jamespetts

Dagworth - can you elaborate on what exactly "full and clear access" is? Are you referring to a particular degree of distance from the edge of the baseboard to the back of the baseboard? Are you referring to vertical separation between multiple levels? Without being able to quantify this, I am really not in any position to do anything meaningful about it. Incidentally, putting things in capital letters will not persuade me. It will make me less likely to accept what you are claiming, as I have found (through your favoured experience) that people who, when pressed for detail or explanations, respond by emphatically repeating the original point without detail or explanations usually do so because they are unable to provide detail or explanations because their claims are unfounded. In relation to testing helices, I may yet do that - whether I use those specific piers I am not sure, but I have not dismissed the idea of testing gradients/helices at all.

Junctionmad, there is no reason to believe that there can be no middle ground between, on the one hand, a full 3D mock-up, and, on the other, just providing some (even approximate) quantification of the specific constraints to which you implicitly refer. I have asked specifically about the extent of level separation necessary quite a number of times now, and have had no clear response (and such responses as I have had have been inconsistent with other sources, and I have not had, despite asking, any explanation of why the views of those expressed here should be preferred to other sources). There is repeated reference to "hidden track". Every time that this is referred to, I point out that there will be very little track that is hidden. Every time that I do so, this is totally ignored, and a few posts later, reference is again made to "hidden track" without reference to what I wrote previously. Do you think that track is "hidden" merely because it is on a baseboard below another level? If so, does this depend on the height separation? If not, that is absurd. If so, what degree of height separation makes the lower level "hidden", and why this number in particular? Again, as to the helix, I have had (and specifically referred to, with links) conflicting information on this. Instead of trying to engage with the conflicting information, you and some others appear to do no more than demand in positively aggressive tones that I unquestioningly defer to your view. I will not unquestioningly defer to anyone's view about anything, ever. It would not be rational to do so, especially in the face of conflicting information. I will consider people's views and the reasons for them and make an independent judgment based on those reasons and such other information available to me about the underlying merits of those views. If you want to persuade me that your views are correct, you will only succeed in doing so by providing a specific empirical and reasoned basis for doing so. The more that you demand that I accept that you are correct without doing so, the more that I will be wary of taking anything that you say seriously. In terms of planning wiring and point motor access, what sort of planning had you in mind? For example, what do you think that planning point motor access entails? Do you do this for your own layouts? If so, it would be helpful to have a description of what you do. Likewise the wiring - do you mean a schematic wiring diagram, or a geographical plan of wiring runs? I was considering preparing a schematic wiring diagram at some point, but I think that I will need to know more precisely how some of the components work in order to do so in a useful way. Again, do you do this for your own layouts? If so, it would again be helpful to know the techniques and systems that you use for this. Edit: Incidentally, it really is very important from a practical perspective to understand precisely the reasons that it is asserted that the lower level will not work. If the problem is the vertical separation between that and the upper level, this can be solved by abandoning the N gauge layout if necessary and increasing the vertical separation (although, of course, I will need to know to a reasonable degree of precision what degree of vertical separation is necessary in order to implement that degree of vertical separation). If the problem is the helix, then this can simply be tested empirically; if this really does not work, then the layout would not need very great changes to separate upper and lower levels as some have suggested, but it would seem silly not even to try when this can fairly easily be changed after the fact and when the information as to whether it is likely to work is conflicting.

Junctionmad - I agree. Although computer software at its current state of advancement might not automatically prevent all layout design flaws, it certainly prevents some and enforces the geometary much more effectively than sketching a track plan with pen and paper ever could. I have found the tool invaluable. As to your last paragraph, as I have stated before many times, I am willing to consider any advice (as can be empirically demonstrated by the extent to which I have in fact changed the layout in response to that advice): but being willing to take advice is not the same as the unquestioning deference to advice, especially non-specific advice, that some appear to demand.

Incidentally, Chimer/Chris - I missed your points 2 and 3 earlier. I have just spent some time attempting to implement your suggestion no. 3, but I cannot get the geometary to work out on the bottom left hand corner: there is not space to have the various crossings without making the curve radii there too tight. As to no. 2, the idea of these crossovers is that they allow a train in the main storage yards to enter on the relief lines. Edit: My posts crossed with Denbridge's and Junctionmad's above posts. Denbridge, it seems that you are the one who does not listen: I have stated repeatedly that bland generalisations are worthless, yet you repeat bland generalisations ever more emphatically. That is not constructive or helpful. Merely stating that "it will not work" tells me nothing about how to know the precise parameters of what does and does not work so that I can plan something that does work, nor does it assist me at all to resolve the conflict between your advice/information and that of other sources. Your more specific contributions have been very useful. Please do not mar that by demanding unquestioning deference to the generalities that you assert. Junctionmad - the reason that I do not combine the upper and lower fiddle yards is that there is not space to do so: either for enough tracks in the fiddle yards, or for the necessary gradients/helices. It really does not assist to repeat over and over that you do not think that the lower level will work without explaining the reasons, and, where the information that you give in support of your reasons conflicts with other sources, give me some empirical basis (or accept that there is none) for preferring your view/information over that from other sources. Without that, I am not equipped to make an independent decision, and you are simply demanding unquestioning deference, which is not helpful at all.

