Looking for Hayling Signals Prototype Advice

[bécasse]

It was quite common Brighton practice to have splitting stop signals well in advance of the facing points to which they applied without having any means of "holding the route", be it a dolly at the divergence or an extra long locking bar. A quick perusal of Wallis' photographs will produce numerous examples. The homes at Hayling in their original position would have fallen well within this practice, particularly if my surmise that the two FPLs were originally on the same lever is correct. I still think that it was extremely unlikely that shunt 7PULL was provided before Southern days. Wallis routinely photographed unusual equipment on the Brighton and I can't think of a single photo of his which shows a PULL/PUSH lever on that system. Once Wimbledon took the reins, of course, such a lever would have been the standard response to meeting a need in a frame which had no spares.

[Martin Shaw]

With respect Becasse, as I pointed out in my earlier post, Wagstaff recorded 7 pull existing in 1920, which is certainly before Southern days. He may well have been wrong, but since this is the primary evidence of it's existence, I would prefer something more substantial to deny it.

 

Mike, would you care to differentiate between a running shunt and it's additional use for route holding.

 

Regards

martin

[The Stationmaster]

11 hours ago, Martin Shaw said:

With respect Becasse, as I pointed out in my earlier post, Wagstaff recorded 7 pull existing in 1920, which is certainly before Southern days. He may well have been wrong, but since this is the primary evidence of it's existence, I would prefer something more substantial to deny it.

 

Mike, would you care to differentiate between a running shunt and it's additional use for route holding.

 

Regards

martin

Martin if the signal is used for route holding it is basically there to keep the facing point locked - normally because the splitting stop signal is too far in rear of the facing point to provide the required length of facing point locki(ing) bar.  So it is effectively a work round to achieve compliant signalling but more particularly compliant locking of the facing point.  Depending on the distance n between the splitting running signal and the route holding signal/point there could still be an element of risk in my view but it is greatly, and obviously acceptably to HMRI, mitigated.

 

A facing shunt is more precisely exactly what it is says - the signal is provided for shunting purposes but lies within a running signal route and can be a considerable distance in advance of the running signal for which it also acts as a running shunt, but it is basically there for shunting purposes.   Undoubtedly far more common in colour light installations where running signals can be a considerable distance in rear of any pointwork over which shunts might set back in order to take a different route after the points are changed.    Interestingly on WR semaphore installations where some facing shunting signals appeared in later years following rationalisation schemes the ground disc forming the running shunt did not detect the position of the facing points - that being considered unnecessary because the points were locked by an FPL.

[Martin Shaw]

Mike

Thanks for your thoughts, in essence though they are one and the same thing, I can't imagine a situation where a running shunt didn't precede the runnng signal in the locking, and if not an afterthought, would most likely lock the points and FPL as well, as well as holding the route. I do recognise that a shunt signal might be provided purely for route holding purposes, but certainly from an SR perspective it would be treated no differently. I suspect this is getting to a point of semantics rather than anything else. I note your comment in respect of the WR, no surprise really.

Regards

Martin

[The Stationmaster]

19 hours ago, Martin Shaw said:

Mike

Thanks for your thoughts, in essence though they are one and the same thing, I can't imagine a situation where a running shunt didn't precede the runnng signal in the locking, and if not an afterthought, would most likely lock the points and FPL as well, as well as holding the route. I do recognise that a shunt signal might be provided purely for route holding purposes, but certainly from an SR perspective it would be treated no differently. I suspect this is getting to a point of semantics rather than anything else. I note your comment in respect of the WR, no surprise really.

Regards

Martin

You'll now get me looking for examples where they weren't cleared in order to clear the running signal. (I wouldn't be surprised if there were some somewhere  ).   Mind you there were definitely places on the GWR where discs were not provided for route holding whereas they probably would have been provided under LBSCR practice.

 

Kingsbridge interestingly used double discs following the 1930 alterations one arm being specifically a running shunt but also obviously used for shunting purposes otherwise running round or bringing an engine off shed to attach to a train in the main platform  would have been unsignalled moves.  So the signals were clearly there for shunting purposes but also had to be cleared in order to avoid a running move passing a signal at danger.  The 1895 signalling included no ground signals at those facing points so there were no route holding ground signals and shunting moves would have been handsignalled.

[Echini]

I found some very interesting free software called SigScribe4 at Modratec.com, which allows you to build a layout with signals, points and FPLs and then allocate levers in a signal frame to each. You then allocate routes for each signal and the software will calculate the resulting locking. You can then run simulations tyo check that the locking is as expected.

