Cost Vs Safety

[MarkyT]

In addition to lamp proving, TPWS equipment has a healthy status output that is monitored by the signal on approach (to rear). If the loops are not powered up correctly with current of the appropriate frequency flowing, then the rear signal remains at red.

 

[SpacePhoenix]

With LED signals there's probably a tolerance for a certain number of the individual LED bulbs to fail, I would imagine that a failure of 25% of the LED bulbs for a single colour of a signal head would probably be considered as a "bulb failure"

 

[carriageline]

Failure of the top yellow (the one not used to display a legitimate single yellow aspect) doesn't affect the previous signal.

I have got my self confused with something I was discussing the other day (blank signals with a yellow in rear in some installations (aka lamp or controls)), but what your saying is correct!

The reason being the top yellow is not lamp proved. So the signal in rear would just see it as a single yellow, and then act accordingly
(IE show a YY instead)


Sent from my iPhone using Tapatalk

 

[Railsigns]

I have got my self confused with something I was discussing the other day (blank signals with a yellow in rear in some installations (aka lamp or controls)), but what your [you're] saying is correct!

The reason being the top yellow is not lamp proved. So the signal in rear would just see it as a single yellow, and then act accordingly
(IE show a YY instead)

That's not right either. The signal in rear would remain at green.

 

[455driver]

If a signal was showing double yellow and one of the yellows fails, it will show a single yellow.

What if the bottom lamp fails?

 

[Railsigns]

What if the bottom lamp fails?

What do you mean? I've already said what will happen if either one of the two yellows fails (the other yellow stays lit, so a single yellow is shown). Only the bottom yellow is lamp proved, however.

 

[455driver]

What do you mean? I've already said what will happen if either one of the two yellows fails (the other yellow stays lit, so a single yellow is shown). Only the bottom yellow is lamp proved, however.

So you are saying the top lamp will light but the bottom one wont, is that correct?
If that is the case then I will categorically state that you are wrong, the bottom lamp must be proved lit before the top lamp will be energised, if the bottom lamp is not proved then the top one cannot light.

This can be seen on filament lamps as the signal will run very quickly from the red to the green via the yellow and double yellow lighting very briefly.

 

[Railsigns]

So you are saying the top lamp will light but the bottom one wont, is that correct?

That's correct.

If that is the case then I will categorically state that you are wrong, the bottom lamp must be proved lit before the top lamp will be energised, if the bottom lamp is not proved then the top one cannot light.

No, it's you who is wrong. The circuit that lights the top yellow lamp only checks the state of the aspect control relays; there's no lamp proving involved whatsoever.

This can be seen on filament lamps as the signal will run very quickly from the red to the green via the yellow and double yellow lighting very briefly.

That's caused by the aspect control relays energising in sequence and not relevant to the matter in hand.

A driver will rarely encounter a signal with the top yellow lit and bottom yellow failed because in most installations the signal in rear would be held at red.

 

[455driver]

today #68 at 2114 you said

That's correct.
A driver will rarely encounter a signal with the top yellow lit and bottom yellow failed because in most installations the signal in rear would be held at red.

So the driver would see green followed by the top yellow lit on its own, yes?
But yesterday #64 at 2221 you said-

That's not right either. The signal in rear would remain at green.

So which is it?

 

[Railsigns]

So which is it?

Look again at what I wrote, but in context (also, you have mixed up the numbers and times of my two posts which you quoted).

In post #64, I was referring to failure of the top yellow.

At the bottom of post #68, I refer to failure of the bottom yellow.

 

[455driver]

Thanks, I have swapped the quotes around, I forgot it multi quotes them in time order not in the order they are tagged.

That's caused by the aspect control relays energising in sequence and not relevant to the matter in hand.

Would the aspect control relays be energised if the lamp isnt proved?

 

[Railsigns]

Would the aspect control relays be energised if the lamp isnt proved?