Mightbe - that is not a constructive or helpful response. What purpose do you think that that serves? Being able to understand the actual nature and degree of the various constraints to which people are referring is not "trivial": it is absolutely necessary if I am to have any use out of what is being suggested at all.

Thank you for your feedback. As ever, the specific details are useful, whereas the more generalised exhortations to be less complex in general or general assertions that what I have designed will not work (without any clue as to how to discern the closest thing to what I have designed that will work, and why the threshold is there rather than anywhere else) are less useful, as I can only sensibly plan within specific constraints. It is impossible to optimise within non-specific constraints. As already noted, the plan is for the helix to have an access from beneath. I do not think that anyone has actually given me any real idea of numbers for level separation and how any degree of level separation actually makes a difference. I should also note that the idea that I have a great amount of "hidden" track seems to have persisted, even though I have made clear that fiddle yard track on the lower level is no more hidden (apart, possibly, from having a curtain of some sort which could easily be moved out of the way) than track on the upper level. One possibility that might be contemplated is to have a greater separation between the upper and lower levels of this layout than between the upper level of this layout and the N gauge layout above, but to have that modularised into perhaps 3-4 sections and readily demountable. Paul: in relation to your layout, what was the depth at the main flat part of the fiddle yard? What was the diagonal to corner distance? It looks as though all of those slightly higher level tracks would readily block access to the rear yards beyond, whereas the plan for this layout is for two essentially flat areas with a substantial separation between them, connected only by an open sided and open bottomed helix at one end. I did look at your thread briefly, but since it was 143 pages long, realised that I could not sensibly read the whole thing, and could not easily find the most important parts. If you could direct me to them, that would be useful. Edit: As to the slow/fast pairings, they are paired by use on both the Great Western Mainline and the Midland Mainline (two of the areas that this layout could represent), so it would be odd to have them paired by direction in this layout. It would also be harder to design the station, since, at present, the relief lines feed the slow platforms and the main lines the main platforms.

Thank you for your replies. Gordon - I should be very interested in the results of your test. Incidentally, what level separation did you originally use?

I am not sure that I have that many cornflake packets.

Thank you for your thoughts. As noted earlier, the plan is to allow access to inside the helix from below by having a hole in the centre, as well, of course, as from the side. I accept that access from the top would be more convenient, but it is hard to see how access from below and the side between them would be impossible or unworkable. As to the radius, I do not think that anything short of data is really going to help me in this regard, given the conflicting information that I have had so far on the point: simply re-stating that it will not work without data really does not assist at all. In relation to lower track access, may I ask: what do you believe is the correct means of calculating the necessary height for access to any given width? Given that I have had conflicting information on this, it is necessary to be rigorously empirical in approach. In relation to the test track - this is different to a smaller layout in that it does not need to have enough time, effort and space invested in it to be operationally satisfying to me: it merely has to serve the utility of allowing me to test things and need not have, for example, scenery or point motors.

TonyMay - thank you for pointing that out. I think that Dagworth's latest design for this seems to be the best so far, so, when I get a moment, I will update my SCARM file to reflect that design, which does not appear to have the flaw to which you referred.

Thank you for your thoughts. I have doubts that a separation of only 150mm between the levels of the OO gauge layout would be workable given the need to access all of the pointwork for wiring/point motors on the upper layer. However, having spoken this evening to somebody at The Model Railway Club, I am assured that a 400mm separation should be sufficient provided that the layout be wired sensibly.

Thank you for all those who have contributed further feedback: it is much appreciated. My apologies for having mis-identified the N gauge helix as HO gauge; I thought that I had searched specifically for OO/HO gauge helices and the gauge was not apparent from the title. However, the fact t hat this particular helix is N gauge does not, of course, tell us anything about whether a 3rd/4th radius helix would be workable in OO gauge. The other link is definitely of an OO gauge railway layout and definitely of a 3rd/4th radius helix and a steam locomotive hauling an 8 carriage train up it, albeit with the assistance of PowerBase. I should note that, if it comes to pass that the trains cannot climb the incline, even with magnetic assistance, on the 3rd radius inner curve, the track could always be singled and the 4th radius outer curve used, albeit at a loss of some capacity. As to the descent to the helix, the 70mm is railhead to railhead height. This could easily be increased to 76mm (the separation height of the commercial helix) whilst allowing the gradient on the flat to remain at a reasonable percentage. I do not understand why there should be any problem in particular with doing this. Thank you for pointing out the error in the original calculations, however. In relation to helix maintenance, the problems described would, of course, be common to all helices. Given that these are a common feature of model railways/railroads and have apparently been common in the US for decades, it is simply not plausible that these are insurmountable obstacles that make such an arrangement "completely impractical". I went along to the Model Railway Club in London this evening, incidentally, and picked up an application form. The people there were most helpful. There was a chap there who knows well the fellow who runs the McKinley Railway, who was confident that a 3rd/4th radius helix (which he advised be laid with Setrack on account of the tendency of flexible track to kink on tight turns, albeit not the steel sort owing to its propensity to rust) would be workable, and who also usefully advised connecting point motors to plugs so that they can easily be replaced in the event of a failure. He suggested building a test track of about 1-2m in length with one or two points to familiarise myself with the skills of track laying and basic wiring (and to have a useful test track separate from the main model), and thought that a 400mm level separation (provided that the wiring be done sensibly, e.g., with plugs for point motors) between layers should be sufficient, which may well make it possible for me to build the N gauge layout after all. Incidentally, I should be very interested to meet a Bhuddist monk with the financial power of Elon Musk, although I doubt that such a person would spend much time building model railways.

