The length of time depends on the settings of how long it takes the route to time out. They vary between 30 seconds and 4 minutes on our route.
For shunting signals not associated with a main aspect signal, where the movement is supposed to take place at slow speeds, the standard time is 30 seconds. Unless there is no train detection (track circuit) and no signalling leading up to a GPL (e.g. a long siding) that protects points that lead onto a main line. Then often the time delay is two minutes.
For main line colour light signals, or for calling on/shunting/position light aspects associated with a main aspect signal, regardless of the number of aspects, the old standard (at least on Western Region) was two minutes. But sometime in the 1980s, three minutes appears to have been adopted.
If there is a long distance between main line signals, the length of the time delay may be extended. The longest that I knew of was at Chipping Sodbury, where the down direction signals on the exit to the long axle counter sections had a six minute timer.
Another thing puzzles me... 3 minutes? A couple of people have mentioned that. Is it not possible for the signaller to simply change the points to the correct route and then - assuming the correct route is clear - the train be good to go immediately? Why the additional 3 minute block?
The signalling system doesn't know why the signal has been put back to red, so is designed on the assumption a train could be approaching at line speed. The delay ensures any such train has stopped at the signal before the points can be moved or, if it was unable to stop, has occupied the track circuit preventing the points being moved under it.
Thanks - I guess that makes sense to the extent that the system doesn't necessarily know that the train concerned has actually stopped in order to query the route. But it seems surprising to me that it's not possible for the signaller to override the delay - to cover cases where the signaller has just been talking to the train driver and therefore knows that it's actually safe for the train to proceed immediately?
The whole reason for a complex signalling system with interlocking controls, is to try and prevent train crashes caused by human error. Either by the signaller or train driver.
A standard set of principles has been developed over the years (although back when BR existed, each region had their own).
One principle is to avoid giving the signaller any override facilities. This is because, when, in the past, such facilities were provided, sooner or later, they were misused. The investigators of various accidents and incidents therefore recommended that the railway should consider very carefully about providing such releases. Even where releases are provided, they are only to be used in a real emergency situation. And the signaller has to follow a set procedure before they are allowed to use the override.
Although various members have provided some good detail on the system, I am going to expand on it a little more.
All modern route interlocking controlled signals where there is either a junction, points, ground frame, swing nose crossing, switch diamond crossing, manually controlled level crossing (including CCTV), or bidirectional signalling, have approach locking. It is this approach locking system that enforces the time delay.
Before I continue to talk about the approach locking system, I need to tell you about the route locking system.
After the signal has been routed and before the signal has actually cleared, the interlocking electrically locks all points and G.F.s to prevent them moving in front of, or under the train. Similarly any level crossing controls are locked to prevent the crossing being open to road users. This protects the path or route that has been set for the train. Thus ensuring that it has safe passage over the points and level crossings. This is called route locking.
After the train passes the signal, even if the signaller cancels the route, the interlocking (route locking) will continue to hold the points and level crossings ahead of (before the train) in the locked state until after the train has passed them.
In addition to the route locking, there is the approach locking system. In fact the approach locking system uses the route locking system to do most of its work.
The approach locking system becomes effective as soon as any part of the signal shows a proceed aspect, be that a route or junction indicator, a calling on or shunting signal (position light aspect) or a main aspect (yellow, double yellow or green).
If the train passes the signal (while showing a proceed aspect) in the normal way, the approach locking will release. Hence in normal running, there is no enforced time delay, because there is no need.
However, if the signaller cancels the route / puts the signal back to red AFTER it has displayed a proceed aspect (or a route or junction indicator, a calling on or shunting signal etc.), then the approach locking will prevent the route locking from releasing. Thus keeping the route ahead of the train in a safe state for the train. No matter how fast the train is travelling. This prevents the signaller from moving any points in front of a moving train, or under a moving train (both of which have occurred in the past and both of which have resulted in train crashes, before approach and route locking were introduced), or raise any barriers/open a level crossing to road users.
If there is a train, either the train will stop at the (now red) signal. Or it will pass the signal. If it passes the signal, the interlocking will continue to keep the points locked that the train is on, preventing them from being moved under the train. It will also keep any level crossings locked.
Regardless if there was a train or not, after the time delay, the approach locking will ‘time out’ and release the route locking, which in turn will release the points and any level crossings from being locked. Only now can the signaller either call another route or move the points via the individual switch or equivalent. Or open any level crossing to road users.
So as you can see, it’s all about safety, preventing human error, and allowing normal operation without causing any delay.
On busy lines, or where it was thought that there was a need, and where the money for a more flexible system was made available, a ‘smarter’ system is used. The system described in the paragraphs above is known as the ‘when operated’ approach locking. Because it operates every time that the signal is put to red by the signaller if a train has not passed it.
The ‘smarter’ system is known as ‘comprehensive approach locking’. For normal running, it works as described above. If a train is detected approaching the signal that has been put back to red, or the train is approaching any signal that would have changed state to a more restrictive aspect, then the time delay works exactly as described in the paragraphs above.
The difference is, if there are
no trains approaching, then there is no need to lock the route and enforce a time delay. The system ‘looks back’ so to speak to see if it actually needs to lock the route. No approaching train means the signaller can cancel the route/put the signal back to red, then immediately re-route it somewhere else.
Of course, if a train is on the approach track circuit (or axle counter), or on the berth track circuit (or axle counter), the system will hold the route locked and enforce the time delay. Because as someone has said, the interlocking system does not have any method of detecting the speed of the train. Or knowing the exact position of the train. Hence when a train has stopped at the signal due to being wrong routed, the signaller has to wait for the enforced time delay to end before they can call another (different) route from the signal.