Baker St Bakerloo BM5

[dave1]

Can anyone explain why signal BM5 won't clear until the signal ahead A4011 has to have a green aspect, the distance between the two is about 435m.

[Nortube]

I can't see any reason for this.

There is a 15 second timer associated with BM5 and so in normal circumstances, BM5 will not clear until the platform track circuit has been occupied for 15 seconds (unless the signalling has changed over the past few years). However, this shouldn't make any difference to B5 clearing after that if the section ahead is clear.

As A4011 has four digits, this suggests that it is an additional signal that has been added at some point, possibly due to sighting reasons. Pre the new signalling, here were a few similar signals on the Northern line (usually added because of sighting around bends) and under normal circumstances, a driver would rarely see these at danger.

I can't remember the exact reason for them now, but I vaguely remember that it was following a recommendation after a crash following a train applying the rule (Stratford?) and was to ensure that a driver carrying out the rule passing two green stop signals before resuming normal speed would not run the risk of potentially running into the train in front.

This may be a similar reason for the working of B5 / 4011.

[Nortube]

Just checked the Stratford crash report of 1953:
[ Click here ]

and it seems that my vague memories were correct. It recommends additional signals on bends. On the Northern line, these signals were generally added by 1957.

[railtechnician]

Jul 2, 2016 10:18:20 GMT dave1 said:

Can anyone explain why signal BM5 won't clear until the signal ahead A4011 has to have a green aspect, the distance between the two is about 435m.

It is many years since I worked on the Bakerloo, the late 1980s resignalling, so I cannot recall the signalling layout even though it is only 12 years or so since working on the lever frame after maintaining the Jubilee ends of the Bakerloo crossovers. However, A4011 sounds like a 'round-the-bender' signal designation and presumably is placed to give early warning of the state of A401 which would otherwise not have enough sighting time should a train be held at it. 435m is about three train lengths on average and if BM5 were allowed to clear with a train still held at A401 it could still be possible for a train leaving the platform and accelerating to trip past A4011 and rear end the berthed train at A401. 

I would suggest that the distance between A4011 and A401 is the critical distance here in terms of safety. That said of course there must be two red signals behind a berthed train so A4011 is the first and BM5 is the second.

On the other hand my explanation could be completely wrong.

[dave1]

Thanks so far, I looked at the report also and yes there were signals installed and they are known as round the benders. What I don't understand is the area was resignalled in the 1990s? and they still put one of these signals in. I looked at the yellow peril for when Baker Street IMR came in and it shows the signalling quite different, the 401 signals were the home signal for Baker Street and there was a 4 digit signal (3991) just over the distance of 4011 but there was a signal before before it (391) so an extra signal. The other difference I found was Regents Park has one home signal but before it had three, could all of this be part of the difference?

[railtechnician]

Jul 3, 2016 12:28:57 GMT dave1 said:

Thanks so far, I looked at the report also and yes there were signals installed and they are known as round the benders. What I don't understand is the area was resignalled in the 1990s? and they still put one of these signals in. I looked at the yellow peril for when Baker Street IMR came in and it shows the signalling quite different, the 401 signals were the home signal for Baker Street and there was a 4 digit signal (3991) just over the distance of 4011 but there was a signal before before it (391) so an extra signal. The other difference I found was Regents Park has one home signal but before it had three, could all of this be part of the difference?

I would have to see the signalling diagram / strip print to comment further upon the old and new arrangements that you refer to.

What I would say is don't get hung up about home signals! In general terms the more home signals approaching a station the better the service in the peak periods as more trains can be packed into an area thus compacting 'time'. Generally multiple home signals are based upon the principle of three homes and a starter but it is not universal, the Picc generally has / had two homes and a starter while the northern had as many as 5 homes and a starter. Multiple home signalling is another subject entirely.

[hangerlanejn]

There was a booklet called 'How the Underground Works' published many years ago and there was a description of multiple Home signalling in it - I will check the loft when I have time and give an extract. (I doubt copies are still generally available.