I wonder whether reducing the separation between all levels to 400mm would suffice to overcome this without the need for pulleys? That would allow the lower level of the OO gauge layout to be at 800mm, the upper level of that layout at 1,200mm and this layout to be at 1,600mm, which is viewable for me. Reaching to the back might then be facilitated with a simple kick stool. Does this seem sane? I should not want to have to abandon this, as this layout is the only way that I can think of of starting with something smaller (in both senses of the word!) than the large OO layout, but at the same time be a layout that (1) will fit in my proposed shed; and (2) will be one that I will actually want to keep in the long term.

Thank you for your replies. I attach for reference an output from SCARM showing just the helix, but both upper and lower levels at once, as it might have been difficult to see the transition between them with the two levels separated. As can be seen, the helix only needs to pass under the first track at 7,100mm from the far left wall (or 400mm from the far right wall), not at the point where I have placed the tunnel start markers. At that point, the level separation is specified as 70mm and the gradient on the pre-helix section is < 2% (as measured by SCARM). As to the height separation between turns on the helix, I have taken this from this commercial product, a 3rd/4th radius helix kit for OO gauge, which gives a separation per turn of 76mm. Somewhere between 5-6 turns would be necessary to achieve the 45cm drop in this case (and if something less than an integral number of turns is necessary, one can always continue the same curve on the flat at the bottom, of course). As to steam locomotives using a 3rd/4th radius helix, is a video of a Hornby B1 hauling an 8 carriage train up just such a helix (look at the end of the video for the haulage section), albeit on the outer (4th radius) curve. This person is using the DCC Concepts Power Base (as revealed in another of his videos). is an American video (of fairly poor quality) showing an HO scale steam locomotive pulling a very long train up a helix of 18" (457.2mm) at some speed. Mightbe - the reason that I do not always unquestioningly defer to all suggestions made here is that they sometimes contradict other information, such as in the videos above. I am afraid that I do not think that there is any reason to criticise empiricism - it is really the only reliable way of dealing with conflicting information. In relation to whether the helix needs to be double track - it would be a lot more convenient for it to be double track, as a single track helix might well create bottlenecks. However, if a single track helix would be the only way of making this work, it might be worth considering. Given the conflicting information and incomplete data about the extent to which a double track helix (with 3rd radius on the ascent) is likely to be workable, I shall probably have to test this (with and without the Power Base) to see whether this is the case before finalising the layout design. I shall be very interested in the results of Denbridge's PowerBase tests in due course. Incidentally, in relation to the level separation: I wonder whether 400mm would be sufficient? That would allow me to have the N gauge layout at 1600mm, which is acceptable for viewing given that I am quite tall, and reaching to the back of that could be facilitated with a simple kick stool. The lower level would then be at 800mm and the upper level of this layout at 120mm. This would also reduce the number of turns required in the helix. Incidentally, well spotted on the single/double slip error: I have now corrected this.

Dagworth - I am not sure that that is right: I have seen a number of examples of successful 3rd/4th radius helices (for example, this one, showing a locomotive haul a train of 14 freight wagons up such a helix and start from a stand without incident). The American forum was discussing helices for very long trains (as is common in US model railroading, seeking to model the very long freight trains that are common in the US; that thread referred to trains of 20 cars' length). The point about "stringlining" appears to be common to all tight turns for very long trains and is not specific to helices. Since I intend to be operating short trains on this helix, it does not seem to be immediately relevant. Do you have any experience of helices yourself, or are you going on what is posted in the forum threads to which you referred? If you do have personal experiences of helices, I should be very interested to learn more about them. Edit: I notice the following from the US forum thread: . For reference, 20" is 508mm, whereas 3rd radius is 505mm.

Thank you for your thoughts - a drawing would be most interesting. Can I ask - has anyone actually tried or used a layout on a pulley or one with an adjustable height? From what I have read, those who have tried them tend to abandon them very quickly, leaving them in the up or down position because hoisting is too difficult or unreliable, converting them to fixed or dismantling them entirely. It would be splendid to be able to have this in addition to the planned OO gauge layout, but not at the cost of making either inaccessible.

