Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Basically, trains are not held at a signal until the crossing is clear, although there are exceptions or situations where this happens. The crossing gates or barriers are closed and the crossing checked to be clear before a protecting signal is cleared: this signal will be cleared sufficiently early for the train to continue at full speed. This is the situation with crossings that have some sort of visual control, such as CCTV or manual.
AHB crossings work independently of the signals and the signals are not affected by them.
Basically, trains are not held at a signal until the crossing is clear, although there are exceptions or situations where this happens. The crossing gates or barriers are closed and the crossing checked to be clear before a protecting signal is cleared: this signal will be cleared sufficiently early for the train to continue at full speed. This is the situation with crossings that have some sort of visual control, such as CCTV or manual.
In some places trains are brought up to a signal at red in a station platform while the barriers or gates are being closed. The idea is the signaller/attendant finishes closing, confirms clear by pressing a button or pulling a lever and then the signal can clear, assuming other safe conditions are proved ahead. With so much road traffic today and experience of accidents, many signal box instructions require crossings to be closed well before the train approaches the station, even if interlocking controls don't enforce that.
Sometimes there is a signal after an AHBC strike-in (the trigger point at a timed distance approaching the crossing). If so, special controls may be required to suppress activation of the road warnings if the signal is held at red for any reason. When the signal is ready to clear again, first the road warnings are started, then a delay in clearance of the signal may be enforced to ensure the legal minimum road warning time is maintained. The reason for this is a desire to avoid exceptionally long warning times for trains that are going to be halted before the crossing for an indeterminate time.
If a platform is present within the strike-in, a signal between it and the crossing may be provided especially for this protection of minimum road warning time. The signaller in this case would have a stopping/non-stopping switch on the panel/instrument shelf which they could preselect for each train, depending on whether it was scheduled to stop or not. The signal would already be cleared on approach for a non-stop and held for a time for one stopping in the platfrom, with the crossing activation also delayed by some means (maybe another timer after passing the non-stop strike-in). Whether this is required or not depends on a detailed study of the site layout, traffic characteristics and achievable train performance, but as a general rule, the closer the platform is to the crossing, the more likely the extra delay signal will be required. On less busy lines with long blocks, many of these signals approaching AHBCs are solely there for this timing control and are not also block signals allowing trains to follow each other more closely. They will be provided with a distant signal with AWS at braking distance of course.
Safety authorities are very concerned about extended road warning times at AHBCs because they are, by definition, half non-barriered, and long warnings without a train showing up are known to convince some impatient road users the crossing may be broken. That's often when attempts to weave around barriers are made. In addition to the minimum road warning time mentioned (usually 27s) which no train should beat, 95% of trains must arrive within 75s. Also 50% must arrive within 50s. These values are mandated in the formal level crossing order document from the Secretary of State, so form part of the legal permission to operate the particular site.
Agreed! Though modern so-called 'manually controlled barrier' (MCB) crossings are bristling with automation today. Auto lower, auto raise, obstacle detection scanners, etc.
Automatic-
AHBC, AOCL, AOCR (and their upgraded +B variants) etc.
Requires no signaller intervention / action for a green aspect to be shown across them
Manual -
MCG, MCB, MCB-CCTV etc..
Require the signaller to do SOMETHING in order to clear the route over them. Even if its only observing the crossing and pressing 'crossing clear'.
Then theres ODs that are somewhere in the middle and just witchcraft...
And of course there will be exceptions to each crossing type because of local arrangements.
Automatic-
AHBC, AOCL, AOCR (and their upgraded +B variants) etc.
Requires no signaller intervention / action for a green aspect to be shown across them
Manual -
MCG, MCB, MCB-CCTV etc..
Require the signaller to do SOMETHING in order to clear the route over them. Even if its only observing the crossing and pressing 'crossing clear'.
Then theres ODs that are somewhere in the middle and just witchcraft...
And of course there will be exceptions to each crossing type because of local arrangements.
