East Barnet Joint Stock - a workbench

[Compound2632]

5 minutes ago, Chas Levin said:

From the point of view of someone as little versed in LNWR lore as I am, may I ask if there were perhaps different roof shapes at different times, so that the Comet one might be a better match for another type - or sub-type?

 

The question was really directed at Jol! As I understand it, the cove roof profile used between 1903 and 1907 was the same as the profile of the main or lower roof of dining and sleeping saloons built from 1892 onwards, with the proviso that I don't know off-hand whether the width at eves changed between 8' 6" wide and 9' 0" wide stock.

[Chas Levin]

2 minutes ago, Compound2632 said:

 

The question was really directed at Jol! As I understand it, the cove roof profile used between 1903 and 1907 was the same as the profile of the main or lower roof of dining and sleeping saloons built from 1892 onwards, with the proviso that I don't know off-hand whether the width at eves changed between 8' 6" wide and 9' 0" wide stock.

No problem: I shall retire to the sidelines on this question and observe the discussion with interest - and no doubt learn a bit too! 🙂

[MikeTrice]

2 hours ago, Chas Levin said:

Thanks for confirming the tumblehome usage - I didn't know it was also used in naval circles...

 

 

The LNER definitely used the term "Turnunder":

 

[Chas Levin]

10 minutes ago, MikeTrice said:

 

The LNER definitely used the term "Turnunder":

 

Oh no!!! I've gotten used to saying 'tumblehome' and, as a modeller primarily focussed on LNER and its constituents, I'll surely have to change now, won't I?

Actually, that's fine, because I've always thought that 'tumblehome' sounds a little odd, whereas 'turnunder' sounds much more relevant and sensibly descriptive: thanks Mike!

[Jol Wilkinson]

LNWR carriage 4mm kits have, to the best of my knowledge, been supplied with one profile of elliptical roof and one profile of cove roof aluminium extrusions. Some kits may have had vacuum formed roofs but the only one I know of is the Mallard Models WCJS 6 wheel fish van. It is possible that the Mallard/Blacksmith 50' kits also had that type of roof but I have never owned one.

 

I have never researched the accuracy of these roofs. There were no alternatives, so you either built the kit as supplied or didn't bother.

 

I have had a quick look through my stock of yet to be built kits, three of which had  aluminium extrusion cove roofs. All the same profile, but three different "manufacturers".  247 developments (when run by Gary Wells), Brian Badger and Barrie Stephenson. Badger kits became BS kits (now part of Squires) and were part of the same West Midlands "group" that included Ralph Jackson, Modellers World, etc. so barely surprising that they used common parts.

 

When building kits supplied with these extrusions, it wasn't uncommon to find that they weren't always a prefect match the the coach ends. Was the roof wrong, the coach etch design wrong or in the case of those that were supplied with cast w/m ends, was the casting wrong (I usually suspected it was and replaced with etched ends if I had any in the bits box as they were "better" anyway)?

 

The attached rather poor photo  (I am not having much success with the results from my Panasonic Lumix lately) shows the etched and cove roofs from two LNWR kits.

 

.

 

Regarding Stephen's (Compound2632) comment about the cove roof matching the clerestory roof main section, Jack Nelson's LNWR Portrayed shows two clerestory roof drawings, the later being for the 1907 WCJS 12wheel stock. The drawings aren't big enough to see what the differences really are.  They only show the overall height and the height at the eaves. Neither Jenkinson's book or Casserley/Millard's West Coast Joint Stock makes any/much reference to specific differences in roof profiles

 

 

Edited March 13, 2023 by Jol Wilkinson
Typo

[Jon4470]

Hi Chas

 

At one time,  French battleships used an exaggerated tumblehome I.e. a slope inwards from the waterline. The link below shows this

 

https://en.wikipedia.org/wiki/French_battleship_Charlemagne#/media/File:Charlemagne-Marius_Bar-img_3122.jpg

 

That is, as I understand it, the naval use of the word. 
 

I try  to remember to use tumblehome for the upper half of a carriage side and turnunder for the bottom half. Now that Mike has shown the LNER diagram…..I’ll try harder🙂

 

Jon

[Chas Levin]

2 hours ago, Jon4470 said:

Hi Chas

 

At one time,  French battleships used an exaggerated tumblehome I.e. a slope inwards from the waterline. The link below shows this

 

https://en.wikipedia.org/wiki/French_battleship_Charlemagne#/media/File:Charlemagne-Marius_Bar-img_3122.jpg

 

That is, as I understand it, the naval use of the word. 
 