Yes. The aspect control relays for any given signal are unaffected by lamp failure of a main aspect in that same signal.

 

[carriageline]

Failure of the top yellow (the one not used to display a legitimate single yellow aspect) doesn't affect the previous signal.

So the driver would come from a green to single yellow? Legitimate question, didn't know that.


Sent from my iPhone using Tapatalk

 

[ac6000cw]

So the driver would come from a green to single yellow? Legitimate question, didn't know that.

Err, that could be the normal sequence anyway in a 3-aspect or distant + stop signal area (if approaching a signal at danger, further down the line). You only get double-yellows in 4-aspect signal areas.

 

[Tomnick]

If there's normally no YY aspect in the sequence though, there'll be braking distance between the Y and R, which wouldn't be the case here. It's not dangerous though - although the driver doesn't know it at the time, the next signal must still be Y or better, so the worst that'll happen is a heavy brake application and a change of underwear...!

 

[ac6000cw]

If there's normally no YY aspect in the sequence though, there'll be braking distance between the Y and R, which wouldn't be the case here. It's not dangerous though - although the driver doesn't know it at the time, the next signal must still be Y or better, so the worst that'll happen is a heavy brake application and a change of underwear...!

...and doubtless on the phone to the signaller pretty quickly

 

[Railsigns]

It appears that someone's after a second opinion...

http://www.simsig.co.uk/

 

[TUC]

Zero deaths is a right thing to aim for, although I would be the first to criticise paranoid approaches to health and safety. However the principle of aiming for zero deaths has to be the right aim. I remember hearing the Medical Director where I used to work saying that he had got criticism at a meeting of clinicians because we had set a target of 'no avoidable deaths'. His response was 'OK. I'll increase the number then. You tell me. How many avoidable deaths is it OK to aim for?'

 

[carriageline]

Err, that could be the normal sequence anyway in a 3-aspect or distant + stop signal area (if approaching a signal at danger, further down the line). You only get double-yellows in 4-aspect signal areas.

Instead of trying to be pedantic, read the posts properly. I do know the correct aspect sequences thank you as it's my job to know that... We are talking about YYs with a blown bulb.




Sent from my iPhone using Tapatalk

 

[SpacePhoenix]

How likely is it that any MK3s that are still slam door will eventually be converted to powered plug doors? It would mean quicker get aways from stations as staff wouldn't have to keep going along the train closing doors

 

[ac6000cw]

Instead of trying to be pedantic, read the posts properly. I do know the correct aspect sequences thank you as it's my job to know that... We are talking about YYs with a blown bulb.

Sorry , but I don't know what your job is (it doesn't say on your forum profile), and my response wasn't meant to be pedantic or sarcastic etc. (and unfortunately I hadn't noticed your earlier posts in the thread).

Not everyone reading the forum understands the principles of UK colour light signalling.

 

[rebmcr]

How likely is it that any MK3s that are still slam door will eventually be converted to powered plug doors? It would mean quicker get aways from stations as staff wouldn't have to keep going along the train closing doors

The majority of them will become sliding pocket doors before too long, by virtue of replacing HSTs with 800s & 801s.

 

[HSTEd]

Its amazing how complicated these systems are - engineering them in the days before CAD must have been.... interesting.

Buildings full of relays.... (did they use telephone exchange style reed relays or other types in these systems?)

 

[ac6000cw]

Its amazing how complicated these systems are - engineering them in the days before CAD must have been.... interesting.

Buildings full of relays.... (did they use telephone exchange style reed relays or other types in these systems?)

Not for railways, but when I first started designing electronics for a living in 1980 it was done with pencils, rulers and stencils (for digital logic gate symbols etc.) on A1 size engineering drawing boards. All accompanied by the smell of ammonia when you had to copy the drawings using a dyeline copier....(the copies were literal 'blueprints').

