Falmouth Branch - a heretic's view

[philspace]

Spurred on by the comments on this Forum (and the half-price sale, of course!), I have recently bought the Falmouth Branch add-on.

I will, no doubt, be branded as a heretic, but I have to say that I was disappointed to see that many aspects of route-building have not progressed, and that the naivity of Kuju in MSTS days, which was still evident in the default routes for Rail Simulator, is still being perpetuated in Railworks.

In some ways, it is unfair to pick on this particular route because the deficiencies I am about to mention appear and far and wide, even in default routes. Most of my observations are generalisations, and exceptions will, no doubt, be thrown up, but they need to be thought about and addressed.

Land/Earthworks:

When the railways were being built, the railway companies had to buy the land they needed. Where cuttings and embankments were needed, extra land had to be bought to contain the earthworks, depending on the type of terrain, and was fenced accordingly - embankments at the bottom, cuttings at the top. However, on flat land with virtually no earthworks, the absolute minimum of land would be bought and would be fenced as close as reasonably possible to the track-bed (and yes, I do realise that the Falmouth Branch was built as broad gauge!) All the infrastructure (signals, telegraph poles, signal-boxes, plate-layers huts etc. would have to be contained within the land owned by the railway company.

Bridges/Tunnels:

Just as land was expensive, the building of bridges and tunnels was expensive in terms of labour and materials. Obviously, the path of the railway had to be kept as reasonably flat as possible - hence the number of spectacular viaducts we have around the place. However, roads and tracks could be easily diverted sideways or deflected up or down to allow for the minimum size of bridge to be built over or under the railway. Bridges built to more than the standard loading-gauge are rare - hence the problems when electrification came along.

The same applies to tunnels - tunnelling is probably the single most expensive form of earth-working - and many tunnels are portrayed as being too wide and tall, often in terrain here a cutting would have sufficed anyway! Strangely enough, the tunnels on this particular route do not appear to have been enginneered for broad-gauge.

I appreciate the need to use existing assets, but there is really no need to routinely have overbridges with 10-15ft headroom above the train - it looks wrong. The stone bridge used on the Falmouth route is completely out of proportion, even allowing for the broad-gauge heritage, and the headroom is far too high in most cases. I am not a civil engineer (quite uncivil, really), but from a personal point of view, the flattened stone arched bridge looks totally wrong - I know Brunel managed it over the Thames, but I suspect that is more to do with the brick and mortar construction and would not work in stone.

An interesting aside is that several over-bridges have a smaller LOD than houses half a mile away - on lower performance machines with reduced detail settings, this leads to track infrastructure objects being apparently missing.

Track:

Ever since the second half of the 19th Century in the UK, the use of hand-operated points on lines running passenger trains has been outlawed, although the practice was re-introduced by Kuju at the start of the 21st century, and later adopted by RS! All turnouts on main running lines have to be interlocked, usually by a signal box, but sometimes by a ground frame.

Equally, connections onto main running lines from loops and sidings have to be protected from shunting errors etc by some form of trap or headshunt, again interlocked in the signal box.

Speed Limits:

There is a plethora of speed limit signs on routes produced by Kuju/RS - unlike on the road, where every idiot has to have it spelt out, a professional train driver would be expected to realise that while the main running line may have a speed limit of 60mph, that dead-end siding may be better approached at a lower speed, without the need for a sign.

Signalling:

Again it is slightly unfair to pick on this route, as it shares many signalling errors with other routes produced by Kuju/RS - there are books and other contributors to this forum who can give a far better insight to signalling than I can, but I will start the ball rolling.

In general terms, this route is gloriously oversignalled, which given the shortcomings of the RW signalling system, detracts in a big way from the illusion created by the route.

The signalling of the passing stations/loops is excessive - the route represents a sleepy branch-line - one sufficiently busy to have signalled passing loops, but not bi-directional loops. In general terms, the preferred path through the passing loop would be signalled with a conventional signal arm, but the alternative paths would be signalled with, at best, a subsidiary arm, or more often than not a shunt signal. One station features a bracketed starter signal on the "wrong-running" platform with the bracket signal protecting the goods-shed!

Splitting distant signals were a relative rarity, giving advanced warning of high speed junctions, but not on sleepy branch lines.

Sidings with hand-operated points cannot have signals - many countries have point levers indicating the setting of the points, but they are not signals as such, and, as far as I know, they were not common in the UK. This is one of the most frustrating errors on many routes because I appreciate the time it takes to instal them and they are utterly pointless (no pun intended!)


