'Printing' a Gloucester wagon

[Andy Vincent]

Making a Start

 

Just to be clear: this thread is about modelling Gloucester wagons. The scratchbuilding tools in this case are (mostly) a CAD package and a resin printer. Whilst tips and techniques might get an occasional look-in, this thread is intended to be much more a discussion about the variations and challenges of modelling the wagon output of the Gloucester Railway Carriage & Wagon Company. These wagons represent something of an enigma: they were recorded through an extraordinary set of works photographs, yet no designs are believed to have survived. 

 

Given the kit ranges that have been lost over the years - and the wagons they enabled us to model - it is a reasonable question to ask whether my time might have been better spent on creating something else. After all, Slaters and Cambrian provide several variants of Gloucester wagons in their ranges. Indeed my modelling interests revolve around the Southern before and after nationalization. Yet there remains the challenge that the Gloucester wagons represent and of whether 3D, and particularly resin, printing can be used to capture more detail and address some of the shortcomings introduced by the injection moulding process.

 

However, successfully producing a 'Gloucester wagon' using these techniques opens up the possibilities of producing far more variants of the 'Gloucester wagon' than ever before. Also, unlike models produced by injection moulding with its high tooling cost, having the 3D design files enables corrections to be made as new information about these wagons comes to light. Ultimately, though, it is the inspirational discussions on Stephen Lea's D299 thread which set me off down this path. And so this thread and whatever value it might have should, at best, be considered as an odd corner lurking in the shadow of the discussions on that mighty thread.

Next time: where to start, and does the RCH 1887 spec give us any clues?

 

[Edited to fix the English!]

Edited April 2, 2022 by Andy Vincent

[MJI]

What cad will you be using?

[Andy Vincent]

9 hours ago, MJI said:

What cad will you be using?

Fusion 360. 

 

I can't claim that my choice of CAD software is based on any sort of rigorous investigation and comparison - it was simply the package recommended by Alan Buttler (of Modelu fame) when he ran a CAD course that I attended at a Missenden Abbey modellers weekend. 

 

What I would say to anyone contemplating CAD is that whilst Fusion 360 may see more complex to learn than some alternatives, in practice you can start with just a few basic operations and for many projects that will be all you need. There is also a wide range of bite-sized training videos on YouTube.

 

However, to me, the capability that sets it apart is that it uses 'parametric modelling'. This records each operation as a separate step where you can move backwards and forwards through the 'timeline' making changes as needed to earlier steps. That ability to go back and change what you did earlier without losing what you did later has a massive benefit.

[Andy Vincent]

Of the RCH, the GRC&W - and Archie Croome

 

The Railway Clearing House got into the 'wagon specification business' after multiple accidents involving private owner wagons. In 1885, following a series of accident reports, it was decided that the vetting of private owner wagon designs was needed. In 1887 'The Railway Clearing House Standard Specifications for 8 and 10 ton Private Owner Wagons' was published. As @jim.snowdon (who knows much more about this than me) has repeatedly stressed to me and others, the RCH did not specify the design of a wagon, rather they provided a set of drawings for components and some principles (timber sizes etc.) which, if adopted, would be likely to speed approval of a wagon design. 

 

Th 1887 specification set in motion a whole raft of developments that shaped the design of private owner wagons right through to the second world war. Aside from an easier approval process, this standard was to lead to operational benefits - if wagons used standard components then they could be repaired more easily since parts did not have to be sent from the original builder. 

 

We now come to the mystery of what component drawings exist in the 1887 specification that we can use as modelers. I know of twp sets of drawings: the first is included in Tony Watt's 'Ince Waggon & Ironworks' book for the HMRS. These, and the specification, are credited to 'Mr T.G. Clayton' who was Chief Engineer of the Midland Railway and Chairman of the Wagon Superintendents Committee. Slightly oddly, these drawings include example registration and reconstruction plates both dated 1903.

 

It is known that there were updated versions of the RCH specification in both 1903 and 1904 (plus others in later years). The latter is available in the reprint of The Railway Engineer fro December 1905 but includes only general arrangements but with the promise of including the detail drawings in a later edition.

