JimC

Do you think, though, the other rivet lines represent tank boundaries or simply fabrication of the bunker assembly? I've taken the viewpoint that only the very closely spaced lines of rivers represent a tank edge, and others are boundaries of plates and/or transverse structures. With my 'no rivets' policy those are best left off. I'm already concerned that the sketches are misleading as there are lines on some original drawings that suggest that what looks superficially like a flat topped tank is actually anything but, and on the inside the tank top slopes down to the floor in the centre. Anyone here familiar with the preserved 5101s and 61xx and can comment?

I'm not clear on what the difference between 3111-3130 and 3131-3150 is.

The 'Victory' class was a class of ten built in 1917 for the Inland Waterways and Docks dept. Post war they were sold off by the Railways Operating Department, mostly to collieries. There's a detailed history here at Planet industrials. The ADR bought two of these from the ROD. They had outside cylinders and were quite powerful locomotives. They were numbered 666/7 on absorption. They received a moderate Swindon rebuild. Another had been purchased by the Brecon & Merthyr. This loco was numbered 2161. It was given a significant overhaul in 1922. The B&M loco was sold in 1929, and lasted to 1951 in colliery service. Both the ADR locos reached British Railways. This sketch of the beast is intended to portray the later GWR configuration on at least some of the class with GWR dome and safety valve cover. They seem to have had new tanks in GWR days with prominent riveting, but I don't do rivets in my sketches. The drawing owes a lot to Planet Industrials and in particular the Don Townsley drawing on the web page for their upcoming model. However the beast is completely redrawn, and, for instance, I've steered something of an intermediate course between the GWR weight diagram B13 and the Townsley drawing on some aspects, notably cab window position.

'cos obviously British Railways polish is utterly different to modern polish. It's daft, but that' s collecting for you.

Excellent. So it's a reasonable supposition that they were all given that layout, which matches the conversions.

Photo extract of a page from Russell - think this distorted image counts as fair use - showing the frankly horrendous routing of pipework for the water collection on the 3150s. The pipes have to exit above the water level in the tanks otherwise they double as drains! So the upward extension of the back tank gets the water level up to match the side tanks. Looks desperate to maintain to me, not the drawing office' finest moment and although I'm not saying I could do any better it's easy to understand how in practice they found it not worth the bother and ripped it out.

Thanks for that. The vertical section at the back of the 3150 tank seems originally to have been associated with water pickup gear, and only given the slope at the top when the water pickup gear was removed. I've found a photo of a later 3150 that wasn't built with water pickup gear which still has the vertical section, but I haven't het found a clear photo of one between 3150 and 3165 as built. Its interesting to look at the County Tank photos in Russell, they had at least three styles of tank!

Some notes, 5100, ex 3100 née 99 seems to have had the small sloping water tank all her life. Its visible in photographs of 99, when she had a vertical back on the bunker. The 3111s all seem to have had the higher flat topped (at least at the sides) tank from new which was unchanged when the bunker extension was added. The 3150s that were built with water pickup apparatus (3166-70) had the vertical section at the back of the tank as shown which accommodated piping for the water pickup apparatus. After the gear was removed and the bunker extended these tanks were altered with a sloping top section of the style shown in the other sketch. At least some, if not all of the 3150s built without water pickup apparatus, both before and after 3166-70, also had the upward extension of the tank. The tank style on diagram A13 for the Collett 3100 doesn't seem to have been constructed on any of the class. Another demonstration of why its wise to be wary of weight diagrams. Photographs show the 3100s with the style shown in the other graphic, presumably unchanged from their previous incarnation as 3150s.

I've sketched out the principal variations I can see in Russell on this blog entry.

I've sketched out the obvious principal variations in Large Prairie bunkers visible in Russell, both drawings and photos. I'm making the assumption that the lines of close spaced rivets on the bunker side did indeed follow the top seam of the water tank. It sees to me there must have been variations in coal and water capacity. I shall have to take a closer look at the RCTS volume.