[dave1]

Jul 3, 2016 20:02:10 GMT railtechnician said:

Jul 3, 2016 12:28:57 GMT dave1 said:

Thanks so far, I looked at the report also and yes there were signals installed and they are known as round the benders. What I don't understand is the area was resignalled in the 1990s? and they still put one of these signals in. I looked at the yellow peril for when Baker Street IMR came in and it shows the signalling quite different, the 401 signals were the home signal for Baker Street and there was a 4 digit signal (3991) just over the distance of 4011 but there was a signal before before it (391) so an extra signal. The other difference I found was Regents Park has one home signal but before it had three, could all of this be part of the difference?

I would have to see the signalling diagram / strip print to comment further upon the old and new arrangements that you refer to.

What I would say is don't get hung up about home signals! In general terms the more home signals approaching a station the better the service in the peak periods as more trains can be packed into an area thus compacting 'time'. Generally multiple home signals are based upon the principle of three homes and a starter but it is not universal, the Picc generally has / had two homes and a starter while the northern had as many as 5 homes and a starter. Multiple home signalling is another subject entirely.

Here is the part diagram from the yellow peril (1977). The new signalling there is no A399 and A3991 is about 4m further.

[railtechnician]

Jul 3, 2016 22:35:02 GMT hangerlanejn said:

There was a booklet called 'How the Underground Works' published many years ago and there was a description of multiple Home signalling in it - I will check the loft when I have time and give an extract. (I doubt copies are still generally available.

I have the 1966 copy of that book in my 'Underground' library. Pages 21-32 give a good layman's overall appreciation of LT Signalling but it is a basic functional descriptive rather than a technical descriptive. There is a nice full page (yep not very big as the book is about one third A5 size!) diagram illustrating the approach to a platform via multiple home signalling. 

[railtechnician]

Jul 4, 2016 7:04:03 GMT dave1 said:

Jul 3, 2016 20:02:10 GMT railtechnician said:

I would have to see the signalling diagram / strip print to comment further upon the old and new arrangements that you refer to.

What I would say is don't get hung up about home signals! In general terms the more home signals approaching a station the better the service in the peak periods as more trains can be packed into an area thus compacting 'time'. Generally multiple home signals are based upon the principle of three homes and a starter but it is not universal, the Picc generally has / had two homes and a starter while the northern had as many as 5 homes and a starter. Multiple home signalling is another subject entirely.

Here is the part diagram from the yellow peril (1977). The new signalling there is no A399 and A3991 is about 4m further.

In the 1939 pre-Jubilee line layout (Harsig signalling diagram) BM5 was the Baker Street platform 8 home signal, BM7 was the starter and the other side of the spring toggle trailing turnout (now 3B points) was A399 before A397A, A397B A397C Regent's Park multiple home signals. A nice simple layout with BM6 as the original platform 7 starter. The controlled signalling back then being from Baker Street signal cabin which I recall was between platforms 7 and 8 at the end of the concourse when I was involved in the 1977 resignalling St John's Wood to Baker St platforms 7 and new platform 10 in my first year of LT service. A3991 was added subsequently following the collision incident cited by Nortube.

The 1977 layout shows platform 8 home signal as BM4 with BM5 as the starter, in between are delta 401^B and BM500, these would have possibly been added as a recommendation following an enquiry into a collision between ballast trains upstairs on the Met at the converging junction from platforms 3 & 5. BM500 is a draw up signal controlled by the delta timing section to ensure that a train approaches the converging junction beyond BM5 at 'walking' pace. A train approaching too fast would not clear BM500 from red to yellow (if the starter is red) and would have to wait for a 4.5 second time relay to completely wind up whereas BM500 would clear to yellow if the train has approached slowly enough for 4.5 seconds wind up time to have already expired. Should BM5 be green then BM500 would clear to green on approach at the correct speed. 
The 15 second time delay that Nortube mentions is to ensure the train has come fully to rest prior to the selection of a conflicting move, in this case a train crossing over to the Bakerloo from Jubilee platform 7.

When an area is resignalled the signal numbering tends to change to reflect changes in layout, generally gaps are left in the numbering which for auto signalling is continuous ascending or descending (both roads doing so in the same direction, one road even numbered and the other road odd numbered by convention) to allow for future resignalling, sometimes modifications to layouts over the years lead to areas where any spare numbers have been used and so in resignalling the numbering will often be changed to restore gaps for future use.

