The last few weeks have been considerably busier than usual with non-railway things, so there’s been less time and less energy for modelling, but some useful research time.
In planning out the side panel lining, I realised that one critical dimension isn't specified on the painting diagrams: the radius of the incurved corners. Not only is it a necessary dimension for the corners themselves, but it also dictates the lengths of the straights.
My first thought was that it wasn't specified because it’s obtainable from the given information and I spent some time in ultimately fruitless but interesting attempts to press a few tattered remnants of schoolboy algebra and geometry into protesting use, before concluding that it can’t be done. A key realisation was that you could reduce or extend the lengths of the straights whilst still maintaining their distances from the panel edges, with predictable consequences on the corner radii.
A quick look online brought relief in the knowledge that I’m not alone; an LNER Info Forum thread (https://www.lner.info/forums/viewtopic.php?f=11&t=10794) involving an enormous collective amount of modelling experience concludes that while this measurement can be found for later styles - BR and perhaps LNER - it hasn't been recorded for earlier pre-grouping GNR. Bill Bedford made the interesting suggestion there that the corner radii aren’t included on diagrams because the paint-shops had templates for them, which sounds very plausible to me.
So: what to do? Having been so excited at having authentic, original measurements from diagrams – all of which remains useful and valid - it seemed like a retrograde step to start measuring photos, but I couldn't think of another way. I entirely realise that this is probably somewhat over the top for the job in hand, but once I got started, it became something of a quest to try and determine exact measurements, even though I realised I'd probably end up measuring more accurately than I'll be able to paint.
The way to do it that occurred to me was to re-size a suitable photo at 4mm:1ft and then mark points to measure on that. I expect there are other ways to do it – please do suggest any that I should try too – but this seemed a good starting point - apologies for the long post...
Scaling & measuring photographs
Starting with as straight a side-on view of a square-tank C2 as I could find, I thought it would be easier to be sure about dimensions if I could mark them accurately in some minutely adjustable way, so I used Microsoft Word, which lets you add lines, shapes, arrows and text-boxes, change their colour, alter their positions incrementally and so forth. It’ll also result in a lining diagram with 4mm scale measurements which might be useful! I'm sure other software would do the job more efficiently but Word is one I already know and have.
In studying and measuring this image - and in comparing it to other similar ones - I came up against one or two peculiarities of optics that required some thought. I found that all the photos exhibited some small variations in dimensions between one end of the loco and the other: tiny, but perfectly perceptible. The height from footplate to bunker top might be half a mm greater than that from footplate to side tank top, for instance. In some cases, this may just be distortion of the original photo, or distortion that's crept in somewhere in the chain between the original and its final appearance in print. I worked from high resolution scans but still couldn't quite eliminate it. At 4mm to the foot, half a mm is 1 1/2” – enough to cause problems!
The other cause - I think - is to do with the width of the viewpoint and it explains why measuring from scale drawings is accurate in a way that doing so from photos usually isn't. When a scale drawing of something long such as a loco is produced, it's done from measurements, directly on to paper and is drawn as if the viewpoint is the same width (or length) as the loco - imagine if your eyes, or your camera aperture - were like a long letter-box, the full length of the loco itself. A photographer however is viewing a long object through the camera lens - a pinpoint, relatively speaking. That means that the distance from the camera aperture to different points on the loco varies - think of the middle compared to either end, which would be the greatest difference.
Another way to visualise the problem is to imagine a loco photographed through a fish-eye lens. That would exaggerate the problem grossly of course, but even the professional photographers who took these types of official portraits, using the best type of camera for such a subject (not my speciality subject so I don't know what they'd be called - full plate?) seem to have retained a very small amount of this perspective distortion, if that's the right term.
I decided the best way to minimise the problem for measurement purposes would be to use as narrow an area of the loco body as I could, so there'd be as little of such perspective distortion as possible. Looking at various photos, it appears to me (and I stress this is only a theory or an impression) that the photographers positioned themselves just slightly forward of the cab doors rather than dead centre of the footplate. The reason I came to that conclusion is because on several photos that were clearly intended to be straight side-on portraits, working forward from the back, measurements - such as the footplate to tank-top one I mentioned earlier - seem to agree from the rear end until somewhere around the middle of the side-tanks, where they then start to reduce. In a couple of cases, there even appears to be a visible curving distortion – like that from a wide-angle lens - of the front end of the loco, as if it's curving away from the plane of the photograph.