 

The second source of drawings comes from a series of articles published in the periodical 'Work' in 1898 by Archie Croome. This series was ostensibly about providing the information to enable the reader to 'construct a wagon to any scale'. The series was published in ten parts from July to November 1898. The articles provided both a complete set of drawings with dimensions but also descriptions of how the components were made and how the wagon was constructed using them.

 

Archie Croome started out as a draughtsman at Metropolitan Railway Carriage and Wagon Co. and went on to win prizes for his railway rolling stock drawings. Subsequently, he moved to 'take up a responsible position with the Midland Railway' and was there when the articles were published. It seems probable that he was known to the aforementioned Mr T.G. Clayton.

 

Comparing the partial set of drawings in the Ince book to the rather more complete set in Archie Croome's articles reveals some small discrepancies that we will come back to later. Suffice it to say, that I decided to use the Archie Croome drawings as the starting point for drawing components for Gloucester wagons that I could use as the basis for comparison to photographs. In terms of building the first Gloucester wagon, I used the T George 5 plank example as my starting point as this was contemporary (built November 1899) with the drawings.

 

Conveniently, the T. George wagon is illustrated on the Lightmoor web site - https://lightmoor.co.uk/books/private-owner-wagons-of-gloucestershire/L9235 

 

The Archie Croome articles were reprinted by Dragonwheel Books in 2005 but copies seem hard to find (oddly my copy came from an eBay seller in the USA) 

 

Next time: establishing which RCH components match those on the Gloucester wagon

Edited April 2, 2022 by Andy Vincent
Fixed English (again!)

[Compound2632]

I'm pretty certain the specification drawings were made either by RCH drawing office staff, if such there were, or failing that the staff of the Carriage & Wagon drawing office of one of the larger railways (which in the context almost certainly means the Midland) under the direction of Clayton as the Chair of the Wagon Superintendents' Committee. The drawing of a five-plank 8 ton wagon in the Ince book has some very Midland features and details, though with some interesting differences, most notably the specification of 12" x 5" timber for the main frame members were Midland wagons used 11" x 4½". The bottom door fastening / release is identical to that that first appeared on Midland wagons in 1882, as far as I can establish.

 

Midland influence is unsurprising since it was the line with the largest volume of mineral traffic and hence the greatest interest in eliminating unsatisfactory PO wagons, first by buying them up and supplying its own wagons in lieu and then through the RCH system of specification, inspection, and registration. Thomas Clayton was therefore a major figure in the world of wagons in the last quarter of the 19th century; it would be odd if he was not Chair of that committee. 

 

I'm beginning to suspect that Archie Croome is a pseudonym - which wouldn't be unknown for railway company employees writing on railway matters in their own time or not as an official representative of their employer. There is no-one of that name recorded as a draughtsman in the Derby Carriage & Wagon drawing office, which would be where his talents would seem to lie. There are two or three Archibald Croomes in the 1891, 1901, 1911, and 1921 censuses; the only one who might fit at all was born in Gloucestershire in 1849 but is recorded as living in Paddington in 1891 and subsequently in Sussex - not trace of a career in Birmingham or Derby!

[drduncan]

Have you got the drawings from the Montague book on Gloucester wagons. I have used them to cad and print a dumb buffered 1887 5 plank wagon. See my 3D workbench thread for discussion but the photos  have been trashed by the server failure. 

Regards

 

Duncan

[Compound2632]

28 minutes ago, drduncan said:

Have you got the drawings from the Montague book on Gloucester wagons. I have used them to cad and print a dumb buffered 1887 5 plank wagon.