Computing ran to HP programmable calculators (Reverse Polish Notation, anyone ?) and a couple of BBC Micros, plus an original IBM PC. Oh, and 'Google' was a bookshelf full of component databooks and circuit design handbooks.

Aren't you glad you missed it (or enjoyed it, depending on age)

--- old post above --- --- new post below ---

Getting back to the railway stuff, yes in the 1920s and 30s in particular the more far-sighted railway signal engineers were probably at the forefront of developing techniques to design and safety 'prove' large relay logic systems. The first large relay interlockings were installed in the 1930s by the LNER. Before then power signalbox interlocking was largely performed by miniature lever or slide frames, so it was essentially using same design techniques that had long been used for full size lever frames, with the addition of more electric detectors/locks for facing points etc. in place of mechanical 'slotting' on the ground.

The actual signal relays I've seen are physically quite large (and no, they are not reed relays), hence the size of the relay rooms....there are a few pictures of BR standard ones here - http://en.wikipedia.org/wiki/Relay#Railway_signalling - but I'm sure an expert will come by soon and provide more up-to-date info.

BTW - it's not just 'did' use relays, there are a lot of relay interlockings still in daily use - if you want a simple piece of digital logic which can handle 'real world' currents and voltages and is intrinsically fail-safe, relays still have a lot going for them

 

[HSTEd]

Is there any particular reason reed relays are not preferred in Railway applications?
The low maintenance and incredibly long life would seem to recommend them or is it simply a result of the development of relay based interlocking logic having predated them becoming practical?

 

[edwin_m]

The railway standard relays are designed above all to be failsafe, so that in any credible failure situation the coil will de-energise and the contacts will go to the appropriate position. Other types of relay have greater probability of failing in the opposite state due to contact weld-up and probably other mechanisms.

This reliance on the relay failure mode is fundamental to the design of railway signaling, as any relay de-energized always represents the more restrictive state. Signals go to red and track circuits are assumed to be occupied, resulting in delays but hopefully not catastrophic accidents!

Each BR939 relay is a bit smaller than a housebrick and no doubt pretty expensive. Relays are still used for small interlockings but I don't believe any large relay interlockings have been commissioned in the UK since Waterloo around 1990. SSI and later computer-based interlockings are simply more cost-effective, not least because they reduce a relay interlocking the size of a tennis court to something the size of a fridge.

 

[Tio Terry]

Not for railways, but when I first started designing electronics for a living in 1980 it was done with pencils, rulers and stencils (for digital logic gate symbols etc.) on A1 size engineering drawing boards. All accompanied by the smell of ammonia when you had to copy the drawings using a dyeline copier....(the copies were literal 'blueprints').

I did my spell in a very similar Drawing Office - but not as modern as you! We didnt have any computer at all, I was the first to have a calculater - a Sinclair - before that it was slide rules.

But I was working on Signalling Systems so was designing the circuits using relays that you are talking about. The pictures you have a link to are relatively modern installations, the older ones were shelf mounted relays. There are still thousands of relays still in use today, they are very reliable which is why they are used in Safety Critical Systems.

 

[carriageline]

At work S&T are trying to trace a long standing interlocking fault. I was pondering over their diagrams for all the locs with them, and you must be some sort of genius to understand it all, nevermind design it! It literally blew my mind, hats off to you all


Sent from my iPhone using Tapatalk

 

[carriageline]

Sorry , but I don't know what your job is (it doesn't say on your forum profile), and my response wasn't meant to be pedantic or sarcastic etc. (and unfortunately I hadn't noticed your earlier posts in the thread).



Not everyone reading the forum understands the principles of UK colour light signalling.

Sorry my post was a bit narky, didn't mean for it to be (long day etc) I never put my job on there, maybe I should!


Sent from my iPhone using Tapatalk

 

[dysonsphere]

American shelf-type signaling relays are the origin of the 19 inch rack

lol and Ive used 19" rack mount all my life and never knew that, just goes to show you can always learn something new.