Route indicator signals

I'm sure I've seen a train move past a signal with the directional thingy lit up but the actual signal itself still showing red. Is this possible?

Most definitely not in normal operation. However, I can visualise a failure situation where such a signal might be passed under rule with permission if the signal itself had failed to clear due to say a signal relay failure. Of course the train would be tripped and the tripcock have to be reset to proceed.

Logically (what a dangerous word to use when you discuss about signalling!) the feathers are linked with the fact that the route (the point) is set in the indicated direction. This must be assessed before the signal is open. At least on the French railways.

If the route is set and if the next block section is still occupied, the red light must logically still be on. Same in the case of trackl circuit failure, hence the exceptional oreder to proceed in such a case.

Still with French railways, once the route indicator (white lights) is set, the main signal would commute from "carré" (two red lignts) to "sémaphore" (one red), the latter being allowed to be passed after a halt and at very reduced speed.

I have had a failure that involved a signal at danger but with a junction route indicator showing a route! Rather than securing up points and what not, as the route ahead showed a proceed aspect, the trains were just sent on to reverse someplace else until it was fixed.
That's up there with 'my strangest faults'!

So no light could get between the two - forming a false "feather"?

No. It is still a valid proceed signal (but of course should be reported)

Right then. We have a signal:



and



Ok, so it's not a feather, nor is it an UndergrounD signal - but it will illustrate the point. Those lenses are 5" diameter, now the optical system is optimised, and the lenses [1] will try and create a beam of nearly collimated light

[1] better if the lens is a stepped fresnel doublet, rather than this example.

This signal would work fine in tunnel - all that would need to happen is that the beam is closely aligned to the line of sight of the T/Op to allow an early view (and allow for both human reaction time and braking to kick in). Imagine now, trying to spot that signal in broad daylight against a moving background - sky, buildings, trees whatever that change as the train moves - it is really quite difficult - especially if the signal stays at the same aspect on approach - the human eye will however note a change, so it aspect steps up or down then the brain will register that.

Signal designers thought long and hard about this, and ended up with a backboard - ta-daaa!:





The keen-eyed will not a similarity with the older signals on the Met and the District - this is because this signal is made by the Union Switch and Signal Company, who were essentially BPRS and Westinghouse in the UK and the firm that provided these for the Met and MDR. The signal is itself a VR design from 1920 - right at the very dawn of light signalling in that continent - there is one large hood:



Signals like this are around the northern side of the circle - Farringdonia and also near Hammersmith. In the ebb and flow of signalling design single and multiple hoods have alternately fell into and out of favour, single hoods seem to be coming back with certain types of LED lamp.

Yes, if it's a shunt signal rather than a feather - as at West Croydon platform 4, where the 378s proceed into the reversing siding after detraining.

The OP still hasn't got the answer to it's question directly I think.
If you look at a junction indicator head on, using an x, y and z axis where x is left to right, y is up and down and z is closer and farther; I think to OP is referring to a metal plate in the z axis linking the hoods of the top and middle white lights.
Although I think the reason for this is the same as what's been explained regarding backplates.

Ah. Remember what I said upthread about US&S.......

Well, if I remember my US signalling history correctly, it is tied in with the early versions of the Pennsylvania RR position light signals.

I might have a note on this somewhere.

Acton town platform 4 station starter has a multi route shunt signal under a colour light junction signal.

Yep that's a tricky one, I can't ever recall seeing a 'full post' so to speak with signals for diverging main routes and diverging subsidiary routes. There are of course plenty of sites with diverging main routes and single shunt route or single main route with multiple shunt routes.
Where diverging main routes and diverging shunt routes are available it is more usual to see a junction indicator and single shunt signal at the main signal with a further shunt and theatre sign further ahead, the latter having to be cleared before the shunt on the main signal post can be cleared.