Points machine detection

[Archie]

Can somebody explain to me the process of points detection?

Not really sure what you are asking. Detection is the name given to the proving that the point switches have really moved to the position which the interlocking has requested- i.e. that the closed switch is sufficiently tightly closed to the stock rail on one side and that at the other side there is a sufficiently wide gap between the open swich and the other stock rail so that there is no risk of the back of the wheel flange contacting it. Also detection will (almost always) prove that the points are physically held in position by a facing point lock (or alternaive means of achieving the same aim).

Detection of mechanically operated points is sometimes still achieved mechanically, bit in general is achieved electrically and therefore there are NKR and RKR repeat relays that are energised when the points are proved fully in one position or the other. It can get a bit more complicated than that where there are multiple detection elements, there is a subtle difference between pure "neat detection" and "correspondence", but this gives the overview. Fundamentally it is the feedback in the control system so that we know the site equipment has truly responded to the command to move. Does that answer what you were meaning?

[Peter]

Can somebody explain to me the process of points detection?

Not really sure what you are asking. Detection is the name given to the proving that the point switches have really moved to the position which the interlocking has requested- i.e. that the closed switch is sufficiently tightly closed to the stock rail on one side and that at the other side there is a sufficiently wide gap between the open swich and the other stock rail so that there is no risk of the back of the wheel flange contacting it. Also detection will (almost always) prove that the points are physically held in position by a facing point lock (or alternaive means of achieving the same aim).

Detection of mechanically operated points is sometimes still achieved mechanically, bit in general is achieved electrically and therefore there are NKR and RKR repeat relays that are energised when the points are proved fully in one position or the other. It can get a bit more complicated than that where there are multiple detection elements, there is a subtle difference between pure "neat detection" and "correspondence", but this gives the overview. Fundamentally it is the feedback in the control system so that we know the site equipment has truly responded to the command to move. Does that answer what you were meaning?

I must admit when I first started to read this, the first thought that I had was that I thought the difference between detection and correspondence was quite important. If the points are in "closed and locked" position, they will be legitimately detected there irrespective of what the interlocking had asked them to do meaning that they are only in correspondence when the state of the detection and the state of the interlocking call agree.

Another area that my thoughts differed slightly was on "open switch detection" because on the UK mainline, we have not normally proved this consistently. OK, the Clamplock has had this feature but until recently, no other operating mechanism has - we have relied on the integrity of the stretcher bars and assumed that if the switch that should be closed is detected, the switch that should be open is open.

[Archie]

Can somebody explain to me the process of points detection?

Not really sure what you are asking. Detection is the name given to the proving that the point switches have really moved to the position which the interlocking has requested- i.e. that the closed switch is sufficiently tightly closed to the stock rail on one side and that at the other side there is a sufficiently wide gap between the open swich and the other stock rail so that there is no risk of the back of the wheel flange contacting it. Also detection will (almost always) prove that the points are physically held in position by a facing point lock (or alternaive means of achieving the same aim).

Detection of mechanically operated points is sometimes still achieved mechanically, bit in general is achieved electrically and therefore there are NKR and RKR repeat relays that are energised when the points are proved fully in one position or the other. It can get a bit more complicated than that where there are multiple detection elements, there is a subtle difference between pure "neat detection" and "correspondence", but this gives the overview. Fundamentally it is the feedback in the control system so that we know the site equipment has truly responded to the command to move. Does that answer what you were meaning?

I must admit when I first started to read this, the first thought that I had was that I thought the difference between detection and correspondence was quite important. If the points are in "closed and locked" position, they will be legitimately detected there irrespective of what the interlocking had asked them to do meaning that they are only in correspondence when the state of the detection and the state of the interlocking call agree.

Another area that my thoughts differed slightly was on "open switch detection" because on the UK mainline, we have not normally proved this consistently. OK, the Clamplock has had this feature but until recently, no other operating mechanism has - we have relied on the integrity of the stretcher bars and assumed that if the switch that should be closed is detected, the switch that should be open is open.

Yes that is what I wanted to know, I kind of knew anyway but just wanted clarification. Could you be so kind to just go over the timer relay (WJR) within the circuit, 7.5 seconds I believe and it's main function, is it to do with in case of an obstruction, and and detection cannot be achieved in this time.

[Peter]

Could you be so kind to just go over the timer relay (WJR) within the circuit, 7.5 seconds I believe and it's main function, is it to do with in case of an obstruction, and and detection cannot be achieved in this time.

Without going into detail of how it works, its purpose is to cut off the motor after a time if the detection is not made up, mainly to stop the motor burning out as would happen if it continued to run until the techs arrive. There is normally some sort of clutch that slips or hydraulic bypass to deal with the time in between the points stopping moving on an obstruction and the motor cutting out.

Why 7.5 secs? Who knows!?!?!?