I can feel the arrows coming my way already - "if you think you can do better..." etc., but I do feel that, in most situations, it is just as difficult (or easy) to do something correctly as it is to do it wrongly

However unfairly, I have used this route as an example because it is the first add-on route I have bought since Rail Simulator/Railworks came along and I was disappointed that long-term flaws are being perpetuated. It would appear that rolling stock addons are being produced to a level approaching perfection, but that the background to the underlying infrastructure is being neglected.

Merry Christmas

Phil

[Kariban]

I should point out that some of the prototype infrastructure was built with the idea at least of the line being doubled at one point.

You're not alone in being a bit disappointed at the presentation, but I'm still messing with that route more than any other atm.

[antonyperks]

About the signalling If you look at the signalling records, then you will find a great many signals actually missing from the route!, the signals used have been I think you will find, used to help pathing of the trains by the sim, and as the Kuju rsc signals are limited they have used what thay had to get the job done , i have re signalled the line using the GWR kit available and it is complex, and even that comprehensive kit lacks some arms as used on the line. So I think this attack opn the signals a Little unfair

[FoggyMorning]

Regarding siding signalling; due to the way the AI calculates block occupation and train paths, each siding, or at the very least each "bank" of sidings requires the minimum of an exit signal or the potential for multiple trains operating becomes horribly limited. There is a compromise here between prototypical operation and operational potential. Two other recent routes take the more prototypical approach (WCML and Bristol-Exeter), with the result that it becomes virtually impossible to have more than one train operating in even the largest of yard complexes. Personally I'd rather have the unprototypical, and relatively unobtrusive, provision of excessive shunting signals than hamstrung operational possibilities.

[smarty2]

I haven't got a bow Phil or any other weapons But I agree about the point of continuing errors consistently especially when the tools are available to rectify it at the time of track laying with a bit of effort, I know I am in a minority about angular horrible totally unrealistic gradient changes, but that's my bugbear for this "new" route since they are so obvious.

[philspace]

Regarding siding signalling; due to the way the AI calculates block occupation and train paths, each siding, or at the very least each "bank" of sidings requires the minimum of an exit signal or the potential for multiple trains operating becomes horribly limited. There is a compromise here between prototypical operation and operational potential. Two other recent routes take the more prototypical approach (WCML and Bristol-Exeter), with the result that it becomes virtually impossible to have more than one train operating in even the largest of yard complexes. Personally I'd rather have the unprototypical, and relatively unobtrusive, provision of excessive shunting signals than hamstrung operational possibilities.

I stand corrected on this point - I hadn't realised this was the reason.

Since Railworks is a compromise between operational and visual aspects, would it not be possible to bury the signals underground - I can't agree that they are "relatively unobtrusive".

Best wishes

Phil

[rabid]

I haven't got a bow Phil or any other weapons But I agree about the point of continuing errors consistently especially when the tools are available to rectify it at the time of track laying with a bit of effort, I know I am in a minority about angular horrible totally unrealistic gradient changes, but that's my bugbear for this "new" route since they are so obvious.

Thou art not alone; such vertical dogleg gradient changes make me cringe also, ever since the days in MSTS.
Maybe some kind talented soul will clone the route and make a more accurate improved version.
Overall it is a fun little route despite it's shortcomings and well worth it's half price tag with the included stock.

[bigvern]

As a route builder (not the Falmouth branch I hasten to add) it's only fair to point out the reason for large UIC or US bridge sizes on UK routes is a result of the course terrain handling/vertices. In general even using a 90 degree cutting angle you cannot get something that looks right for the UK loading gauge. So you end up using wide bridges and to avoid these looking odd, tend to set them higher over track level. As the OP says, this is a bit of Kujuism that was also present in MSTS which shared the same inability to get smooth terrain hugging the track. There's a few things we can do to improve appearance such as use of the smooth tool, but until such time as RW gets a finer terrain mesh we are essentially stuck with the problem.

I've already offered my views on the signalling, not having an actual plan in front of me I can't comment if it's oversignalled but it is still quite ludicrous to have semaphores defaulting to "off", clearing behind you for opposite direction moves etc. I suspect the reaon why main aspects were used for sidings is because the default disc signals don't functiion properly and/or still muddled in which type serves which purpose. However given RSC has at least one person on their staff who understands the placement of signals (the guy who kindly redone my signalling on Tanigumi) you would have thought this could have been tidied up. Obviously it's not going to eliminate the dafter failures of the AI and scripting once out on the line, but would help the illusion along a bit.

We still don't know who actually built the route, let's just hope they can take on board the feedback for their next project.