Therefore, I decided to focus on the bunker side as my main measuring area. It's a fairly short section - relative to the whole loco length - and should therefore be reasonably uniformly reproduced in the photos.
After making a reasonably high-res scan of as straight a side-on view as I could find in decent quality, the next job was to re-size it to print out at 4mm to the foot. Using the measurement from the footplate top surface to the underside of the bunker beading as my datum point (it's prominent, clear, uncluttered by other fittings and fairly easy to measure by going in from behind the rear of the loco with the calipers, plus the change in background behind the loco aids accurate placing of the caliper jaws) the first thing I found was that this dimension is 17mm on the Isinglass drawing but 17.4 on my LRM model. I think that's down to my inaccurate construction, but I've used 17.4 (or 52 1/3” on the prototype) to scale the photo, as the point here it to produce lining measurements for use on this particular model.
I then added in 'marker lines', in various colours (in order to aid distinguishing between them where things get a bit crowded). These were positioned by zooming in on the photo to maximum size and, for final positioning, altering the stored position numerically for each line or, in the case of the circles used to match the incurved corners' diameters, altering their sizes numerically: you can go into the properties for an individual line once it's been created and alter its vertical or horizontal position by steps of 0.01mm. Almost everything can be marked by lines added to the bunker side, except of course the horizontal lengths of the side-tank lining.
Here's what I eventually came up with - I hope this is OK from a copyright point of view, as the original image is incorporated in another piece of own. Please note the image below is a cropped, low-resolution screenshot of the document for viewing purposes and is not to scale:
(All versions edited September 20th, to correct a labelling typo and to change the colour of the 'A-F' letters from white to red, so they show up on the 'loco-less' version).
Here's the same thing in a high resolution 4mm:1ft scale-print size, in pdf format, in case anyone wants to take a closer look - if you download this and print this one at actual size - no scaling - it will come out at 4mm scale:
GNR C2 1013 scale-print lining CL 20210920 (1).pdf
Another issue I came up against is the tendency of the line-drawing in Word to exhibit slight off-vertical or off-horizontal tilts which can't be corrected by typing into a ‘Properties’ box, as can the size and overall position. This can mean that the distance between pairs of coloured Word lines may vary, depending where along their length the measurements taken. My solution to this was to take all measurements from the point on the Word lines where they actually sit above what they're marking in the photo, positioning that was done as accurately as possible, under extreme zoom-in: for instance, the blue lines that mark the outer white lining on the loco were measured at the point where they actually intersect with the white lining on the photo.
I found this quite difficult to do at first because the very thin (¼ point) coloured lines are hard to see with the loco photo behind them... until it occurred to me that having got the coloured lines accurately positioned and saved, I could simply save another copy of the document with the original photo of the loco deleted, leaving the coloured lines much more easy to measure against a plain white background - again, please note this is a cropped, low-resolution screenshot, not to scale:
And as above, here's the same thing as a high-resolution pdf, which will print out - at Actual Size - at 4mm to the foot:
GNR C2 1013 scale-print lining CL 20210920 No Loco (1).pdf
Looking at the loco-less lines on paper, it also occurred to me to try putting in a suitably re-sized scan of another square tank C2 under my measuring lines to see whether they’d line up reasonably well. This didn't go quite as smoothly as I’d hoped, perhaps because - I suspect - the perspective distortion I mentioned above is either more pronounced in this photo, or because it varies between photos and has affected this one differently. Also, the only other suitable C2 photo (of 1009) I've found so far is more blurry than the one of 1013. But, allowing for some variation between the two, things lined up well enough to confirm the crucial measurements of curve radii and lining lengths. I was able to line up all the measurements on the all-important bunker side, even though further forward along the loco things were a little adrift, which again makes me think the perspective of that 1009 photo is slightly different to that of 1013.
At last: the elusive incurved corner radius measured!
So, some actual figures: I measure the GNR C2 lining’s outer corner radius as 1.25mm (3.75” on the prototype) and the inner corner radius as 1.55mm (4.65” on the prototype).
This makes the inner radius 24% larger than the outer one. (If I've got my calculations wrong, please don't hesitate to reply and let me know!). This is quite visible in the original photo, where the inner curves are clearly significantly broader - the upper right-hand side tank inner corner curve radius actually looks nearly flat, but I'm guessing that's some sort of photographic distortion and that it would have matched the others in reality.