[PJW]

Could you be so kind to just go over the timer relay (WJR) within the circuit, 7.5 seconds I believe and it's main function, is it to do with in case of an obstruction, and and detection cannot be achieved in this time.

Without going into detail of how it works, its purpose is to cut off the motor after a time if the detection is not made up, mainly to stop the motor burning out as would happen if it continued to run until the techs arrive. There is normally some sort of clutch that slips or hydraulic bypass to deal with the time in between the points stopping moving on an obstruction and the motor cutting out.

Why 7.5 secs? Who knows!?!?!?

My guess is that some people wanted 5 and others 10 so it was a British compromise! To be honest with longer heavier switches 7.5seconds is a bit too short. I remember that at Cogload we originally had to put two in series to get 15 seconds as the WJRs are non adjustable- I think that the SSI TFM time value is slightly more. It might even be selectable on a "per point" basis with certain interlockings.

Re detection- I would argue that detection must theoretically be both the open and closed switch, although I agree with Peter that the open switch position is often not explictly measured but only inferred by the integity of the multiple stretchers (and whereas this seemed adequate for very many decades, Potters Bar and particularly Lambrigg provide contrary evidence).
BR Western Region has always distinguished between neat detection (that only states where the points actually are) and correspondence (which proves that it accords with the requested position). Other Regions generally incorporated the "called position" into the very first detection relay and therefore made no such distinction. The latest NR circuits has gone against that policy. Hence the words are used inconsistently depending on context which is confusing to the newcomer but something one learns to live with; the industry is not completely homogeneous. Signallers almost always say "detection" when they actually mean "correspondence" and so do many signal engineers including myself- I find myself only really using the two terms when the difference is important in the context and I tend then to say "pure detection" to emnphasise that it isn't "correspondence". Probably more than Achie wanted to know!

[Archie]

Could you be so kind to just go over the timer relay (WJR) within the circuit, 7.5 seconds I believe and it's main function, is it to do with in case of an obstruction, and and detection cannot be achieved in this time.

Without going into detail of how it works, its purpose is to cut off the motor after a time if the detection is not made up, mainly to stop the motor burning out as would happen if it continued to run until the techs arrive. There is normally some sort of clutch that slips or hydraulic bypass to deal with the time in between the points stopping moving on an obstruction and the motor cutting out.

Why 7.5 secs? Who knows!?!?!?

My guess is that some people wanted 5 and others 10 so it was a British compromise! To be honest with longer heavier switches 7.5seconds is a bit too short. I remember that at Cogload we originally had to put two in series to get 15 seconds as the WJRs are non adjustable- I think that the SSI TFM time value is slightly more. It might even be selectable on a "per point" basis with certain interlockings.

Re detection- I would argue that detection must theoretically be both the open and closed switch, although I agree with Peter that the open switch position is often not explictly measured but only inferred by the integity of the multiple stretchers (and whereas this seemed adequate for very many decades, Potters Bar and particularly Lambrigg provide contrary evidence).
BR Western Region has always distinguished between neat detection (that only states where the points actually are) and correspondence (which proves that it accords with the requested position). Other Regions generally incorporated the "called position" into the very first detection relay and therefore made no such distinction. The latest NR circuits has gone against that policy. Hence the words are used inconsistently depending on context which is confusing to the newcomer but something one learns to live with; the industry is not completely homogeneous. Signallers almost always say "detection" when they actually mean "correspondence" and so do many signal engineers including myself- I find myself only really using the two terms when the difference is important in the context and I tend then to say "pure detection" to emnphasise that it isn't "correspondence". Probably more than Achie wanted to know!

Many thanks PJW for your input, I like to learn different inputs so it can never be more than I wanted to know, but I found it quite funny the point you made about signallers saying "detection." I used to be a points operator on weekend possessions, and when it came to given the worksite back I would obviously go out and make sure that all sets of points are normalised and powered back up, and when I went through the process of handing them back to the signaller I have always found they say 'detection,' and never heard them say 'correspondance.'

So is it right in saying in lehman's terms that correspondance is the last phase of this process.

[PJW]

Many thanks PJW for your input, I like to learn different inputs so it can never be more than I wanted to know, but I found it quite funny the point you made about signallers saying "detection." I used to be a points operator on weekend possessions, and when it came to given the worksite back I would obviously go out and make sure that all sets of points are normalised and powered back up, and when I went through the process of handing them back to the signaller I have always found they say 'detection,' and never heard them say 'correspondance.'

So is it right in saying in layman's terms that correspondance is the last phase of this process.

Certainly in the sense that the contacts at the points are individually detecting individual bits of the points and these are then summated to form the detection for an entire point end and then an entire point number but by the time it is used within tthe interlocking it is almost always correpondence and the end result when the signaller gets an indication then it is most certainly correspondence (even though it is often referred to still as "detection")