Thank you all for your ongoing feedback on this: it really is most helpful. The tight helix seems to have been the most important problem with the last design. I have found that, by adding a small bulge to the baseboards near the end of the shed, I can have a 3rd/4th radius helix offset sufficiently from the reversing loop on the other side of the room to allow (just about) sufficient space to pass between the two sides of the layout, with judicious angles used in the baseboard outline. I have thus produced a revised plan. Here is the upper section: and here is the lower section: . I have also made some other minor changes: I have added an extra road to the engine shed, and altered the sidings around the engine shed a little as I was having trouble working out how the locomotive coal train would be shunted in the previous arrangement (the pilot engine would have had difficulty running around the locomotive coal train's brake van in the engine shed area). I have also added three auxiliary storage sidings in the fiddle yards so that rarely used trains (such as the locomotive coal train) can reverse into them and then be stabled out of the way to give some extra storage space. Finally, I have added an extra crossing in the carriage sidings to allow one pair of carriage sidings to be used more readily for dividing/combining shorter rakes of carriages/loose carriages/strengtheners. Denbridge - I should be very interested in the results of your branch gradient PowerBase tests. Your previous tests were very useful and interesting. Thank you for the information on GWR carriages. I think that that is broadly in accordance with what I had imagined the position to be from what I had been able to find out so far. In relation to the tracks at the back of the terminus, save for those which descend into the now enlarged helix, the furthest reach is ~1.0m just opposite the new bulge (which might only be on the lower level), otherwise 950mm for most of the length of the station. Since these boards will be lower (1.2-1.3m high, depending on whether I can build an N gauge layout above it; I will respond to the suggestion regarding the shelf in the N gauge layout thread), this should not be beyond my ability to reach. I will have to make a note to make sure to make any overall roof removable. In relation to a platform without access to the carriage sidings, I have checked the plan, and cannot find such a platform. Can you let me know which one that you think that it is? As to the descent to the helix, this is intended to start almost immediately after the junction, and that gives me a <2% gradient on the upper level, which should look realistically shallow as well as be reasonably easy for the trains to negotiate. In relation to building time, it is very difficult indeed to know what to do with imprecise information. If there is not a linear relationship between any given measure of complexity/size and build time, then I might as well build a large layout rather than a small layout, as there may be little (proportionate) difference in build time between the two. The only precise information that I have suggests a manageable ~177 hours of work to complete the track laying and wiring of the upper level. As to how long that I should like to spend building it - that is hard to answer. In one way, it would be splendid to be able to summon elves to build it all in the twinkling of an eye and have it up and running instantly; but on the other hand, I suspect that I should then be missing a certain joy of building it myself and seeing how all the components fit together, planning the logic of the wiring and so forth. The testing, too, I imagine will be most interesting. In terms of data, I am looking at some well known model railway Youtube channels. Dean Park is a good example: he started in June 2012 according to his first video posted in December 2012, by which time he had completed the baseboards and a single circuit of track, together with most of the tracks for the station. Later in December, by the second video, he had completed the second loop. By March 2013, the basic track layout had been completed and he had started work on the platforms. The ballasting and scenery appear to have been largely complete by October 2014, albeit the layout was further refined for sometime afterwards. I looked at the videos of the original Everard Junction, but the first video shows the track already in place, so there is no good data point for this layout, and the current layout has not been completed, although I note that what appears to be a gargantuan 14 road fiddle yard has now been laid, corked and wired (at least so far as the droppers are concerned). He described that as having taken a long time, stating that he started work on the fiddle yards "shortly after" completing the back scenes in his previous video. That earlier video was posted on the 5th of December 2017. Silver Sidelines - you suggest that using DCC will make hauling trains up gradients more difficult than with analogue. Can you elaborate on why this should be? I have not heard of this issue before. The videos show a Hornby BR 8p and 7 carriages managing a gradient on a curve (albeit slowly), although it is not clear what the gradient or the curve is from the video. I was not planning on having trains longer than about 6 carriages (or the equivalent length in wagons) using the helix. Hopefully, this should be manageable with a 3rd/4th radius helix (although unfortunately the upward direction will have to be the 3rd radius).