I try  to remember to use tumblehome for the upper half of a carriage side and turnunder for the bottom half. Now that Mike has shown the LNER diagram…..I’ll try harder🙂

 

Jon

Very interesting Jon, thank you. I've noticed that exaggerated shape - which I now know I should call a tumblehome - before on ships, but I wasn't aware that it was a particularly French thing. Well, the French do have an eye for a beautiful outline...

 

I too shall try to use the two terms correctly!

 

And starting as we mean to go on, it seems to me that we tend to be looking at - and trying to model - turnunders more often than tumblehomes; that the majority of coaching stock from the 19th and 20th centuries tended to have straight upper panels.

 

Is that a fair statement, or am I being unobservant? Perhaps I should limit it to LNER and constituents, as I spend more time looking at their stock than that of any other company.

[Jon4470]

39 minutes ago, Chas Levin said:

 

And starting as we mean to go on, it seems to me that we tend to be looking at - and trying to model - turnunders more often than tumblehomes; that the majority of coaching stock from the 19th and 20th centuries tended to have straight upper panels.

 

Is that a fair statement, or am I being unobservant? Perhaps I should limit it to LNER and constituents, as I spend more time looking at their stock than that of any other company.

The turnunder on a Gresley carriage is certainly more pronounced than the tumblehome. The top half of the carriage was slightly angled in from the waist….I don’t have the exact figure with me, but I think it was 1”-2”, The upper side is straight though…the turn under is curved.
Hopefully that a) makes sense and b) is correct 🙂

[Compound2632]

On carriages up to 8' 6" wide, it was usual for the side above the waist to be vertical; on 9' 0" wide carriages, the width at the eves was usually a bit less, so they end up having both turnunder and tumblehome, much like a middle-aged man seen in profile.

[Chas Levin]

10 hours ago, Jon4470 said:

The turnunder on a Gresley carriage is certainly more pronounced than the tumblehome. The top half of the carriage was slightly angled in from the waist….I don’t have the exact figure with me, but I think it was 1”-2”, The upper side is straight though…the turn under is curved.
Hopefully that a) makes sense and b) is correct 🙂

Thanks Jon - I can confirm that it makes perfect sense and to me it sounds correct, in that I think what I was thinking of is the way the upper section usually seems to be straight in appearance, but that of course is different to whether or not the upper section is angled inwards, something I hadn't allowed for in my mental picturing.

[Chas Levin]

10 hours ago, Compound2632 said:

On carriages up to 8' 6" wide, it was usual for the side above the waist to be vertical; on 9' 0" wide carriages, the width at the eves was usually a bit less, so they end up having both turnunder and tumblehome, much like a middle-aged man seen in profile.

Thanks Stephen; that's a very effective analogy. A little too effective, perhaps.

 

(Note to self: stop using the exercise bike as a clothes horse).

 

Presumably the addition of a tumblehome for wider stock was to do with the narrow British loading gauge?

 

On our recent French trip, I was very struck by the profusion of double decker coaches. I then happened to watch a compilation the other day of Classic Railway Newsreels (hosted by one of my favourite railway personalities, John Huntley - memories of Saturday afternoons at the NFT...) which had an item on the brief and not very successful trial of such stock over here, the Southern Region SR Class 4DD EMUs, which were more split-level than true double-decker, because of the loading gauge restrictions. Funny how these things become the norm in a particular region. I assume that a sufficiently large amount of UK infrastructure had been built for smaller, earlier stock to make it completely uneconomical to undertake the work that would be needed to increase the gauge once passenger numbers began to warrant it. I wonder why that situation didn't occur on the Continent? Apart form anything else, they started many operations in the early years (early 19th century) with quite a lot of British built stock - locos, at any rate - so I wonder what prompted them to build to a larger gauge. Is it just a function - perhaps subconsciously - of the awareness of more space: do we, on a relatively small island, husband our resources and automatically build to the minimum specifications, where on the Continent they spread out without concern?

[Compound2632]

11 minutes ago, Chas Levin said:

I wonder why that situation didn't occur on the Continent? 

 

Fewer overbridges?

[Chas Levin]

6 minutes ago, Compound2632 said:

 

Fewer overbridges?

And are you implying that that would be for the more space / more spread out reason, Stephen? Was it as simple as having much more space meaning that roads and rails didn't need to cross over so often? Please excuse me if this is obvious: I find learning about why things have come about endlessly fascinating.

[Compound2632]

8 minutes ago, Chas Levin said:

And are you implying 

 

It's just a guess. In what limited European rail travel I've done, I've been through some very flat areas - northern France, central Germany, western and central Poland. I wonder, very speculatively, if the different legal system, and/or different customs re. rights of way, may also be a factor?