You forgot the new Automatic Full Barrier Crossing Locally monitored (AFBCL) type which also uses OD scanners and enables local monitoring of the crossing by the driver, through clear sightlines and road surface illumination on approach. Correct operation of equipment and sensor clearance is signalled via the same flashing red/white indicator used for ABCL. This type can reduce road closure time compared to the MCB-OD, though, like ABCL, is limited to a maximum approach speed of 55mph given good sighting and considerably less where sightlines are constrained. That often doesn't matter much on minor rural lines, especially around stations where all trains stop, where they are either already slowing or just starting away from rest.
Where in one direction an ABCL is immediately after a station platform, the warning sequence is often started manually by the driver pressing a button (plunger) on the platform, or even pulling a chain connected to a switch through the cab window! Everything after that is automatic though, including raising barriers after passage of the train.
This may be a purely hypothetical question because I don't know if such a situation actually exists, but how do things work when there's an automatic crossing immediately after the exit from a loop? If a long, heavy freight train is held in the loop then, once given the signal to proceed, it will take several seconds to release its brakes and will then move very slowly off the loop, therefore taking a long time to reach the crossing. On the other hand, a light locomotive or track maintenance machine will 'take off' as soon as the signal clears and will be on the crossing within a few seconds. So how do you minimise the waiting time at the crossing when faced with these two extremes?
This may be a purely hypothetical question because I don't know if such a situation actually exists, but how do things work when there's an automatic crossing immediately after the exit from a loop? If a long, heavy freight train is held in the loop then, once given the signal to proceed, it will take several seconds to release its brakes and will then move very slowly off the loop, therefore taking a long time to reach the crossing. On the other hand, a light locomotive or track maintenance machine will 'take off' as soon as the signal clears and will be on the crossing within a few seconds. So how do you minimise the waiting time at the crossing when faced with these two extremes?
Traditionally, you had to assume the fastest possible acceleration, as a train reaching the crossing without adequate warning was a big no-no. If that meant the crossing closed for longer for trains with slower acceleration, tough. In the days when I did such calculations, we used the acceleration curve for a light class 47. And we assumed that the loco was running up to the signal at red, so not accelerating from rest.
Nowadays I understand that there are systems that can measure the speed of an approaching train, and adjust the crossing initiation accordingly. But I think they still have to assume the train might accelerate, presumably at the fastest possible rate.
== Doublepost prevention - post automatically merged: ==
Yes, but......
Where a signal is immediately adjacent to an automatic crossing that is proved by flashing white light, such that the flashing white light is mounted on or adjacent to the same post as the signal, then the white light won't start flashing unless the signal is ready to clear, and the signal won't clear unless the crossing is proved closed and the white light ready to start flashing. This avoids the driver being given conflicting stop/go information. So in these rare cases, the automatic crossing is effectively protected by the signal.
Generally, with controlled crossings, the crossing initiation point has to be sufficiently far in advance that the crossing can close and be checked to be clear, in sufficient time for the protecting signal to clear and the signals in rear to step up to green, without the driver of an approaching train being checked at a cautionary aspect. This can mean the road is closed for a significant time, depending on how far back the first caution signal is, and how slow the approaching train.
With automatic crossings, the crossing just has to close with sufficient warning time (about 27 seconds) for road users. With AOCL and ABCL, trains speeds should be slow enough that the train could stop if there is an obstruction, or the crossing fails to operate and the white light doesn't flash. AHBs however are used on higher speed lines, where there is little or no chance of a train stopping within 27 seconds or less, so they don't bother providing a flashing white light.
I understand that it is policy to replace AHBs with controlled crossings, which are undoubtedly safer, but can involve significant extra delay to road users.
== Doublepost prevention - post automatically merged: ==
You forgot the new Automatic Full Barrier Crossing Locally monitored (AFBCL) type which also uses OD scanners and enables local monitoring of the crossing by the driver, through clear sightlines and road surface illumination on approach. Correct operation of equipment and sensor clearance is signalled via the same flashing red/white indicator used for ABCL.
I think that they have greatly overcomplicated the AFBCL. The Isle of Man has automatic full barrier crossings that don't have OD scanners. Like at AOCL and ABCLs, the driver checks the crossing is clear without needing expensive scanners. And because they are full barriers, they aren't over worried about the crossing staying closed for too long - drivers can't ignore the lights, as there is a full barrier in the way. This means the strike-in and strike-out arrangements are very simple - no track circuits, just one treadle either side of the crossing. Simples!