Yes, I had a fair bit of agonising over bunkers when I did my series of drawings. I'm not sure that there weren't three depths over all the large prairies, but I don't seem to have recorded my thinking. Maybe it was a last vestige of the bunker free for all earlier in the century? And later - I'm also seeing variations in the rear tank as witnessed from the rivet line. If you have Russell handy compare fig 512, 515, 518, 523 and 546. Note also that the rivet line on the photos of 8102 and 8103 is quite different from the tank profile sketched in fig 549.

I'm sure Mr Pareto's law comes into play here. 90% of accidents were/are pretty easy to diagnose, but the other 10% are a problem. And a popular but less scientifically defined law tells you that if you don't get every last scrap of evidence it will turn out to be one of the 10%

There were at least two balancing schemes on the 56s weren't there? RCTS says that cranks were altered to the "Stroudley" pattern with crank and coupling rods in line to reduce wear on the driving axleboxes. What effect would this have had on the fore and aft motion felt by the train? I suppose, thinking about it, all outside cylinder locomotives are balanced on this pattern. I can't get a picture in my head of what difference this would make to eventual loads.

I recall a schoolmate's father had a very large (treble garage sized) 00 layout which was outside rail pickup, so it existed. Certainly looked a lot better than Hornby 3 rail track!

My understanding is there was one specific bridge or viaduct with a weight limit, and it was later upgraded.

As an exercise I've just been attempting to draw relationships of wheels with track with different wheel arrangements. It speedily became apparent to me that this is a far more complicated subject than I had ever realised, and that a reasonable study would involve some very serious mathematics, well beyond my competence, and large tables of data to present the results, which would be rather indigestible in any format other than textbook or technical paper. However my little study served to educate me in the very beginnings of the subject, so I thought I would pass the results on. The attached sketch should blow up large enough to be able to make some sense of it. Here are some notes, most of which will probably be more than obvious to any professional railwayman, or for that matter trained engineer, and most likely most chassis building modellers too. For simplicities sake I have limited this to three arrangements. Green is a fixed wheelbase 0-4-2 (or 2-4-0 or 0-6-0). Red is an 0-4-2 with a radial trailing axle, and the wheels spaced evenly. Blue is an 0-4-2 with the driving wheels closely spaced in the manner of the 3521s and others above, and still a radial axle. The 4th sketch is all three superimposed. You'll need to enlarge the thumbnail and use a big display to get much from this I fear. My first realisation was just how bad the 0-6-0 arrangement is for spreading the track and wearing rails and flanges. In any given fixed wheelbase an odd number of wheels is the worst case. I now understand better than before flangeless centre driving wheels on ten coupled locomotives. On the other hand its the best for lateral stability and minimising throwover on curves. My second surprise was that shortening the wheelbase on the driving wheels has a considerable effect, more than I would have thought. I think Dean must have been seeking to reduce flange and track wear on the 3521s and their cousins. But the price that must come with it is lateral instability and increased throwover on curves. I also started thinking about how side control to return the radial axle to the centre must have a significant effect on the wear on track and wheels, especially leading wheels. My understanding is that sideways control, and how best to achieve it, was an engineering challenge well into the 20th century, so its unsurprising that Deans design team struggled with what must have been a new topic. Perhaps the 3521s needed a deGlehn style bogie? Another interesting thought is how uneven wheel spacing may have advantages. My little thought experiment started with an 0-6-0 design with all 3'6in wheels evenly spaced on a 16 foot wheelbase - 8ft + 8ft, and 12 tons weight on each wheel. I then moved the leading pair of wheels from eight feet apart to four feet apart, shortening the overall wheelbase to 14 feet, so the wheel base is 4ft +10ft. There should still be 12 tons weight on each driving wheel, but not only is the locomotive kinder on the track because the wheelbase is shorter, there's also less of a spreading effect on the track as the centre drivers are not on the apex of the curve. The complications and tradeoffs of different arrangements seem almost endless.