[Chas Levin]

1 minute ago, Compound2632 said:

 

It's just a guess. In what limited European rail travel I've done, I've been through some very flat areas - northern France, central Germany, western and central Poland. I wonder, very speculatively, if the different legal system, and/or different customs re. rights of way, may also be a factor?

Very interesting thoughts; like you, I've done only limited continental rail travel but I'd agree there are a lot of very flat, very open areas.

I know that the French system was built up in a fairly piecemeal way originally, with different independent companies building out from Paris and little crossover: capital to coast, sort of thinking, as I understood it.

Very interesting idea though that it may also be to do with legal / rights of way issues.

 

Opposite me, in the current 'to-read' pile, is a large coffee-table type tome called 'The Golden Age of European Railways' (a holiday puchase recently from a second-hand bookshop in Norwich) so we'll see whether that has much actual history, or whether it's mainly a pretty picture book...

[Chas Levin]

Right, armed with the Isinglass drawing I thought I'd take a closer look at the Sentinel roof profile question and the shape and height of the centre arc. The first thing I wanted to check was how the Isinglass and Yeadon drawings compared - no disrespect intended to either authority, I know both are very highly respected and rightly so, but there's always the possibility of different sources having being used and finding their way into later drawings, or of multiple versions of original drawings and so forth, so with a lot of thought being expended on whether the prototype roof profile matches any or all of the available model options, I thought I'd try some comparisons.

 

I took copies of both Isinglass and Yeadon (the Diagram 97 in my copy of Yeadon came out clearer than the 96 for some reason, but as the only differences between the two diagrams pertain to windows, cabin lengths and door arrangements I couldn't see a problem whichever was used), cropped them, resized one to match the other and superimposed one over the other. I'm sure no-one will be surprised to learn that they're a near perfect match, allowing for slight variance from the thicknesses of the drawn lines at different points.

 

It's not always easy to see the alignment of the roof arcs because they're both drawn in a very similar way; I found it easier to see by varying the opacity of the upper (Yeadon) layer, starting with it quite transparent so that you're mainly looking at the lower, Isinglass one, then increasing the opacity of the Yeadon layer so that it gradually darkens, 'taking over' the roof shape. Hopefully that makes sense - here are three versions, the first with the upper Yeadon drawing at 40% opacity, the second with it at 60% and the third with it at 80%. You may find any one of them easier to take in, or you may find as I did that moving through the three in sequence is the most illuminating. And remember, we're only looking at the line of the roof arc for this test:

 

 

 

 

I realise that other areas of the two drawings don't match up precisely; partly that's because I used the top of the roof vent as a datum line for the upper crop point before noticing that the two drawings show slightly differently sized vents, partly because there are other slight differences between the two drawings, but the important point is that the main roof profile can be made to align very well and that therefore shows that the two drawings agree.

 

I then thought about whether these two drawings in turn agree with the prototype - same remarks as above about my not intending any disrespect to the drawing originators, but errors do sometimes creep in.

 

There, I hit a slight snag, in that straight, end-on views of these railcars are very uncommon - almost everyone took three-quarter views, or occasionally side-on ones. The only true end-on one I could find is unfortunately so low resolution that cropping and enlarging sufficiently to compare with the drawings produced such an uneven roof arc line as to negate the exercise - here it is though (it was already freely online), as it's quite a nice view and even at this resolution it does provide some idea of the roof shape from the front; it shows car 2217 Royal Charlotte at Tweedmouth, some time in the early 1930s:

 

Atmospheric as it may be however, it doesn't fit the bill for comparison with the drawings, so I had to look elsewhere.

The closest to an end-on view I've found so far in sufficiently high resolution is again from the Yeadon Sentinel volume, a photo of 51908 Expedition near Leeds (co-incidentally probably the car I'm going to model), around the same time as the photo above. Here's a fairly tight crop - the original's quite a wide view of the car and it's surroundings - and it's immediately obvious that this isn't an end-on view, but it's closer to end-on than most of the three-quarter views and it therefore gives a fairly good idea of the roof profile:

 

 

I followed the same procedure as I had when comparing the two drawings (done in an old version of Photoshop, by the way) and superimposed a tight crop of the end of this car on top of the two drawings, with a 60% opacity seeming to provide the best compromise between seeing the drawings' roof lines underneath, and seeing the photo's roof line on top, to compare them with.

 

Making allowances for the fact that the Expedition photo is not an end-on view (you can see that the windows are out of alignment with the drawings' ones, and the edges of the roof corners don't align) I think this does still show that the shape of the main, centre arc is a very good match for both the Yeadon and Isinglass drawings:

 

 

Not the most scientific experiment, I admit, but enough to make me feel perfectly confident in basing my roof arc decisions on the available drawings.