 

That's the drawing that had me doubting washer plates in the corners. However, it needs to be treated with caution as an exemplar of Gloucester construction. It is the wagon the the Gloucester company submitted to the Great Western in 1894 for inspection and registration, in full knowledge that it did not comply with the RCH specification (not least because of the dumb buffers). It was possibly originally built in 1872. The Great Western refused it, leading to a case before the Railway and Canal Commissioners. This was all set up as a test case with Gloucester acting for the private wagon trade generally, there being a great deal of discontent with the RCH registration system. The trade contended that they were being required to build wagons to a higher standard than those offered by the railway companies for the same traffic. As evidence of this, Gloucester captured and photographed the grottiest, most decrepit Midland ex-PO wagon they could find [Midland Wagons plates 18-20]. This, of course, was one of the 66,000+ PO wagons the Midland had bought up from 1882 onwards, precisely to rid its system of the sort of sub-standard wagon that Gloucester were submitting for registration. Although the Midland's plan was to withdraw and replace such bought-up wagons with ones of its own design, even it could not build as fast as it bought, so the ex-PO wagons remained in service, in the case of the better ones, at least to the turn of the century, it would appear. The outcome of the case seems to have been some relenting by the railway companies through the RCH on the conditions to be met for a "reconstructed" wagon.

 

This is mostly what I've gleaned from Len Tavender's Coal Trade Wagons; he cites a 1968 history of the RCH by P. Bagnell and an HMRS Journal article by A.R. Croall (who he? @Penlan) from 1983. It would be interesting to track down the transcript of the case.

[Andy Vincent]

2 hours ago, Compound2632 said:

There are two or three Archibald Croomes in the 1891, 1901, 1911, and 1921 censuses; the only one who might fit at all was born in Gloucestershire in 1849 but is recorded as living in Paddington in 1891 and subsequently in Sussex - not trace of a career in Birmingham or Derby!

The 'biography' (if that is what it is) that was attached to the articles says that he was 'a native of the old cathedral city of Peterborough'. In terms of his time at the Midland Railway it says that this was in London but I dont know what functions the Midland had in London. Not sure that gets us much further forward! From my perspective, having a set of drawings with a known date is the most useful element along with the extra detail in the narrative. 

[Andy Vincent]

Anatomy of a wagon

 

In order to work out how closely a Gloucester wagon might adopt elements of the RCH 1887 wagon we need to select an example. Mine was the T George wagon shown on Lightmoor's site (https://lightmoor.co.uk/books/private-owner-wagons-of-gloucestershire/L9235). This example was built in 1899 but there are other identical examples from around 1897 through to around 1907.

 

The Gloucester photograph gives the (internal) dimensions as 14'5" x 6'11" x 3'1".  The corresponding dimensions for the GA of a wagon drawn to illustrate the RCH 1887 specification are 14'6" x 7'0" (which equates to 14'11" x 7'5" externally). There would be a one inch discrepancy in overall length and width if Gloucester used the minimum 2½" boards specified by the RCH. However, if they used 3" boards then the external length and width measurements would be identical to the RCH drawings so we seem to be in the same area.

 

A wagon underframe is bounded by the two solebars and two headstocks. Running across the wagon are two 'cross-bearers' which are tenoned into the solebars which help ensure that the solebars remain parallel. These cross-bearers also provide the mounting point for the foot of the 'body knees' (long L shaped wrought-iron brackets) which support the sides either side of the door.

 

Whilst interior views of wagons are notoriously difficult to find, we know the position of the knees and hence the cross bearers by the position of the external strapping that receives the bolts which pass through the knees and planks. On the T. George wagon, scaling from photographs shows that the centreline of the strapping is 5' 2½" from the ends which is identical to the RCH GA. Similar use of scaling shows that the side door also matches the RCH GA at 4' 1".

 

The mortice and tenon joints between cross-bearers and the solebars are held tight by a pair of cross rods (typically ¾" diameter) that passes through both solebars and run just inside the two cross-bearers. The ends of these rods are threaded and their position can often be identified since they have slightly larger nuts. The expected position of these cross rods looks to coincide with the bottom nut on the (internal) diagonal strapping

 

The ends of the solebars are connected to the headstocks by wrought-iron brackets ('diagonal knees') through which the buffer shaft also passes. However, a set of four longitudinal rods (typically 1" diameter and arranged as two pairs) run from the headstock to the inside face of the cross-bearers at each end. Once again we can identify the position of these by the larger nuts. The first pair also serve to hold the top of the draw-bar plates in position. The second pair pass through the end pillars and then the headstocks - usually the second bolt up from the bottom of the end pillar.