[smarty2]

I haven't got a bow Phil or any other weapons But I agree about the point of continuing errors consistently especially when the tools are available to rectify it at the time of track laying with a bit of effort, I know I am in a minority about angular horrible totally unrealistic gradient changes, but that's my bugbear for this "new" route since they are so obvious.

Thou art not alone; such vertical dogleg gradient changes make me cringe also, ever since the days in MSTS.
Maybe some kind talented soul will clone the route and make a more accurate improved version.
Overall it is a fun little route despite it's shortcomings and well worth it's half price tag with the included stock.

Aaah a fellow anti-angle type person, I was feeling a little out in the cold! Ta.

[philspace]

Good morning everyone,

Many thanks to those who have posted answers to my observations regarding visual aspects of route design.

As Vern has explained, the issue of bridge width being tied in with the terrain mesh is insurmoutable in many cases, given the current design of the graphics system. However, there are still many instances, even on default routes, of bridges being placed unrealistically high in relation to the track.

However, I am slightly more confused about the issue of signals being used in sidings to allow AI interaction. It is obviously true - the practice is so widespread - but I have to confess that I have never seen it referred to in the manuals and videos, or even discussed in any detail on the Forum, so I have to apologise for my lack of knowledge on the matter.

I mentioned earlier that I thought the signals were, in my opinion, visually obtrusive, but I do accept the need to expand the operational possibilities of a route. I wonder, therefore, whether the visual aspect could be improved by either pushing the signals underground or developing an invisible version! They do not appear to present any useful input to the Player Driver, except to indicate the setting of the points, which, in a yard with hand-operated points, would be done visually anyway.

Likewise, I gather the use of manually operated points on running lines is also tied up with being able to perform certain manoeuvres in scenarios. To improve the visual appearance of a route, would it not be possible to push the operating lever under the ground so that it was barely visible. An alternative, and possibly better solution could be that, in certain situations, the visible lever is removed and the point changed by clicking on the tie bar, for example.

The comments about gradient changes are interesting and understandable - the mechanism exists within the editor to automatically follow the terrain with the track, so presumably it would be an extension to the same part of the code to automatically smooth a transition in gradient.

Best wishes

Phil

[bigvern]

The issue of accessing sidings off running lines remote from the signalbox was exercising my mind while working on a little single track project. Normally this would be done by provision of a "ground frame" which, on a single track line, would be released by an Annetts key in conjunction with the token or train staff. Once the train is in the siding and points normalised and FPL'ed (locked) the token can be put in a remote machine thus releasing the main line for other train movements. Once the train is ready to come out, the reverse applies. Obtain the token, Annetts release, out on the "main" line secure the points and proceed with the token authority to the next block post.

However for practical applications in RW we don't have anything approaching correct GF operation, we don't even have a dummy GF we can put lineside. However I 100% agree that substituting hand points should not be done in these circumstances. In fact, on the little route I'm working on I'm going through and correcting that at locations where I put in manual points off the passenger running line. Manual points are *not* an appropriate substitute for a ground frame. Quite how we should approach this in a sim is an interesting subject. If you use an automatic point then the AI probably needs signal controlling access to and/or from the siding, whereas in reality - on a single track line - you wouldn't have signals. Possession of the token is authority to proceed.

[philspace]

Good morning Vern

I'm just thinking out loud here (and probably teaching my grandfather to suck eggs as well!), but have you tried dragging the operating lever away from the point and cobbling a ground-frame together with dummy levers for locking - the lever still functions half a mile away from the point itself!

I've successfully sunk the operating levers into the ground on the default S&D, and just left the top 6 inches showing to allow operation without the worst of the visual distraction.

I'm going to go back and try sinking the ground signals in sidings on Hedborough, for example - the problem is finding a scenario in which it makes a difference.

Best wishes

Phil

[bigvern]

I'm holding out for realistic operation Phil (!), probably a distant hope when driving along and you get a "Petrolhead" achievement. Still, if it's any consolation, MSTS didn't do it any better and Trainz effectively they're all hand points!

[Kromaatikse]

This is not something I've tried, but I believe you can get away without signals for a siding provided you never try to use it as a refuge. So in principle you can do shunting within station limits (or in section) under Absolute Block, but you can't shunt off to let another train pass unless you have a signal protecting the siding exit.

[Kariban]

Good morning Vern

I'm just thinking out loud here (and probably teaching my grandfather to suck eggs as well!), but have you tried dragging the operating lever away from the point and cobbling a ground-frame together with dummy levers for locking - the lever still functions half a mile away from the point itself!

If it comes to that, drag the base underground and then all the levers together to make the ground frame would have a hard time finding the right lever to actually use, I guess.