If anyone does have a decently high resolution end-on view of a Sentinel that they're happy to share for a repeat of these tests, please would you post it on here - or PM me if you prefer not to post it, in case we can achieve a closer match...🙂

[Chas Levin]

Ok, so having established that the Sentinel railcar roof profiles as shown the Isinglass (and Yeadon) drawings are easily sufficiently accurate to the prototype to use (and once again, I would stress that I do realise that both Isinglass and Yeadon didn't really need checking, I was just being - probably overly - cautious and didn't intend any disrespect), the next question is to what extent the available roof sections match the drawings.

Photographing how the sectional profile of a length of extruded aluminium matvches a drawing is a little awkward and I know that a better way would be to cut off a small section, but I'm a little loath to do so at the moment, in case I end up using these sections to build a long coach, so I've done the best I can with the full lengths for the Comet C10 and the MJT 2970 BR Mk1; for the MJT 2971 Gresley, I already had a short piece, left over from a previous build (a D&S LNER Pigeon Van) and the Nu-Cast white-metal roof is shorter anyway.

 

First up then, the Nu-Cast white-metal roof that's part of the kit, shown below on the Isinglass drawing and I think we'd all agree that it's nowhere near highly domed enough, nor wide enough:

 

 

And here are the other three, with handwritten labels to show which is which:

 

 

 

 

Verdict? All three fit the main arc of the drawing very well indeed and surprisingly similarly, given that they're meant to be different profiles. That shows - I think - that the differences are more to do with the lower sections (the guttering / cantrails are certainly very different between the three, as you can see in my earlier posts comparing the pieces to each other) and that the overall height of the arcs (i.e. the height from the lowest edge underside to the apex of the arc) is where the main differences between the three shapes actually lie, while the shapes of the central areas of the main arcs of all three are actually quite similar.

 

They're also very similar in width and, just like the Nu-Cast casting, they're none of them quite wide enough. It looks to me as if the MJT 2971 Gresley is ever so slightly narrower at the corner curves (a very difficult point to measure accurately and to compare across the three pieces) and that the Comet C10 and the MJT 2970 fill those corner curves just slightly better. 

 

I haven't yet experimented with moving the roof pieces slightly higher, so that although the centre of the arc would be a little high for the prototype (by perhaps 0-5 to 1mm) the slightly wider bases would also move up and occupy those gaps at the sides... 🤔

 

My view at this point is that given that I'd have had to live with a far too low and non-prototypical centre arc height if I'd used the Nu-Cast version (not to mention the considerable weight penalty), it's worth a little compromise over the edge corners to have the main arc right and lose so much weight at the same time.

 

Which one I eventually use - the Comet C10 or the MJT 2970 - will depend on how the tops of the sides line up, because the two roofs have different 'footings', gutter/cantrail sections and so forth and one might be a smoother and easier fit with the tops of the sides than the other.

 

So that's probably about it on roof work, until I have the Worsley Works etches and can begin deciding how to build up the body. There isn't much more I can do in fact until that point, though I'll build the steps once I have some suitable NS or brass strip and I'm still looking at buffer options...

[MikeTrice]

Would a 3D printed roof be an option?

[Chas Levin]

4 minutes ago, MikeTrice said:

Would a 3D printed roof be an option?

Hello Mike, it is certainly an option, thanks for the suggestion; although I have to admit I do prefer working in metal where I can, it might be a way to overcome the combined issues of arc height and width in this case.

 

I think any decision probably has to wait until I have the basic body built, so that I can look at how the various roof options might fit to it. 34016 had width problems when he built recently, using the Nu-Cast body and roof, which suggests that their own roof was actually too narrow for their own kit (though I wondered whether there might have been different batches of castings with slight variation in production - he clearly had far cleaner ones with far better surfaces than mine).

 

I think any 3D printed roof would need to be printed after the body had been made, to fit the precise dimensions of the body as built... On the other hand, if I end up building the body partly from WW etches, I may end up with a fractionally narrower body that might marry up very well with the aluminium roof options.

 

Although I realise it's not the most prototypical approach, I'd be quite happy - if it proves the best option - to make the body ever so slightly narrower, to marry up with an aluminium roof; I think the overall appearance would be worth it, and a body narrower by a couple of mm would look still fine I think.

 

Have you ever done that - knowingly built a coach or other vehicle for practical reasons one or two mm narrower than it should have been, without it being obvious to the eye? (Please excuse me if that seems an impertinent question, no cheekiness intended😉).