 

The last item worthy of note are the strap bolts at either end of the solebars. There are three of these strap bolts at each end with one on the outside (where the solebar bolts pass through to the diagonal knees) and two on the inside where the strap bolt passes through the headstock and then through the buffer guides. The position of the two internal strap bolts is given away by the nuts on the outside. The spacing between these also tells us the vertical separation of the nuts on the base of the buffer guide and from that, the position of the inner two nuts since the buffer guide base is (usually) symmetrical. 

 

Having dealt with the basic framing, we can turn to bits mounted on it. The brake hanger and the axleboxes are uniquely Gloucester and we will come back to these.

 

Turning to the 8 leaf axle springs, these are identical to the RCH 1887 drawings. The bearing spring shoe is also identical to the cast spring shoe drawing. Looking closely at the axle guard, there are subtle differences to the RCH drawing: the side wings join the legs at a different position and the legs are longer. The crown plate and side wing washer plates appear to be the same as the RCH versions.

 

The last item on the solebar, the towing hook, is clearly longer on Gloucester wagons (the loop on the RCH version is 3¼" and 9¼ overall).

 

Clearly, the 1887 RCH specification and drawings form a reasonable starting point but some components will need to be derived from photographs.

 

Next: tackling the axle box - and time to learn from USB?

 

 

 

 

 

Edited April 13, 2022 by Andy Vincent
Removed ambiguity

[Andy Vincent]

Firstly, a big apology for the crazy break in postings on this thread. In part, this was because of domestic pressures - but also in part because the wagon that this thread covers, plus five other variants, two choices of axleboxes with springs and two choices of buffers are all actually going on sale at the Warley show this weekend! This wasnt the original plan and the amount of work needed to complete the design work and get set up to produce them in appropriate quantities pushed other things too far into the background.

 

Now that things have settled down, I will pick up the story over the coming weekend . . . 

[Compound2632]

1 minute ago, Andy Vincent said:

Firstly, a big apology for the crazy break in postings on this thread. In part, this was because of domestic pressures - but also in part because the wagon that this thread covers, plus five other variants, two choices of axleboxes with springs and two choices of buffers are all actually going on sale at the Warley show this weekend! This wasnt the original plan and the amount of work needed to complete the design work and get set up to produce them in appropriate quantities pushed other things too far into the background.

 

Now that things have settled down, I will pick up the story over the coming weekend . . . 

 

I will look forward to seeing them - and you? at Warley on Saturday.

 

I believe I saw one at Scaleforum?

[Andy Vincent]

On 24/11/2022 at 14:15, Compound2632 said:

 

I will look forward to seeing them - and you? at Warley on Saturday.

 

I believe I saw one at Scaleforum?

Sadly not, my kids had a day trip to the WW1 battlefields in Belgium this weekend that requiresd drop-off and collection at around 2am on Saturday / Sunday - adding a drive to Warley in between the two was a step too far!

 

Whilst I wont be there, the bodies were on the Brassmasters stand along with alternative buffers and axleboxes/springs. I am told that they sold well which is gratifying. There are more versions - and other wagons - to follow in the coming weeks

Edited November 28, 2022 by Andy Vincent
Fixed confused tenses!

[Compound2632]

13 minutes ago, Andy Vincent said:

I am told that they sold well which is gratifying. 

 

One to me!

[Andy Vincent]

Way back sometime, I said that I would next talk about axleboxes - and USB, so . . . . 

 

I find axleboxes one of the most difficult items to draw when no drawings survives. Even when not covered in layers of grime, they are painted black and lurk in the shadows in photographs. To compound matters, side on photographs are understandably rare. However, axleboxes do leave some cues.

 

The only real advantage of front on photographs is that the overall height and width of the axlebox are fairly easy to establish relative the wheel diameter (3'1") and, perhaps to a less certain degree in early wagons, the distance between bolt centres on the crown plate (1'0⅝").

 

The other thing we know is that most early wagons had a journal size of 8" by 3¾". Whilst we cant see the journal, we can often see the retaining bolt on the side. In the era of the Gloucester wagon, this bolt was located such that it just touched the lower surface of the journal and was located at the mid-point lengthwise. We also know (from early drawings) that the distance was 6'6" between these mid-points on either side. Since we know that the front face of the axleguard is aligned with the inside face of the solebar (aside from anything joggled!) then we have a second measurement that allows us to start to scale the depth.