It's interesting how the Barry rang the changes with various chassis and tanks around a basically standard boiler and I think cylinders, 0-6-0T, 0-6-2T, 0-6-0ST, 2-4-2T, 0-4-4T, but if only we had a record of why the CME decided on each config and of how they worked out.

I think it will be reasonable to distinguish the long wheelbase 0-42T and 0-4-4T from the others. 3521, 3541, 1345 and 34/35 all seem to be long overall wheelbase/short driving wheelbase locomotives. I'll speculate that flexibility to handle curved track may well have been part of the design aim. The drawback seems to have been that at least some of the designs were terrible riders, with a track record of derailments. No 9, although in many ways more related to tender locomotive development - see @MikeOxonblog for an interesting side by side of a No 9 model and a Dean single, seems to take the flexibility and also the poor riding a stage further. The original form of No 1 also seems to be about flexibility. The Stellas and the rebuilt No 1 seem of a type chassis wise, and, in spite of quite different frame design, perhaps with much in common with the Joseph Armstrong 2-4-0T. I need to think further about the others. Got to beware though, its perilously easy to construct great card houses of speculation about designs, only for them to fall when further facts become available.

Most likely the outline was readily visible as a guide. For a modern example look at photos of 5322s tender as currently painted black - the outline of ROD 5322 is clearly visible.

They were part of a whole group of tender and tank engines that are very similar, with 17*26 cylinders. Looking at the build date they were surely designed as a group. As a reasonably conventional class they're not so much what I was thinking of. 3201 2-4-0, Lot 65. 3201 was built before the others, in Dec 1884, and immediately sold to the Pembroke and Tenby Railway, returned1896. 3501-10 2-4-0T Broad Gauge, Lot 64 built 3/1885 - 5/1885. Converted to Std gauge and 2-4-0 1892 3511-20 2-4-0T, Lot 64, built 5/1885- 9/1885. Altered to 2-4-0 1894/5 3202-3205 2-4-0, Lot 65 built 6/1885 - 8/1885 2361-2380 0-6-0. Lot 67, built 9/1885 - 5/1886 1661-1680 0-6-0ST, Lot 69, built 6/1886-11/1886 1661 1680-1700,0-6-0ST, Lot 71, built 11/1886 - 6/1887

This is the full on Swindon version of the F class, with the Barry boiler replaced by a Standard 9, and of course pannier tanks. This sketch is based on GWR diagram B60, and its interesting that the inside frame profile on that drawing is different from the other GWR and Barry diagrams. Its difficult to work out this kind of detail in photographs where the frames are in shadows, but I have a suspicion this style of frame may be associated with at least some of the later lots. I do need to say that although I've given the sketch a cabside number of 780, I do not actually know which style of frames 780 was fitted with.

This was an interesting one to draw too. The nicely drawn Barry weight diagram is dimensioned with a front overhang of 6ft 7in, but the drawing scales some 5 inches less! I've gone with the written dimension, which is the safer option with workshop drawings.

I'm trying to work up a sketch format that I can do reasonably quickly and which doesn't imply a level of accuracy that doesn't exist, yet is sufficient to be usable for comparison purposes. How does this seem? This is the 1865 Sturrock built by Avonside for the GNR. The first of Deans experimental types No 1 started life with a very unconventional front bogie, which was a unique and rather complicated suspension type. Later it was converted into a conventional 2-4-0T as per @Miss Prismpost above. This is a representation of the notorious yet elusive 4-2-4T No 9 which is reputed to have never left the factory. Only the sketchiest of information survives on this locomotive, and even that is debatable. LA Summers and Mike Flemming make a convincing argument that the recorded bogie dimensions are transposed front to rear, and I have followed their lead. The tanks are reported to have stretched beyond the smokebox, but all else is guesswork. No 13 in its original form as a 2-4-2 with a back tank and a well tank. And this is (roughly from a photo) one of the first few 3521s. This is 34/35 as built, with a saddle tank. Here's 34/35 in their 0-4-4 days with a back tank. And this is the 1345 class, the ex Monmouthshire 0-6-0ST modified to 0-4-4.