It seems that little changed when the area was resignalled except that the numbering altered slightly and distances may have altered by a few metres. When an area is signalled everything has to be measured and laid out according to the signalling scale plans which are the result of mathematical calculation of signalling overlaps according to line speed, stock braking characteristics, gradient etc. Also changes to basic signalling technology and incidents and events may alter signalling standard practices to improve the safety integrity and so there will always be slight changes to the lengths and positions of signalling overlaps, the location and lengths of timing sections e.g. where a signal may have multiple overlaps to take account of varying levels of track occupancy i.e. the somewhat congested peak with 90 second headway or off peak with extended intervals where trains can motor and require longer full speed overlaps. As a young technician new to the job there were several occasions where I was assisting a signal supervisor marking up the locations of signals and blockjoints to the scale plans and determining the exact positions for the associated trainstops and signal equipment locations. Sometimes the scale plans did not quite fit the tunnels and so new signals that would have been in more or less the same location as the old could be three or four metres ahead or in rear of the existing, a determination usually made by the supervisor on site. Marking up tunnels to the scale plans with a 100' foot tape measure could be interesting especially on a tight curve and that could also account for the slight discrepancy between what was on paper and reality, it really depended upon where the tape was laid and whether that was the same place that the surveyor for the tunnel distances on the scale plan had laid his tape! We always used the rails for measurements when marking up but of course the lengths of two parallel rails also vary anywhere not on a straight! 

[Nortube]

Sorry this post is a bit lengthy, I've tried to cover everything as I understand it, starting from the beginning, and I think I've got it right Hope it all makes sense!

Baker St was resignalled either late 1989 or by June 1990. Presumably this is when that section between 3B points and the approach to Regents Park SB platform was changed to Block jointless track circuits).

Looking at the changes to the signalling compared to as it was in the yellow peril (I'll call this OLD) and as it was from 1990 (I'll call this NEW):
BM500 - platform draw-up - removed
A399 - replaced by A4011 (in a similar position)
R397AB - removed
R397BC / A397A - becomes A395
A397B - removed
A397C - removed

A395 is the only home signal for Regents Park SB
This is followed by A391 which is the SB starter for Regents Park.

Thus, between Baker St inner home and Regents Park SB (starter to starter), there are the following signals:

OLD signalling - at time of yellow peril:
A401 (not shown on YP) (Baker St plat 8 outer home), or possibly A401A (outer) and A401B (intermediate)?
RBM500 / BM4/400 (inner home)
BM500 (platform draw-up
BM5 (starter)
A399
A3991
R307AB
R307BC / A397A
not shown on YP clip, but presumably followed by:
A397B
A397C
A391 (Regents Park starter)

NEW signalling - 1990:
A425 (outer home Baker St plat 8)
BM4 (inner home)
BM5 (starter)
A4011 / R395
A395 (home)
A391 (Regents Park starter)


As I see it, under the OLD system, BM500 would act as a draw-up signal that protected 3 crossover. Presumably BM500 would go to yellow after X seconds and thus an approaching train would either:

1    have slowed down (to perhaps 10 or 15mph) and so reduced the train's braking distance. BM5 would then be protecting the crossover
or
2    would be going too fast and so reaching BM500 before it went to yellow and so tripping the train.

This is the standard method used where it is desired to be able to allow a train into the platform when the crossover is reversed - i.e. two trains can be travelling in the same direction at the same time. This saves time over the other method of holding a train outside the platform while a train crosses over from the other platform or waiting until the approaching train has stopping in the platform before allowing the train from the other platform to proceed.
This was common at Kennington SB City where B800 in the platform slowed the train because the points ahead were set for a SB Charing Cross train towards Morden.

(B800 was under the platform ledge and poorly sighted with no repeater. In days of old, it was common for drivers not to notice it was red and get tripped at it. A suitable choice of words was uttered by the driver before they got down on the track to reset the trip and carried on. Golders Green NB has a similar layout, but it was worse for the driver then because a driver getting tripped was usually greeted by a rousing cheer from other crews on the platform )


Under the NEW system, BM4 protects the crossover and will only clear if 3B points are normal. BM5 is approach-controlled and will not clear until a train has occupied the platform track for 15 seconds. If there is a train in the platform, BM5 will be protecting the crossover.