The reason that I am trying to calculate how long that building the layout will take is because it has been suggested that building the layout will take a long time and that I should take that into consideration when choosing what sort of layout to build. I need to be able to quantify that to some extent to be able to make any meaningful decisions based on it. In any event, I have now further revised the layout design. For some reason, I cannot attach files, so I have had to upload them to a server. Here is the revised upper level: and here is the revised lower level: As you will note, I have moved the engine shed to the far corner of the layout. I was concerned about reach/access distances, but, by altering the station throat layout and moving the lower level station back, I have been able to achieve the recommended 900mm front to back reach distance in the engine shed area. This has allowed for a slightly larger engine shed and one with more direct access to the station throat without conflicting with the carriage sidings. I have also shortened the platforms a little: the longest pair of platforms can still take 12 carriage trains (indeed, it could take a 13 carriage train, but that would not fit in the carriage sidings), and the next longest two pairs of platforms can take 11 carriage trains. The carriage sidings has one road that can take 12 carriages, and the others (bar the short one) can take 11 carriages. Moving the depot has enabled me to straighten the fiddle yard sidings as recommended, which are shown on the plan. I have also altered the layout of the fiddle yard so that the main line trains run around the outer reversing loop whereas the relief line trains run around the inner reversing loop. I have extended the length of the fiddle yards slightly so that the ends of them will oversail the workbench a little more, but at ~1.2m in height, this should not be too much of a problem. I have also slackened the curves at the left hand side of the layout plan. The result of all that is that trains on the main lines (as opposed to the relief lines) and in the carriage sidings never have to negotiate a curve of less than 572mm radius (4th radius in Hornby/Setrack terms), and never have to negotiate, even in the fiddle yards, short radius points. Trains on the relief lines in the fiddle yards must be able to negotiate 3rd radius curves and short radius points, but an express train dispatched from the station onto the relief lines for any reason can cross onto the main lines before circumnavigating the reversing loop. Short and medium radius points are, as before, confined to the fiddle yard/non-scenic areas. Only trains using the lower level will have to negotiate 438mm (2nd radius equivalent) curves and only trains passing between the upper and lower levels (via the helix) need to traverse any sort of gradient. Trains using the lower levels will either be electric stock or tank engine hauled suburban stock with trains no longer than 6 carriages, which (perhaps with a bit of magnetic assistance on the helix) should be able to manage the tight curves from what I have been able to discover so far. Hopefully, with the long trains (whether in the sidings or on the main line) no longer having to traverse any curve of <572mm, it should not be necessary to use magnetic assistance other than on the helix, and only a limited range of stock will need to use this. This revised plan is still usable with GWR stock owing to the ability to use 12 carriage trains. It remains a "track heavy" layout, but that is really a matter of taste rather than practicality. Moving the depot does not allow the town scene that I had imagined, but has helped to solve a lot of practical issues with the previous version, so it is probably worthwhile. If I take a fancy to Midland/LMS/ex-Midland stock in the future, I could always just switch out the stock on this layout and have a whole new operating experience. (Indeed, one might even do the same with the LNER/GNR, as that, too, had a connexion to the City Widened Lines). In any event, I should be grateful for any feedback on the revised plan. Your feedback so far has been most invaluable.

Thank you all for your continued thoughts on this. I am becoming a little concerned that it might not be possible to fit this and the other proposed layout in the shed and have a sensible height clearance between all three levels (one level for this one, and two for the other), and, if forced to choose, I should prefer the other layout to this. The issue is that the minimum recommended distance between levels is, from what I have found so far, about 45cm. The lowest sensible level for the lower layer is 80cm, which would mean that the upper layer of the other layout would have to be at 125cm high. Adding a further 45cm to that would make this layout 170mm off the floor, which would be too high for anyone other than a giant: I am 185cm tall, so could see the front of the layout, but would not be able sensibly to reach to the back. Somebody suggested a pulley system, but briefly looking into this suggests that this is frought with problems. Can anyone think of a creative practical way around these issues? I should not want to have to abandon this idea.

Thank you for all of your replies. I am aware of the King's Cross layout, as I saw the Youtube videos over a year ago now. It is indeed impressive. I counted 113 turnouts (including diamonds and slips) on that layout, although I may be off by one or two. If one is using this to calibrate the rate of progress that I am likely to make on this layout, that does not seem to be plausible, as that would suggest a rate of one turnout every 3 months on average for the last 30 years (installation and wiring; I use turnout count rather than linear meters as I do not have the measurement of the trackage on that layout in linear meters). I note that it is an analogue which would make the wiring on a larger scale layout more complex than it would be in DCC (not that DCC wiring is not capable of being complex). This seems to be entirely out of line with the other data point so far available, and we do not know how intensively that this was worked on over that time: there might have been gaps decades long with no work at all. In relation to curve radii, the data by those who have very helpfully tested on this thread suggests that this is not an insurmountable obstacle, although it does pose challenges and may require extra adhesion on some locomotives. In relation to baseboards, as I have stated previously, I intend to have these made professionally, as I do not have any skill for woodwork, so the difficulties of the baseboard construction would not of themselves be something with which I need directly concern myself. I do not think it practical to attempt to make this layout modular or portable. In relation to a pulley system for the possible N gauge layout, this seems to be fraught with difficulty. If it were achievable within reasonable parameters, that would be potentially attractive, but I do not know whether the roof could support the weight, whether it would be too awkward to use, whether a sufficiently reliable system could be devised, or whether the equipment would need more clearance than is reasonably available.