Traditionally, you had to assume the fastest possible acceleration, as a train reaching the crossing without adequate warning was a big no-no. If that meant the crossing closed for longer for trains with slower acceleration, tough. In the days when I did such calculations, we used the acceleration curve for a light class 47. And we assumed that the loco was running up to the signal at red, so not accelerating from rest.
Nowadays I understand that there are systems that can measure the speed of an approaching train, and adjust the crossing initiation accordingly. But I think they still have to assume the train might accelerate, presumably at the fastest possible rate.
Predictors are widely used in the USA. I only know of a few UK examples in East Anglia on the Cromer line. The technology monitors the changing electrical characteristics of an approach track circuit as a train moves along it. Using the output as a proxy for speed, the appropriate point to start the road warning is determined. Difficult to deploy through junction areas I understand.
I understand that it is policy to replace AHBs with controlled crossings, which are undoubtedly safer, but can involve significant extra delay to road users.
I've not seen that specifically mentioned anywhere as policy, but it seems every time an AHBC is renewed it's upgraded to OD. Maybe that just falls out automatically on relative risk assessment.
I think that they have greatly overcomplicated the AFBCL. The Isle of Man has automatic full barrier crossings that don't have OD scanners. Like at AOCL and ABCLs, the driver checks the crossing is clear without needing expensive scanners. And because they are full barriers, they aren't over worried about the crossing staying closed for too long - drivers can't ignore the lights, as there is a full barrier in the way. This means the strike-in and strike-out arrangements are very simple - no track circuits, just one treadle either side of the crossing. Simples!
I believe they were most worried about vehicles and pedestrians being trapped between barriers as there's no escape route with full closure. The offside booms don't close until the road surface has been scanned clear, and I think that's monitored until they're fully closed. Perhaps a simpler type without OD might suffice on quiet lines in locations where there are few pedestrians. Or cheaper scanning tech might be investigated. I only recently found out some of the latest pedestrian road crossings incorporate sensors looking down from the poles for proving the crossing is clear of people before traffic signals change to proceed. Pedex, Puffin and Green Person Authority crossings are current variants:
In terms of protection, I would class them as protected by signals. They should have fixed distants on the approach. Traditionally the gates themselves were the stop signals, with big round red discs mounted on the gates. Nowadays, some of them have more conventional stop boards as well.
== Doublepost prevention - post automatically merged: ==
At one time, NR and the ORR seemed to be quite keen on them, but they don't seem to be quite the in-thing anymore. Wasn't there an incident on one of the trial crossings where an accelerating train managed to get to the crossing with insufficient warning time?
I believe they were most worried about vehicles and pedestrians being trapped between barriers as there's no escape route with full closure. The offside booms don't close until the road surface has been scanned clear, and I think that's monitored until they're fully closed. Perhaps a simpler type without OD might suffice on quiet lines in locations where there are few pedestrians. Or cheaper scanning tech might be investigated.
This is exactly what I meant by over complicating the AFBCL. The Isle of Man use their version without any OD in both rural low usage locations, but some of them (such as the one in the middle of Port Erin next to the children's play area and giving access to the big housing estate) have a very high pedestrian usage. In the extremely unlikely event that a vehicle or person should get caught between the barriers, I should imagine that the train would stop, the train driver would have a few choice words, and then manually open the crossing and release whoever has got trapped.
As I understand it, like AOCL and ABCL, it is an inherent feature of these crossings that the train must be able to stop if the driver observes that the white light isn't flashing or there is someone on the crossing, so I really don't see the need for OD.
I only recently found out some of the latest pedestrian road crossings incorporate sensors looking down from the poles for proving the crossing is clear of people before traffic signals change to proceed. Pedex, Puffin and Green Person Authority crossings are current variants:
We are aware of an issue with emails from the Forum to Microsoft-based email accounts (hotmail/outlook/live.com email addresses). This is being looked into currently, thanks for your patience meanwhile.