On the LNER Info Forum thread mentioned above, there are several calculations and suggestions for these radii. 'John Coffin' suggests an inner radius 7” and an outer of 5” (a 40% increase, based on his study of a Doncaster 2-2-2 drawing and his considerable experience in GNR lining). Allowing for the change from 2” to 2 1/8” for the overall white-black-white lining (where I think we’ve referred to differing sources) this assumes that the two radii match in curvature, so that one can be calculated from the other by adding – or subtracting – the total lining width. That certainly looks to be the case on many LNER loco photos, where the inner and outer out-curved corners appear to match in curvature. As mentioned in my previous paragraph however, the C2 photos I've been studying show the outer curves as very clearly tighter than the inner ones. This is very apparent on 1013: the inner curves look to me to be like small sections taken from a far larger radius circle. I wonder if this might be another area of variation between different locos or paint shops, perhaps more varied in GNR days, being more standardised after Grouping?
'61962' refers in the thread to another painting diagram in Leech & Boddy's “The Stirling Singles” (a book I hadn't come across before, but which sounds interesting). Like other diagrams referred to, it doesn't give radii either, but he calculates them - by scaling from the measurements given - as being 3 1/2” inside, 5 5/8” outside. Unless that's a typo - reversing inner and outer? - this suggests to me that there were changes in the lining style and that perhaps this diagram was from an earlier style, where the inner curve was the tighter of the two.
I took a look through the Stirling volume of RCTS; to my eye, the Singles do seem to have incurved corners of a very similar looking radius, which seemed to support the theory of different earlier styles... until I carried on through the volume! The next locos, 0-4-2 Well Tanks, have a lined panel on the bunker with the same feature as the C2, where the inner curve appears to come from a far larger radius circle. The same is true of the long side-tank panels on the 0-4-4T G3s and G4s, F2s and F7s and the 7'6” 2-2-2s. Were the Singles a special case? Well, whatever the case for them - and for their painting diagrams - I believe it's the case that in general, the inner curves of GNR incurved corners were a wider radius than the outer curves.
Outer lining to panel edge distances
I measure the GNR C2 vertical spacing of the outer white lining, measuring in from the tank front and from the bunker end (and also from each side of the door), as 2mm, or 6” on the prototype and all the square tank C2 photos I can find appear to show these vertical spacings being consistently the same.
This doesn’t agree with the painting diagrams I have which give 8” for these distances, but that’s on the side of a tender - a different loco - and we’ve already noted that these diagrams may show earlier and slightly different lining styles.
Whilst not directly relevant to this square tank C2 build, I learnt from studying various photos that the round tank locos are not at all consistent in this measurement. Measuring from straight side-on photos of 1504, 1509, 1529 and 1535, while the spacing from the outer white lining to each side of the door is the same (it would look odd if it weren't!), the spacing of the outer white line from the tank front is generally 10-12% less than that from the door, while the spacing from the rear bunker edge is anywhere between 30-50% narrower. I thought this variation might be due to lining being applied at different paint shops, but according to the RCTS book all 60 of the class were built at Doncaster, so it must simply be that such dimensions were not fully standardised at this date.
One spacing that is fully consistent across both square and round tank C2 versions in photographs is the horizontal alignment of the lining panels: I measure the spacing between the outer white lining and the upper and lower panel edges at 1mm, or 3” on the prototype, for both side tank and bunker lining.
This is at odds with the painting diagrams I have, which give 4 1/2” from the lower white down to the footplate but 3 1/8” upwards from the upper white line to the panel edge. However, these figures are for the lining on a what I believe is an earlier Stirling tender, which has additional lining bands between the outer dark green area and the start of the tender flare, so it's effectively a different lining design. Having checked this measurement on six good side-on photos of C2s (two square, four round) and found it consistent, I think I'm safe in taking 1mm / 3” as being correct.
Having now established the missing dimensions, I think I'll also do some proper practice - and compare the Phoenix Precision and Humbrol gloss whites for lining use - by going back to the GNR green test panel, which I used earlier to practice the red-black lining - the abutting white-black-white lining required is a step up from the black-red on the chassis. There's space for the two panels - side-tank and bunker - in scale, on the piece of test brass, a happy accident!