691 for that matter) would have dropped to "Stop and Proceed". But once the switch has been reversed signal 702 stays up even when the switch is returned to normal, as long as the track circuit is still occupied:
The westbound is now clear of the mainline and the track circuits and pulls to the west end of the siding waiting for the meet. Note that since this meet took place with the westbound - inferior - train taking siding, the eastbound would all the while have been able to continue all the way to signal 692 since signal 702 did not at any time show less than "Approach".
As with other types of ABS the westbound train is invisible to the ABS as there is no track circuit in the siding. With the siding switch back to normal and the track circuits clear the ABS is now back to its normal state with all signals "Clear":
If the train taking siding is inferior by class (i.e. a lower class train) it must clear the main track at least 5 minutes before the superior train is scheduled to leave. For maximum efficiency as far as the ABS goes the westbound should clear the main track in time for signal 722 to change to "Clear" before the superior eastbound train sees it at "Approach", thus avoiding the eastbound having to slow down:
If instead the train taking siding is inferior only by direction (i.e. same class as the superior train) it must only clear the main track before the superior (eastbound) train is scheduled to leave. If the superior train arrives at the meet first its crew may reverse the siding switch for the westbound and allow it to enter the siding without stopping:
However the timing of the meet came to be, once both trains arrived at the meet they can proceed if they have Authority to do so. The eastbound train is on the main track and can proceed according to the indication of signal 692. The westbound train is in the siding and needs to ensure that no train is approaching on the main from either direction. Since it just met the eastbound nothing can be approaching from the east. And with signal 701 "Clear" nothing is approaching from the west either so the eastbound can pull out of the siding, trailing the spring switch:
Both trains are now on their way again, which concludes this description of AT&SF's single track ABS as upgraded in the late 1950s. Thanks to Rex Seedig and Jon Roma for providing source documents and schematics to assist the research.
Text, Images, HTML: Carsten S. Lundsten.