[MikeTrice]

The diagrams show the roof as being 9'0" over the cornices so any variation from that dimension would be down to the fusion of parts or quality of the kit. As you say one of the existing metal roofs might actually work out ok. LNER roofs are slightly smaller over cornices than the quoted 9'0" so it might work if additional strips were added to the cornice to make them thicker.

 

To curve the ends saw a series of slots along the length of the roof to match the length over which to roof curves down and bend the aluminium to suit. Back fill with car filler.

 

To do a 3D printed roof I would suggest the body is built then placed top down on a scanner to allow an accurate profile to be matched to both length and width and end curves.

[MikeTrice]

Regarding sawing a series of slots you form the roof like the Kemilway products illustrated here by @Rob Pulham in his Flickr collection:

 

[Chas Levin]

27 minutes ago, MikeTrice said:

The diagrams show the roof as being 9'0" over the cornices so any variation from that dimension would be down to the fusion of parts or quality of the kit. As you say one of the existing metal roofs might actually work out ok. LNER roofs are slightly smaller over cornices than the quoted 9'0" so it might work if additional strips were added to the cornice to make them thicker.

 

Yep, agreed - I suspect that some discreet cornice enhancement might be needed...

 

28 minutes ago, MikeTrice said:

To curve the ends saw a series of slots along the length of the roof to match the length over which to roof curves down and bend the aluminium to suit. Back fill with car filler.

 

Excellent suggestion, thanks Mike - I've seen that done but not with aluminium and I hadn't thought of that: I'd been planning on splicing the Nu-Cast white-metal ends onto the aluminium length and ad been starting to think about how best to secure the (difficult) join, thinking at the same time that it would partly negate the advantage of the aluminium's smoothness.

 

31 minutes ago, MikeTrice said:

To do a 3D printed roof I would suggest the body is built then placed top down on a scanner to allow an accurate profile to be matched to both length and width and end curves.

 

That's a very good idea too: I don't think I'll go down the 3D printed route on this vehicle but if not for this one, that's a really good way to plot a customised shape - I hadn't thought of using a scanner! That might be useful too for judging the fit of the aluminium ones, just to compare scans of the top of the body and the underside of the roofs...

[drmditch]

Re: Tumblehome

This is a very old term in naval architecture. The inward slope of a ships sides above the waterline has to do with:-

     Ships (and especially sailing ships) being hardly ever vertical in the water. (1)

     Reducing topweight and keeping the Centre of Gravity low. (2)

     Allowing the shrouds (which are the ropes which provide lateral support to the masts) to run in a straight line to the tops (at least of the lower masts).

     Shrouds are secured to the hull, and in the 18th and earlier 19th centuries, were offset outwards by platforms known as 'channels'.

     The stress patterns of a sailing ship are carried not just by the hull, but by the hull masts and rigging together. (3)

     Guns (should one be carrying any!) on the upper decks will still be in that plane of the structure directly supported by the water even at a 10 degree angle of heel. (4)

     The 'frames' (acting like ribs) of the ship  can form a curved structure, which is inherently stronger than a straight one. (5)

 

I would suggest that the last one of these factors is also relevant to railway carriages, and such carriages were, presumably the earliest landbound large moving structures.

Perhaps it is significant that earliest (and small) carriages tended to have slab sides, and as they grew larger they tended to have 'tunblehome' and 'turnunder'

 

Readers of the Jack Aubrey books will be aware of his dislike for the 'slab sided' (ie no tumblehome) HMS Worcester!

 

Notes:

1. I think railway vehicles will not rock from side to side to the same extent.

2. Might also be relevant to railway vehicles.

3. Apart from the famous examples on the Spurn Point railway, few railway vehicles required complex masts, and sails.

4. As far as I know, no railway carriage ever carried a broadside armament!

5. And can be formed from several pieces of timber clamped and glued together.

 

Sorry for the distraction. I will now return to ensuring that the sides of the Covered Van I am constructing are definitely flat!

 

 

Edited March 16, 2023 by drmditch

[Chas Levin]

33 minutes ago, MikeTrice said:

Regarding sawing a series of slots you form the roof like the Kemilway products illustrated here by @Rob Pulham in his Flickr collection:

 

Nice job - I hadn't seen this particular one but I've seen others.

 

Isn't the extruded aluminium quite thick to do this with though? Have you ever done it with aluminium Mike?

[Compound2632]

The meaning of tumblehome in naval architecture is all very well but to my mind, the question is, what terminology was used by the builders of road carriages of the 1820s? It was one such, Thomas Clarke Worsdell, who designed (over tea with George Stephenson) and built the first proper railway carriages, for the L&M.