Within this software, I built the Hayling Island signal plan with the ten levers from the small signal box. The result was as follows:

 

This illustrates the position of the levers for the No. 2 Road Down Home. The "L" and "F" below the levers indicates if they are Locked or Free.

Lever 7 controls the 7Push ground signal, but the system does not allow for a Push/Pull lever. That was not a problem as I just used Lever No. 11 for 7Pull. However, that would not work because the simulation will only allow one route based on any one signal. 7Pull is reversed for both No. 1 Road Down Home and No. 2 Road Down Home according to the Hayling mechanical locking chart. In addition, if the siding points and catch point No. 6 is reversed, then logically 7Pull and 7Push will be free to be able to indicate if the engine is either cleared to leave the sidings (7Push into the frame) or to go into the sidings (7Pull out of the frame). That means that for 7Pull out of the frame at "Safe to proceed", there are no less than three possible routes behind the signal.

When running round the carriages, the engine would go past 7Push (into the frame), over the points at 6 reversed and then stop before going back past 7Pull (out of the frame) over points at 6 normal onto No.2 Road to reconnect to the carriages. The bracket signals carrying the two Down Home signals is 118 yards beyond the 7Pull ground signal and they certainly never went that far. So, there would be nothing (other than routine) indicating to the driver which of the three possible routes was being cleared by 7Pull. Bearing in mind that on occasion, in addition to going onto the No. 2 Road to reconnect to the carriages, the engine might go back into the sidings to pick up trucks or it might shunt the whole train back into No.1 Road and the bay platform to await the arrival of the next train from Havant.

That seems to be unusual and it does not seem surprising that the Modratec software would not allow it.

Other that that, the simulation works correctly and it is capable of simulating very much larger and complex layouts. It also appears that Modratec will build the required levers and locking for you for a price!

 

[RailWest]

Although in early years - given that many frame designs did not allow for conditional locking - it tended to be the practice that each shunt signal worked for one route only (usually the diverging one), in later SR days it became quite common for them to work for multiple routes, so I would say that the Hayling situation with 7PULL is common not unusual. IIRC similar existed in the 1930s new box at Seaton.

 

One possibility for you.....there were instances in SR days where one shunt was worked by two levers, one for each route - usually, but not always, where the original installation of two discs side-by-side was reduced to one, but the separate levers retained to avoid any need to alter the locking. So why not get around the simulator limitations by assigning three levers to 7PULL, one for each route?

[Titanius Anglesmith]

The Modratec software is very clever, but it does have some limitations. One thing which irks me about it is that it seems to omit reciprocal locking between points. In your example above, 10 lever should lock 5 lever reverse. In practice 5 lever would be locked against 6 lever as you would never need 5 and 6 points both reverse at the same time. Therefore with 10 reverse, 6 will be locked normal by virtue of 5 being locked reverse. This was very common practice as it greatly simplifies the locking required. 

Modratec seems to omit this, and unless you intervene it will use 10 lever to lock both 5 and 6, which is an unnecessary complication. 

 

Unless I’m doing it wrong, which is quite likely.... 

[RailWest]

I think it depends upon how you 'read' such things. Also, with mechanical locking a lot of the 'reverse' locking is done automatically by virtue of the way that it works, whereas with electrical locking it has to be stated specifically.

 

For example, 5 and 6 must lock each other to avoid conflicting moves. This could be done in practice by a single double-ended 'dog' between the two adjacent tappets - either lever is 'free' to pull, but whichever is pulled automatically locks the other. 10 is released by 5, but the act of reversing 10 automatically then prevents 5 from being put back - you don't need to state specifically that 10R locks 5R, as it is implicit by the fact that '5 releases 10'.

[Echini]

Thanks for the reply Railwest. Unfortunately three levers for 7Pull would not work as the "routes" in the simulation are based upon the signals, not the levers. 

That doesn't particularly worry me because with the interlocking as per the simulation, 7Pull will only be at "Stop" if 7Push is reversed or signal 8 the Up Advanced Starter is reversed - in all other situations 7Pull can be reversed - or at "safe to proceed" (except technically if all ten levers are normal). So, in looking at building a working lever set, rather then complicate the physical interlocking, which is difficult enough anyway, it seems easier to control 7Pull electrically and to stick with lever 7 controlling 7Push.

The general idea is to build a control box with a working set of levers and interlocking, which provides an electrical output that, in turn, works servos under the model signal box to move the levers in the box and work the signals and points through wires and rodding. Quite a challenge!