 

Armed with this information (ideally taken from more than one photograph) we can produce the best representation we can. However, this is where the trouble starts!

 

Aside from the journal region, the end of the axle were it passes through the wheel and also at the outer end is five and three sixteenths diameter and typical contemporary axlebox has an overall width of 6¾" - 2.25mm in 4mm scale. Herein lies the challenge. Railway modelling has standardised on 2mm diameter bearings and with the best will in the world, having only 0.125mm of material either side of the bearing isnt going to work!

 

In my axleboxes for the Gloucester wagons (the 4N curved base and the ubiquitous 4S rectangular one) I have generally settled for a 0.25mm wall either side of the bearing. The rear of these are slotted so that they work with suspension - which brings me to my USB point!

 

I really admire P4 models but it does seem odd to me that their use of 2mm bearings requires axleboxes to be over scale width. In the same way that USB connectors have shrunk over time, then there must be a case for those who model the pre-grouping era in P4 to use smaller bearings and stepped axles!

[Miss Prism]

Waisted bearings don't protrude too much into a scale axlebox.

[Compound2632]

15 minutes ago, Andy Vincent said:

Way back sometime, I said that I would next talk about axleboxes - and USB, so . . . . 

 

@Miss Prism just beat me with the waisted bearings there. @billbedford's Mousa kits are designed to use these, chiefly because of the design of springing he uses, but this does enable him to use scale-width axleboxes - and he has passed comment on overwidth axleboxes, such as on the Slaters Midland wagon kits. The otherwise excellent MJT range of cast whitemetal axleboxes make roughly the same accommodation you are making. 

 

20 minutes ago, Andy Vincent said:

I find axleboxes one of the most difficult items to draw when no drawings survives. Even when not covered in layers of grime, they are painted black and lurk in the shadows in photographs. To compound matters, side on photographs are understandably rare. However, axleboxes do leave some cues.

 

I'm sure you are aware of the drawings in A.J. Watts, Private Owner Wagons from the Ince Waggon & Ironworks Co. (HMRS, 1998). These include various axlexboxes, including Attock's and Ellis' patent grease boxes. The 6' 6" centres-of-journals dimension formed part of the RCH 1887 Specification. Do you have a copy of that specification?

[Andy Vincent]

9 minutes ago, Miss Prism said:

Waisted bearings don't protrude too much into a scale axlebox.

Agreed (and I assume that they will be used when I make the slot in the back) but it is the 2mm diameter that causes the bigger issue when the axlebox is barely any wider than that as the full width part of the bearing still partially protrudes into the axlebox

[Andy Vincent]

2 minutes ago, Compound2632 said:

I'm sure you are aware of the drawings in A.J. Watts, Private Owner Wagons from the Ince Waggon & Ironworks Co. (HMRS, 1998). These include various axlexboxes, including Attock's and Ellis' patent grease boxes. The 6' 6" centres-of-journals dimension formed part of the RCH 1887 Specification. Do you have a copy of that specification?

 

I have the Ince book and also the RCH 1887 spec which, in the copy I have, includes the Ellis axlebox. 

 

The drawings in the Ince book puzzle me: the text relates to the 1887 spec but taking the buffer shank drawing, this shows a buffer shank length of 10" rather than the 9½" length shown in my copy of the 1887 spec. My suspicion is that those drawings (or at least the buffer shank) relates to the 1903 spec.

[Compound2632]

1 minute ago, Andy Vincent said:

My suspicion is that those drawings (or at least the buffer shank) relates to the 1903 spec.

 

I agree; the dates of the drawings are not all as clear as they might be.

[billbedford]

1 hour ago, Andy Vincent said:

The other thing we know is that most early wagons had a journal size of 8" by 3¾". Whilst we cant see the journal, we can often see the retaining bolt on the side. In the era of the Gloucester wagon, this bolt was located such that it just touched the lower surface of the journal and was located at the mid-point lengthwise. We also know (from early drawings) that the distance was 6'6" between these mid-points on either side. Since we know that the front face of the axleguard is aligned with the inside face of the solebar (aside from anything joggled!) then we have a second measurement that allows us to start to scale the depth.