This is another common method used and examples on the Northern line were Moorgate NB and Kings Cross NB. An approaching NB train would either be held outside the station while the train ahead was reversing, or the train would be held in the platform and once it was detected that it had stopped, the crossover ahead would be reversed.


A4011 must be green before either starter BM2 (route 2 to cross from Jubilee to Bakerloo line) or BM5 can be cleared. This effectively means that the track ahead must be clear as far as the overlap of A395 - the start of the Regents Park platform track circuit. I.e., if following a train, BM5 will not clear until the train ahead has berthed in Regents Park platform.

As far as I can work out, A4011 serves no real purpose as far as track capacity is concerned and would make no difference if it wasn't there, other than being one additional stop signal for the purpose of applying the rule. A4011 should never be seen at danger by an approaching train unless there is a signal failure.

Assuming that A3991 on the OLD signalling was a round the bend signal, I assume that it still serves the same purpose now that it has been renumbered to A4011.

What is a little strange is why they renamed the signal to A4011 when there was no longer a 401 signal in use. Perhaps they just liked the idea of keeping it four digits, but why not leave it as A3991?

[Nortube]

my post crossed in the post with RT's posting

I'd forgotten about looking to see if there were any Harsig Bakerloo diagrams

[railtechnician]

Jul 4, 2016 18:20:48 GMT Nortube said:

my post crossed in the post with RT's posting

I'd forgotten about looking to see if there were any Harsig Bakerloo diagrams

I concur with your comments, of course the only way to be 100% on the various functions is to see the signalling layout and associated signalling circuits side by side.

The designation BM4 / 400 in the old layout suggests to me that this semi-auto signal could also work automatically when 3B points are normal and the line ahead is clear. There were/are similar arrangements on the Picc at Arnos Grove for the E/B platform and centre platform E/B routes when the line ahead is clear.

With regard to A4011 it must still be there for additional 'sighting' of the signal ahead or it would have a 3 digit designation except that R395 is at the same location which rather obviates that idea. It is one of those queries that a look at the prints should clear up but I suspect you'd only ever see double green there and never red as you say or indeed green and yellow except perhaps in the event of a failure.
So yes A4011 may simply be there to provide appropriate limits of control for BM5 and BM2 starters. Without it the limits of control for those signals would be up to and including the overlap of A395. Yet as A4011 needs to be green to clear either starter there is perhaps more to it unless this is a case of extra 'belt and braces' after simplifying the earlier signalling layout.

ISTR that there was more than one reason for auto signals with 4 digit numbering and I just can't recall what the other reason was!

[Nortube]

I got a copy of How the Underground Works on a visit to London a few years before I actually joined the Underground Green cover with yellow writing. From what I recall (it's up in the loft somewhere!), I managed to keep it in a pretty good condition. I found it very interesting at the time.

There's a description of how multiple home signals work, with some examples based on the old Northern line signalling, here:
[ Click here ]

I think the information is correct, although there is a typo on the the Basic Signalling notes (page 4, Absolute overlap, paragraph 1)

[railtechnician]

Jul 5, 2016 20:17:01 GMT Nortube said:

I got a copy of How the Underground Works on a visit to London a few years before I actually joined the Underground Green cover with yellow writing. From what I recall (it's up in the loft somewhere!), I managed to keep it in a pretty good condition. I found it very interesting at the time.

There's a description of how multiple home signals work, with some examples based on the old Northern line signalling, here:
[ Click here ]

I think the information is correct, although there is a typo on the the Basic Signalling notes (page 4, Absolute overlap, paragraph 1)

Actually my copy is grey with yellow writing and is 1968 rather than 1966 as I initially quoted.

Yes "This ensures that the signal does not remain at danger with a train in the overlap." is clearly a typo!

Of course blockjoints are not always at the signal but generally are within a metre or so but it does rather depend upon line speed e.g. on the west end of the Picc a blockjoint can be several metres beyond the signal and trainstop and the same is true on the Met.

I am still wracking my brains trying to recall the other reason for a signal in standard LT signalling to have a four digit designation. I may have to dig out my old T2S training notes although I am not certain the answer will be there as my vague recollection is that it is rarer than 'round-the-benders' and something that may have cropped up several years after I passed out as a signal lineman.

***Got it!*** The other reason for a 4-digit auto signal designation is for an auto draw up signal which is relatively rare.