The first thing that intrigued me was the 1345 rebuilds. Why rebuild 0-6-0ST as 0-4-4s? Looking at RCTS they are noted as having been unsteady at speed, rather like the 3521s. The first of them were built by Avonside for the Monmouthshire, a line that seemed to have had a very odd collection of locomotives. That got me looking at what else came from Monmouth, and I noted three similar Avonside 0-4-4Ts with back tanks, rebuilt by Swindon with saddle tanks 1897/9. That in turn made me look for Avonside, and I turned up this page: https://www.gnrsociety.com/locomotive-class/sturrock-0-4-2-suburban-tanks/ Doesn't the 241 series carry a bit of a look of the 3521s? Sturrock had been principal assistant to Gooch on the GWR, and the 241s were put together to work the Metropolitan lines when the GWR had a disagreement with the Metropolitan. It seems very unlikely that Dean would have been directly influenced by an 1865 design in 1887 though. I shall try and figure out a style for basic sketches of some of these types. I just don't have the information to do my normal style of sketch without them appearing to be more accurate than they really are.

This is something of a followup from discussion in another Blog entry, https://www.rmweb.co.uk/community/index.php?/blogs/entry/24891-gwr-no-34-1890/ and is also relevant to this one. https://www.rmweb.co.uk/community/index.php?/blogs/entry/24922-gwr-3521-0-4-2t-and-0-4-4t/. As I said, I'm beginning to further appreciate what a weird and largely unsuccessful bunch Dean's larger tank engines were, and what a contrast in style they were from the smaller 6 wheeled engines, conventional, successful and very long lived, and heavily based on Armstrong originals. So this is a sort of brain dump/request for comments. What struck me in the previous thread is that there does seem to be something of a common style to the larger Dean tanks, quite separate to the Armstrong derived Metros and Metro derivatives, very conventional and with six or eight reasonably evenly spaced wheels. It seems to me that I can categorise the larger ones as: 1880, No 1, 4-4-0T. Rebuilt 1882 as a more conventional 2-4-0 and in that form survived until 1924. 1881, No 9, 4-2-4T. Never went into service, and has attracted a good amount of writing and speculation. 1885, 3501 Class 2-4-0T. Broad gauge convertible versions of the Stella Tank. 1885, 3511 Class 2-4-0T. 'Stella tank'. A tank engine version of the Stella 2-4-0 and part of a group of closely related locomotives that also included the 1661 Class 0-6-0ST and 2361 class 0-6-0. 1886, No 13, 2-4-2WT. Rebuilt 1897 as a 4-4-0ST and in that form survived until1926. 1887, 3521 Class, 0-4-2T. Converted to 0-4-4T 1891/2 and rebuilt as 4-4-0 tender engines from 1899. 1888, 3541 Class, 0-4-2ST. Broad Gauge half sisters of the 3521s, converted to 0-4-4 1890/1, narrow gauge 1891/2 and rebuilt as 4-4-0 tender engines from 1899. 1890, 34 Class, 0-4-2ST, converted to 0-4-4ST in 1895 and as such ran until 1906/8. 1891, 1345 Class. 0-4-4S/WT. Rebuilt from ex Monmouthshire Railway 0-6-0ST absorbed in 1880. Withdrawn between 1908 and 1913. 1898, No 1490. 4-4-0PT sold 1907. I don't think I'll consider the Stella family much in this exercise, it seems to me that they are a separate line of development. What I'm particularly interested in for this is the Swindon built 0-4-2/0-4-4T locomotives which seem to me to embody a common style.