In looking at the multiple routes for 7Pull, I was more concerned with how it actually worked in practice - it would have required a "conversation" between the engine driver and the signalman on each occasion to make sure that the engine ended up in the right place. Although I am sure that the "conversation" would have been no more that a gesture over time!

 

[RailWest]

>>>7Pull will only be at "Stop" if 7Push is reversed or signal 8 the Up Advanced Starter is reversed...

...or 9 or 10 are reversed.

 

>>>Unfortunately three levers for 7Pull would not work as the "routes" in the simulation are based upon the signals, not the levers....

I would envisage a scenario where you 'imagine' that 7PULL is in fact three discs side-by-side (or stacked), one of each of the 3 possible routes. Assign each 'virtual' shunt an individual physical lever in your 'control box', so that you can work out the locking quite easily using the simulator. However in practice there would be one physical disc on the model, controlled by the one lever in the model signal-box, which could be operated by any one of the three levers in the control box.

[Echini]

Does anyone have any reference manuals/books that refer to the manufacturer or type of interlocking frame at Hayling? The space would have been very limited for the interlocking frame. Would there be a frame interlocking or Dog chart anywhere on record?

[RailWest]

According to SRS records this was a LB&SCR GF pattern tappet locking with levers at 5" centres and probably dating from about 1884. That doesn't tell you much at the moment  but gives you a start for searching the specific details.

Edited July 25, 2019 by RailWest

[bécasse]

The physical locking was probably remarkably simple. Lever 4, the FPL on point 5 (and originally point 6 as well) would have acted as a direction lever, facilitating movements in the down direction when normal and in the up direction when reversed. Once lever 1 became the FPL on point 6 it would effectively had mirrored the actions of lever 4 in this respect. Adjacent levers 2 and 3 and also 9 and 10 mutually lock against each other, another very simple locking action - and, of course, 7PULL and 7PUSH mutually lock against each other without needing any actual locking to do so.

[Nick C]

From what I understand it was quite common in SR practice for one shunt signal to apply to multiple diverging routes - as you say SigScribe doesn't allow for this, although you can still get it to lock correctly by setting the locking for the signal lever to apply to the point normal or reversed. You can have interdependent locking on points levers too (e.g. point 5 requiring point 6 to be normal) - it's somewhere in the right-click option list IIRC. It can't cope with push-pull levers or selected signals though!

[Echini]

Hayling Island Signals.

Continuing my work on Hayling Island signals, I wanted to see if I could build an effective interlocking system that could operate the signals in a realistic way based upon the 1957 plan.

The first thing that I did was to create a simple Excel spreadsheet to simulate the interlocking for the ten levers. Initially I needed six cross bars but gradually managed to get that down to only four bars.

The next problem was to work out how to build the physical interlocking system. After a lot of research, I decided that the best solution was 3D printing and acquired a Monoprice Select Mini V2 3D printer. It can print up to a 120mm cube, so only relatively small parts, but certainly enough for my needs. It was also not expensive. Creating each part is a three step process.

The first step is to design the part in 3D. I used FreeCAD, a 3D CAD system that is free to download and quite easy to use for the simple kind of parts that I needed. The final design has to be exported to a special .stl file, which is then opened in a “slicing” program that creates a further .gcode file that tells the printer how to print the part. Cura is another free to download program that seems to be the most popular “slicing” program, which is very intuitive to use.

After reading the manual, watching several YouTube videos, and carefully levelling the base plate, I designed and printed the base of the interlocking system. Because of the size of the base, I printed it in two pieces that would interlock together.

The print process is very precise and also takes quite a lot of time. The base part took over three hours. With a second part created, the two could be fitted together to make a base for ten levers and four bars.

The next step was to design and print the lever tappets with notches in the appropriate places.

Finally the four cross bars were designed and printed.

The parts could then be assembled and tested.

Of course, it was never as simple as that as the first test highlighted a couple of issues where the interlocking was not correct. This illustrated the strength of 3D printing as all I had to do was redesign the specific incorrect part and re-print it. I also found that Home signals could still be set if point 6 was set against them, this was solved by adding a second layer (under the square block) providing an interlock between levers 6 and 7.

