Signal Time Out

'Looking back' to see if a train is approaching requires more design work and, in the case of relay interlocking, additional circuits. Presumably at London Bridge the cost could be justified, but not elsewhere.

I can understand the four minute time-out in areas where trains run at 125 mph. What I do not understand is why a signal will time-out with no train anywhere? Does anyone know the history behind this?

The part of Kings Cross that was resignalled quite possibly now has "Comprehensive" approach locking.

I believe the parts which are a straight recontrol don't have TORR on the new workstations, so I'd be surprised if they've fiddle with the approach locking.

"Simple" approach control, where the signal just times out when the route is pulled, only requires a couple of relays and all the logic is in one place.

When you say 'no TORR' for the recontrolled workstations, is it that the workstation software itself sends a route cancel command automatically once certain conditions are met, as opposed to it happening in the interlocking?

NSW relay interlockings that were re-controlled to computer workstations do it that way

Giantray in post 154534 said:

I can understand the four minute time-out in areas where trains run at 125 mph. What I do not understand is why a signal will time-out with no train anywhere? Does anyone know the history behind this?

"Simple" approach control, where the signal just times out when the route is pulled, only requires a couple of relays and all the logic is in one place.

"Comprehensive" approach control, where it only times out if a train is approaching, requires much more logic because it has to check three or four sections back, which could require lots of circuitry to deal with pointwork (is the approaching train actually heading for this signal? is track circuit DK actually occupied because a train is crossing over a diamond and again, not heading for this signal? how many different signals could change aspect if this one goes from green to red [1]). A lot more relays and a lot more cost.

I've also got a vague idea that comprehensive approach locking was a relatively late addition to UK signalling, with London Bridge being the first PSB to have it. But I could be wrong there.

[1] The Down Fast at Peterborough through the station has, if I recall correctly, five signals in a row that all have to have a route set before any of them will come off on a clear line. So the look-back can be a lot further than you think.

clive in post 154541 said:

"Simple" approach control, where the signal just times out when the route is pulled, only requires a couple of relays and all the logic is in one place.

"Locking", not control!

So the general rule outside of the Southern is unless a signal has an auto button, it won't have comprehensive approach locking. Which brings me to the Donny station sim. I haven't visited or worked there, so i can't say for sure - but i was very surprised there is no comprhensive approach locking on the auto buttoned signals - if this is just something that got overlooked in the sim. then so be it - but i would be very surprised if that was the case in real life - as having to time out often negates or at least severely hinders the reason for providing an auto button in the first place.

Thanks for that Geoff. I must say i am surprised they bothered to put auto buttons on approach locked when cleared signals - but i didn't say it couldn't happen, LOL.

As for the shunt timeout associated with main aspect.....i can't say for sure,as i am a bit out of date - but i always understood that the longer timeout was used because of the poosible approach speed from the signal in rear. A 30 second timeout for a train that could have passed a single yellow at 50 - 60 mph seems a little short to prove the train at a stand before route release (which is the timer's raison d'etre), so the type of interlocking is irrelevant. But, i could be wrong :)

My understanding of a call-on on a signal capable of showing main aspects was that in relay days a single timer would be used so the longest timeout required would apply to all types of routes from that signal.

In CBI days this is different for the different types of route (30s for call-on/shunt; 120 or more for main aspect; 60s for a bay platform).

Surely it should be the route class for the approaching train, not for the route set for the signal, that matters. I can see Izzy's point that if a train is approaching a red at main-line speeds and braking, you want the full 120 seconds even if it's a calling-on route that's been pulled. But equally, if the train is coming out of a siding on two whites, 30 seconds should be enough even if it's a main route that's been pulled.

GeoffM in post 154617 said:


My understanding of a call-on on a signal capable of showing main aspects was that in relay days a single timer would be used so the longest timeout required would apply to all types of routes from that signal.

Surely it should be the route class for the approaching train, not for the route set for the signal, that matters. I can see Izzy's point that if a train is approaching a red at main-line speeds and braking, you want the full 120 seconds even if it's a calling-on route that's been pulled. But equally, if the train is coming out of a siding on two whites, 30 seconds should be enough even if it's a main route that's been pulled.

As TORR crept in to the post - TORR was designed originally for the (original) London Bridge scheme, so is very rare on relay interlockings outside the Southern and even then only a few were fully provided with it. TORR is much easier to provide on CBI's (computer based interlockings), and i believe it is now a standard provision. Interestingly - the TORR relay cicuitry isn't that complicated either, it just wasn't done until somebody thought it might be handy at London Bridge.

PS. Clive - comprehensive app locking wasn't first on London Bridge - a lot of the earlier southern power schemes had it in selected places/signals - Dartford and Feltham had it.

"Different timeout values seems to have been normal on CBIs since at least the late 1980s. Bethnal Green, for example, has 60s for warner routes and 120s for main routes which I think dates from 1989 if memory serves. Recent re-signalling seem to have anywhere from 2 to 4 minutes for main routes, for example 3 minutes is reportedly common on the forthcoming Colwich re-interlocking."

Yes, but my point is that a sub signal associated with a main aspect should have the main aspect time off - as it is a main signal. That is why it is called a subsidiary, and not a shunt signal head.

I'd be interested to know of somewhere where the sub route of a main signal only times for 30 seconds - it goes against the principle of making sure an approaching train is at a stand before releasing the onward route. The timeout value is designed with approach speed in mind, not only what class of route is set. Or at least that's how i was taught to do it.

A main signal with warning class route can have 2 timer values - the main 60-240s value. Or a warning route reduced timout - this is justified by the fact we can guarantee any approaching train has been slowed down before the delayed yellow aspect is displayed, so we don't need the full 120, 180 or 240 second timeout for maximum permissible speed.