 

The spring is centred over the journal. So the front face of the axlebox is around 6" from the spring centre. Actually 6 1/8" for MR or 6 1/4" for LNWR with raised beading on the front face. 

[Miss Prism]

2 hours ago, Andy Vincent said:

as the full width part of the bearing still partially protrudes into the axlebox

 

A bit of discreet countersinking on the rear of the axlebox might be necessary, but if the 2mm part of a waisted bearing is protruding significantly into your axlebox, you are mounting your solebars/W-irons too close.

 

 

Edited November 29, 2022 by Miss Prism

[Andy Vincent]

9 hours ago, Miss Prism said:

 

A bit of discreet countersinking on the rear of the axlebox might be necessary, but if the 2mm part of a waisted bearing is protruding significantly into your axlebox, you are mounting your solebars/W-irons too close.

 

 

I think we probably all largely agree with each other but I went back to look at my calculations again - which were based on Russ Elliott's 2004 re-drawing of Mike Trice's 1982 specification that was referenced earlier - to check numbers.

 

The drawing shows that the 2mm part as projecting a maximum of 0.6mm behind the outer rim and that the bearing should narrow to 1mm diameter at or before a projection of 1mm behind that same outer rim.

 

Assuming an etch of 0.3mm thickness and the spring carrier being half etched then this comes to a maximum of 0.45mm (or possibly 0.3mm without springing if we assume the same width between solebars). This leads to the conclusion that the full width protrusion into the axlebox assembly is a maximum of 0.3mm and that the narrowing to 1mm diameter should happen at a maximum protrusion of 0.7mm. Incidentally, this narrow part should protrude into the axlebox by a maximum of 1.4mm (or 1.25 with springing).

 

The front part of the axlebox that retains it between the axleguard legs is specified as being a minimum of 1" and flaring slightly towards the axlebox body. Thus it is at least 0.33mm from the axleguard which should accommodate the maximum 0.3mm projection, albeit with very little tolerance, and we can assume that diameter has begun to narrow towards 1mm by the time we encounter the body of the axlebox, albeit that it wont be much below 2mm at that point. 

 

The approach used by @billbedford is, cleverly, slightly different in that his version of the non-sprung case still uses the (printed) spring carrier so gains some extra tolerance - the same would be true of using fixed axleguards but adding a half etch washer between bearing and axleguard inner face.

 

All of which is probably a long winded way of saying that we are all in agreement but that a worse case exists of fixed brass axleguards used without a washer. Maybe we are at the point where we need to move on from considerations of those who used fixed axleguards without that extra washer ('just bend them in a bit') so that we can achieve that higher accuracy.

[Miss Prism]

Do you want me to ask Russ Elliott to do a clarifying sketch?

 

[Andy Vincent]

6 minutes ago, Miss Prism said:

Do you want me to ask Russ Elliott to do a clarifying sketch?

 

No, I am happy with the one he has on CLAG but thanks for the offer.

 

As an aside, what would be interesting to know is how well those waisted bearings produced today conform to the spec. Admittedly the manufacturers seem to have not changed much but I get the impression that the rear face needs more cleaning up than they used to. I actually print a hole on my axlebox/spring carrier to make it easier for people to hold the bearing whilst filing the back flat for just that reason

[Compound2632]

45 minutes ago, Andy Vincent said:

As an aside, what would be interesting to know is how well those waisted bearings produced today conform to the spec. Admittedly the manufacturers seem to have not changed much but I get the impression that the rear face needs more cleaning up than they used to. I actually print a hole on my axlebox/spring carrier to make it easier for people to hold the bearing whilst filing the back flat for just that reason

 

Some time ago, when I built my first Mousa kit - LNWR D32 - I tries Alan Gibson waisted bearings but ended up with axleguard splay. So I tried the MJT ones instead, which turned out to be just right. I think the issue was the depth of the inner cone, rather than the external shape. I may just have been unlucky with the batch of Gibson ones but I have been twice or maybe now three times luck with the MJT ones. Perhaps it's a case of it paying to go back to the original!