Thank you all for your replies. To deal with the issues in an orderly manner: the first difficult question is the extent to which the shed should be built to fit the model railway, or the model railway should be built to fit the shed. The latter may result in a quicker construction of the shed, but would be unsatisfactory if a small adjustment to the shed that would be relatively inexpensive to order now but which would be impractical to alter after the shed has been built would make a large difference to what I am able to do. The main constraint is the width: to fit between my newly installed, concreted in washing line and the fence, the maximum external width has to be 2.9m (giving an internal width of 2.7m; I might have mistakenly given the internal width as 2.5m above). The length is a little more flexible: the current planned length of 7.7m external (7.5m internal) was arrived at after consideration of what would not excessively overshadow the rear of the house or conservatory (albeit I do not much use the conservatory). It could possibly go up to about 8.0m externally (7.8m internally). Within that, there are then the constraints inherent in the idea that I am trying to implement, being a main line terminus in London with an element of the Underground (sub-surface lines), together with engine sheds and carriage sidings for the main line station. This would inevitably require a through station for the Underground, but a terminus station for the main line. That is why I thought initially of something like Paddington, with its platforms 15/16 (and originally also 13/14 - and I note that the last time that I visited Paddington, platform 13 had been demolished, but I digress). However, this would then require somewhere for the trains disappearing down the Underground lines to go, which is why I had designed initially a height separated reversing loop and later (on realising that the height separation was inadequate) a helix. If there is to be a height separated level, I might as well make use of that and have an interesting area on the lower level, too, with an Underground station representing somewhere on the City Widened Lines (which would also have the advantage of allowing the representation of multiple companies' rolling stock, such as the LNER N2s and suburban carriages, to add some interest and variety). Given the desire to automate the layout, reversing loops are essential, and there is only so much width available, so it is not possible to have two 3rd/4th radius loops opposite one another, albeit it is (with a narrow section as depicted in the latest design) possible to have one of the 3rd/4th radius reversing loops opposite a 2nd/3rd radius helix. The need to have the level separation at the buffer end of the terminus for the Underground lines then dictates that there needs to be a linear gradient between the main lines and the buffer end of the terminus, which then puts a constraint on the minimum distance between the point in the station throat when the lines to the Underground diverge and the end of the platforms. On the current design, that is either 2.3% or 1.9% (depending on whether one starts from where the track becomes straight or whether one starts further back towards the junction), so there is room for this to be shorter, but not much shorter. With Midland length platforms, it would be at around 2.6%, which is probably acceptable for the shorter trains that would be using it. The initial idea of a Paddington like arrangement with a separate Underground line into the station and locomotive change facilities for suburban trains was potentially problematic because of the need to rise from the lower level, have flat platforms and junctions, and then descend again, there being insufficient space to do this satisfactorily, so I revised the design to something more akin to St. Pancras or King's Cross's access to the City Widened Lines (as distinct from Paddington's access to the Inner Circle), removing the upper Underground platforms and electrification, and sending steam trains down into the lower section to join a more interesting semi-terminus on the City Widened Lines section. My strong preference to avoid hatches and duck-unders and strong dislike of unrealistically short trains also impose constraints on the design, the former requiring a U-shaped layout and the latter requiring a certain minimum platform length. It is from those basic ideas and constraints that the current design has evolved. As indicated above, a further possible evolution would be to shorten the platforms and carriage sidings and represent a Midland/LMS terminus rather than a GWR terminus, as the Midland ran shorter trains (no more than 9 carriages from what I have discovered). This would potentially allow the engine shed to be moved onto the same plane as the station and the carriage sidings to be altered (although I am not sure quite how or to what extent) to allow for a straighter fiddle yard and earlier termination of the station throat so as to give gentler curves on the main line section (although there would still be the need for the 3rd radius curves in the reversing loop). One possibility that I have yet to evaluate properly is whether this would also allow the helix to be offset from the upper level fiddle yard reversing loop sufficiently (by shortening the fiddle yard) to allow the upper/lower level helix to be 3rd/4th radius, which would reduce both curvature and gradient, although it is not clear that the awkwardness of this arrangement, even if offset, would outweigh such difficulties as there might be with magnetic adhesion systems for the relatively small number of locomotives that would have to be equipped with them. What I will need to do is produce a modified version of the layout plan based on Midland train lengths to see what this would allow me to do. I would then need to consider whether slightly lengthening the shed would allow me to achieve similar advantages for GWR train lengths and what the preferable option between the two is. I note that one of the advantages of modelling the Midland mainline in LMS days is that there is very good availability of suburban rolling stock, whereas this is not the case for the GWR, where I should either have to use old (Triang era) suburban stock, or convert B set brake carriages with custom sides. As to the idea of a cassette: given my preference for automation and the space that a cassette would take, as well as its lack of reversing ability, I do not see that as being feasible at present. In relation to build time, I note that the total length of trackage in the upper section alone is 177,374mm (or 177.3m), which, at 3h/m would take 177 days at 3h/day, so the upper section could, if this extrapolation is correct to within a factor of about 2.5, sensibly be completed within a year, and the upper section would be operable in its own right even if the lower level has not been completed. I agree with the idea that there is great advantage to proceeding in a modular fashion. As stated above, I am not quite sure what to do about attempting to build a smaller layout first. I had been quite keen on the idea of the N gauge layout proposed above, but I am increasingly concerned about whether it is possible to get three levels (whether connected or not) all with adequate height separation. Taking the suggestion to research US multi-deck model railroads, I find that the recommended separation between levels is 45-50cm, which is good for two decks, but makes the top deck too high (almost as tall as me, and I am quite tall) for three decks. I now wonder whether the N gauge layout idea would cause more problems than it solves and whether I should be better off either just experimenting with loose track or perhaps wiring up one of the old layouts that I have in my attic to practise wiring skills at least before moving onto the ultimate planned layout. As to cunning plans, I should be very grateful to know more about such things, as I am generally fond of cunning plans. As to model railway clubs, I belonged to one as a teenager, but not at present. I am considering joining one, however; either one near where I live (which appears to be very small and has not been able to put on any exhibitions lately owing to lack of space, and which also appears to concentrate on finescale), or possibly The Model Railway Club, as its premises are not too far from where I work in central London. Finally, as to prefabricated platforms, the idea was very much to start with prefabricated things and improve on that as and when my modelling ability and time develops and allows respectively (which is why completing the upper section, requiring no Underground stock, first and modelling the Midland, which has a full range of ready to run suburban carriages, seems attractive at present).