While this may not be the simplest interlocking possible, it does work and provides all the interlocking protection required and indicated in the 1957 plan. Previous discussion on this forum indicated that levers 1 and 4 provided direction control for up or down trains. My reading of the plan is that levers 1 and 4 lock points 5 and 6 both as normal and reversed and so they are symmetrical. The only lever that is clearly directional is lever 7 where 7 pull provides for down trains and 7 mid for up trains. 7 push requires point 6 to be reversed and logically requires that all other signals are off.

The next step was to start work on a design for the lever chassis and the levers themselves to include micro switches so that the unit can drive the actual signals and points. The easiest design was to create ten individual modules so that levers,  switches and links can be installed as you go and then the ten units bolted together.

This was the first attempt and I have just redesigned it to give the microswitch sufficient clearance and one or two other minor adjustments. Unfortunately I am away from home for a few days – so it will be a week or so before I can re-print it along with a prototype lever!

More to come.

All the best, Echini.

[Titanius Anglesmith]

3D printing a locking frame, that’s very clever! 

[Echini]

Been working on designing and 3D printing the Signal Box lever chassis. Each lever has its own chassis, which holds the lever, link and micro switch.

Each chassis takes over two hours to print, so printing ten of them has taken a while.
For the time being, the lever is unlocked by lifting it and locked by dropping it into the groove at either end of its travel..
 
The resulting set of levers correctly reflects the Hayling Island Signal Box and their interlocking. As each lever has a micro switch, all the signals can be driven from the levers. The two points can switch the Cobalt IP Digital points using the left/common/right terminals on the micro switch. As the two FPL levers must be used to release the points to change, their micro switches will interrupt the common of the appropriate point switch so that there is always a gap between the change from one side to the other as seen by the IP Digital Motor.
Been giving some thought to how best to mount the levers on the layout. Luckily, there is a place where a control panel used to be, before I made a new one for the whole layout, and it overlooks the points and signals controlled by the diminutive Hayling Signal Box. So, I think that it will work well built in to that space. Tried it and it fits perfectly. The signals here are the old "dumb£ ones from Hornby, soon to be replaced by working signals.
 
Clearly, I will need to do something cosmetic to hide the woodwork but the position is perfect for controlling the station and its approaches.

The levers are wired to two Megapoints servo boards, which provides "bounce" for the signals. The points will be wired directly to terminals 7, 8 & 9 of the Cobalt IP Digital motors - with the FPL levers providing an interrupt between changes. I understand from DCC Concepts that it is not a problem for the 7/8 or 8/9 switching to be constant (rather than momentary) so long as the change from one side to the other is not instantaneous. Using the FPL (which must be reversed before the point can be changed) to interrupt the "common", will ensure that there is always a gap between the change from one side to the other.
The down home junction signal for Hayling, the two up starting signals and the advanced starter signal are coming on but in the last week I have been concentrating on the two ground signals. I acquired a copy of "The Pictorial Record of Southern Signals" by G. Pryor, published by Oxford Press. The book is a great source of information including many photographs, diagrams and specifications. They also do a similar book on the LNER, not sure if they do other areas
I copied the diagram of a typical Southern ground signal from the book and scaled it to be exactly ten times 1/76th scale - so that 1 cm on the diagram equals 1mm for the model. Wow! It is small! After about ten versions produced on the 3D printer, trying to get the angles, sizes and internal passages right, I finally produced something that seems to work. As with the other signals that I am building, all the bearings are 1mm OD tube with an 0.5mm ID and with 0.4mm rod running inside. This gives quite a stable axel despite the fact that some of the bearings are only 2.5mm long.
The light is a DCC Concepts engine lamp mounted in the top of the signal. The plate and arm are Wizard etched brass and the base is a standard base that I have created for all the signals that will fit flush into a 25mm hole in the baseboard. The base contains the servo, it can also house two servos for a junction signal. While the hole in the baseboard is quite large, it does mean that the whole unit can be dropped in from above without having to hook up servos or anything else (apart from long cables) underneath the baseboard afterwards.
Finally, this morning, I managed to get one up and running. Lever 7 on the signal box is a push/pull that controls two ground signals, so I hooked the new ground signal up to lever 7 pull and it worked fine. I did make a video, but it was too large to upload, so this is a still of the signal.

 

The signal is just 10mm high. The servo works brilliantly with a pause as the lever is pulled and a bounce as the lever is returned.

Regards

Echini

 

[The Bigbee Line]

Just found this most interesting thread.  At some point I will get to signals...

 

I particularly like the LBSC lower quadrant signals, the handrail from the platform is in my opinion a most elegant, if not truly functional solution.

 

Thanks for all the entries.

 

Keep safe

 

Ernie