Thank you for the suggestion. I should really want to avoid a lifting flap if at all possible to make the space easier to use (and easier to escape in case of fire); also, I am very keen on modelling a terminus; but did you mean only for the Underground lines to be continuous? How would you imagine that working; having the Underground lines come off on the inside?

Thank you all again for your further replies. Taking these things one at a time: SCARM tells me that the total length of all track in this layout amounts to 316,942mm (that is, 316m). At 3h/meter, that would work out as 950 hours, or 316 days at 3 hours a day, suggesting that the track and wiring could be completed in 1-3 years, rather than 5-10 as earlier suggested. I realise that this is a very approximate measure (which is why I have used a magnitude of 3 margin of error), but it is the best available from such empirical data that there are at present. This might be an overestimate owing to the length of the platforms and carriage sidings in comparison to the points, although that might be balanced out by my lack of experience (although that significance of that would diminsh with time, of course). In relation to helices - I was indeed referring to UK rail modelling: it is most interesting that they have been around in the US for rather longer. It is interesting that even in railway modelling/"model railroading", fashions drift eastwards across the Atlantic with time. In relation to the SCARM file, I cannot find any way of exporting (as opposed to printing) with a grid. However, I can give you the measurements: the basedboards (marked in the blue outline) are, at the top, 7.5m on the horizontal and 1.1m wide at the widest point, narrowing to 0.9m at the narrowest part of the station throat. On the left, the width of the straight section is 0.4m. At the bottom, the width opposite the turntable is 0.67m, the width opposite the carriage sidings is 0.84m and the width at the widest point of the helix is 1.2m. The length of the lower section is 6.5m at its longest. One thought has occurred to me this evening about one possible way of making a layout to this basic specification consume less space, which is to model the Midland mainline rather than the Great Western mainline. Trains on the Midland mainline had, by my research, only up to about 9 carriages (rather than up to about 13/14 for the main expresses on the GWR), and so would allow a platform length of ~3.1m, shaving about 0.6m off the ends of both platforms and carriage sidings.I do not think that that would allow carriage sidings on the same plane as the station, but it should at least be easy to fit in the engine shed in the same plane as the station. The carriage sidings might possibly partly occupy the rear wall area and thus leave the entire window side wall free for storage yards only. A Midland mainline has some useful features: it allows for a plausible connexion to the City Widened Lines, there is an excellent range of ready to run stock that would have run on it (the main missing item being the Fowler 3p condensing tanks - but the equivalent are also missing for the GWR, albeit available in kit and 3d printed form, the latter taking a standard Bachmannn 57xx chassis, thus requiring only painting skills to complete, only the LNER/GN having a condensing tank of any sort), and there are good beginnings of Midland Railway era items of rolling stock to enable me to Edwardianise the layout in good time. However, I wonder whether this would make much of a difference. The engine shed might be moved to the top left hand corner in any event (and access made easier by shaving off more space at the front of the baseboard in this area, skewing the lower level station inwards). This might well not allow for gentler curvature, although with 9 carriages, this might not matter much according to Denbridge's tests, and this difficulty looks likely to be able to be overcome with magnetic adhesion technology and/or ballasting. The most significant difficulties aside from corner traction/friction that have been identified all relate to the relationship between the upper and lower level and/or general size of the project, neither of which would be (or be much) alleviated by the expedient of shortening the platforms and train formations. Moving the engine shed as described above would, by my estimation, allow a minimum curve radius of circa 580mm (22.8") on the fast lines (where the long trains are likely to be). Given that I am particularly keen on a layout incorporating a main line terminus with a mixture of express and suburban passenger operation incorporating a connexion to the Underground, it seems difficult to do away with the two level design. My real concern at present is the height separation between the base boards, which is more of an issue of the number of baseboards that I am planning to stack vertically (a total of 3 including the proposed N gauge layout) than the fact of having two connected levels in and of itself. I am not quite sure how to resolve that at present - it would be a pity to have to abandon the N gauge project, as it would be a wonderful thing to model railways as I remember them as a nipper as well as something from rather older times, and both I and others on the forum have spent some time in designing and redesigning the layout for that. I wonder whether somehow having the N gauge layout portable or collapsible (but able to be set up ~30cm above this layout when in use) would work? One issue that I have not considered in detail yet is that of baseboard supports: it would not be ideal if access to the lower level were marred by large frontal support struts; it would be really very bad if access to the upper level were also marred by these support struts on account of an even higher N gauge layout. I had not considered this in detail as I was hoping to get the baseboards built professionally and therefore hoped that someone else might be able to solve this problem (rear supports with strong brackets, perhaps, or possibly wall mounting the N gauge layout). I wonder in the circumstances whether I am better off not using the N gauge layout as the basis for practising with things, and whether I am better just buying a little Setrack and practising wiring and automation on the floor (can surface mounted point motors work with layout just resting on the floor?) before the shed is even built, although I am not sure how useful that that would be. Incidentally, as to a stickied post - I can see that the layout building time might be too approximate to sticky, but the results of the locomotive haulage tests are surely useful to lots of people? Edit: In relation to Peco platforms, my approach was to use the prefabricated platforms initially, and, if those prove unsatisfactory, replace them in due course with something more bespoke, but for that work to fall into the category of "a lifetime of slight improvements" rather than work necessary for it to reach a basic stage of completion, which is the important milestone. I had planned something similar with the condensing tanks, to use the 57xx (or, if modelling the Midland, the Fowler 4p) until such time as I am able to acquire the skills to build the appropriate type of condensing tank from a kit/3d print.

Thank you all for your continued feedback: that is most helpful. Junctionmad - I note your reference to servo motors in particular. From my very brief research on the topic, it seems that these are considerably more complex to wire than, say, a Cobolt motor, which has built-in frog polarity switches (albeit those switches still need to be wired in, of course). Can you (or anyone else) estimate how much longer that it takes to wire a servo motor compared to a more ordinary point motor? Also, it would be extremely helpful if you could let me know (approximately) the number of turnouts on the layout in question so that I can try to use that as a basis for calibration rather than linear length of track, as trying two separate extrapolation methods seems to be sensible in the circumstances. Incidentally, we have some extremely useful data in this thread now (the results of testing locomotives' haulage capacity with long trains and corners, and layout building time estimates); I wonder whether it would be helpful to put these in a new topic on their own and ask the moderators to pin it so that people can find these data without having to wade through pages of discussion about my layout plan in particular? As to RR & Co. Train Controller, I am aware of this and that it is well received. I was looking into something similar, the iTrain, which has the great advantage of running on Linux, which is what I use on the small portable desktop computer that I plan to use in the shed, and I am told has features comparable to the RR & Co. software. In relation to height separation, I could probably re-arrange the heights to get a height separation of 35cm between each board with none of the three boards (the two for this layout and the one for the N gauge layout on one wall) at an unreasonable height above the floor, but any greater separation would not be possible within reasonable limits without having only two boards, which would either prevent me from building the/a N gauge layout or requiring having one level only for the OO gauge layout. I wonder how the US layouts manage three boards each with a 60cm separation? That would require a total height of 180cm between upper and lower boards, not including the depth of the boards themselves, which would be very difficult to use for anyone without some sort of raised walkway for the upper level (a kick stool would be an easy thing to get and use for viewing, but it would not be practical to have to use this generally rather than for occasional access to rear fiddle yards). Modular baseboards might need to be considered, but I imagine that it would be very hard to get the track plan for a substantial station throat to a state where there are no joins that do not bisect a point without making each module unwieldy in size.

Denbridge - thank you for the tests - that is most interesting and useful. It is very useful to have empirical data. However, please stop repeating the falsehood that I am not receptive to advice. Exactly what part of planning to test curves with long trains and gradients with short trains before building a layout relying on them because of advice received here amounts to ignoring that very advice? In relation to the time extrapolation, 316 hours' work for 300ft of track would be ~1h/foot, would it not, on average, or ~3h/meter? I note that there was reference to installing point motors/frog switches in cramped conditions as being significant. Can you elaborate a little on what sort of cramped conditions that you mean? (I.e., do you mean that the points are cramped because they are close together, or that the baseboards are cramped and therefore have limited access to them?) In relation to hidden areas, the lower area on this plan is intended to be ~90cm above ground level, so a similar height to a desk (perhaps slightly higher), so the lower level would not be inaccessible on account of its height. The rear of the lower level has no pointwork on the far wall (the top of the diagram), albeit there is some in the fiddle yard further back. Is a layout at ~90cm high with another at ~120cm high really that inaccessible so as to be comparable to an area hidden behind scenery? Edit: Incidentally, as to the traction, the locomotives seem to be quite close to pulling the full amount of carriages, suggesting that ballasting (as recommended by the people who run the McKinley Railway) or a magnetic adhesion system (as produced by DCC Concepts, the "PowerBase") might well be sufficient to overcome these issues, although that would also need testing, of course. Another option would be to reduce the maximum length to 10 carriages.