Downton Train Crash of 1884 in EM Gauge

[Herb Garden]

What seems like ages ago, this thread was kicked off by a slightly random and perhaps ill-considered purchase of a K's Milestone kit for a Falcon Class loco. This engine if built would require coaches not available ready-to-run. which thanks to a BoT Accident Report and the work of Gordon Weddell, now exist for a specific but unfortunate train - though currently only in a virtual world. On the day of the accident, the train engine was Vesuvius Class 294, which was a development of the Falcon Class, of similar size, yet differing in many details, so really I need to consider the work needed for the loco next.

With almost all of the crude K's underpinnings missing, the next task was to develop some sort of etched chassis that would enhance the kit and could be built to EM (or P4) Gauge. The original idea was to quickly knock up a simple Perseverance style chassis, but I very quickly established that this isn't the prototype for anything remotely sensible! So, using the Hollywood Foundry guidelines for dealing with cusp, and the PPD guidelines for colouring the artwork, I've used QCAD over the last three weeks to design several different sets of components - these being (so far):
- Smokebox Front and Cosmetic Frames
- Crossheads and Slidebar Assembly
- Loco Frames, Rods and Brake Gear
- Tender Frames and Brake Gear
- WG Beattie (Stirling style) wrapover Cab
View attachment 221472
Following tradition I thought I'd start with the tender and using a 'road kill' layout these are my plans for an etched chassis in 0.4mm nickel silver for the tiny Beattie 1950g tender. I'll be compensating it using twin beams with a trim screw above the trailing axle set in a dedicated frame spacer, with the wheels mounted in High Level Kits Miniblox - hence the 4mm wide half etched slots. These tenders have been photographed with large wooden brake blocks, but the brake arrangement shown here is a later development. The third layer of brake shoe is there as an option to add chunkiness. Hangers appear to be quite delicate in photos and drawings, and I'll be thinning mine slightly as on reflection I've overcooked them. It took me a while to get to grips with the guard irons, until I realised that they could only be an extension of the brake hangers. At the front I've provided something to hang the brake actuating shaft from and an operating arm to go below the brake standard. All of this fits inside the unmodified white metal tender body and I have had to move the top brake pivots forward slightly or the rear one would have been off the end of the frames and buried in whitemetal.

The mid-blue is half-etched from the front and will need to be pure blue in the final artwork. The red is half-etched from the rear. As a novice this colour scheme is something I can get my head around. Confusingly QCAD 'colouring in' is described as 'hatching', it's very sensitive to surrounding lines not being joined 'properly' leading to some rework. Sometimes the 'hatching' completes, but it is then lost (with an unhelpful error message) when the file is saved. This is also due to line intersections needing rework, and always in my experience where a curve joins a straight end on. All this is going to make adding the moats and tabs required for responsible chemical etching quite tricky! I hope to retain the 'road kill' layout as it makes the etch and required assembly easy to understand.

Currently each QCAD drawing for etching has five layers:
- Finished Size
- Cusp Allowance (added or subtracted)
- No Etch - Black
- Half Etch from Front - Blue
- Half Etch from Rear - Red

This approach allows me to keep tabs on where I am in making allowance for cusp as I can always see a 'before' picture when making any changes. I have also made a GA drawing where each logical set of components is on its own separately coloured layer. I use this to understand how the different parts relate to eachother as I can easily switch the visibility of sets of components and bitmap images in and out by layer.

Luckily the excellent Hollywood Foundry (HF) guidelines for etching had been safely squirelled away whilst still available. In my researches I'd seen guidelines that suggested adding 20% for cusp, but that seemed to be over egging it and I was thinking that in theory 15% might be nearer the mark. I was therefore reassured that HF did indeed suggest adding 15% (o.o6mm in this case) all round, apart from slots where to avoid the difficult task of opening out a slot you should remove 15% on each side, but not ends. It seems reasonable that holes could be a special case of a slot, and for the holes in the tender I've yet to make any allowance for cusp.

I'm new to this so correction on any aspect would be appreciated!

Your approach looks good can't see any faults as a someone who also designs etches.

I also have a Ks falcon that needs rebuilding so do let me know if you accidentally make a second set of etches and want rid of them....

 

[AJC]

Very interesting, Steve - I don't know much about artwork drawing (I just build the results!), but I can't see anything that would trip you up. Would a hole etched through the brake shoe overlay, in order that you could insert wire all the way through for location, be an improvement? I think it might. Obviously the fold-over tab does allow for that, but a hole through is belt and braces. I see that the actuation gear is hidden behind some convenient steps: even I would ignore that. Looks very positive.

Adam

 

[Stevers]

Would a hole etched through the brake shoe overlay, in order that you could insert wire all the way through for location, be an improvement? I think it might. Obviously the fold-over tab does allow for that, but a hole through is belt and braces. I see that the actuation gear is hidden behind some convenient steps: even I would ignore that.

In general it all depends whether I've ended up doing this for me, a few hardy and experienced souls, or a very secret army of Victorian era LSWR modellers. The two layer approach would suit the latter best, but I don't think that's something I need to worry about! When I hand fret brake blocks and hangers, I start by putting in the holes to line them up for assembly, so I'll make that change, as the hardy souls will like that too.

 

[Stevers]

The 'finished' tender chassis fret:

In a surprise development a 3D printed body to go on it. Jolly useful for a model of Lion Class Stour, or for anyone planning to convert a Beattie Well Tank to tender form, as the LSWR did when these tanks were retired from suburban duties.

I'd like to think I'm about half way through, but am only too aware that's not how it works. Interestingly the K's tender castings hold up well to scrutiny and I still plan to use them for 294. The K's rivets are a little bigger and a bit further apart and I'm in two minds as to whether that's better or not in a model where the prototype rivets are so tightly packed.

There's a vertical handrail on the other side that I may want to etch some brackets for, in which case I'll add those to my 'finished' chassis etch.

 

[Stevers]

We're at the point where I am starting to produce STL files that could be printed.



This is a view of the top with it's little shanty town of tool boxes. The tool box at the back is metal and has stops on the hinges to stop the lid from falling back. I'll probably organise an etch for these, with a recess in the lid to take the etch. There should also be a staple to secure the lid. Forward of this are two wooden chests that I've sized and positioned to match those on 294. On some tenders these were longer and of slightly different design. The lids on these were I think hinged, but I've yet to represent hinges or fastenings. In almost all pictures these boxes are partially obscured by a carefully constructed mountain of coal. Forward of that is a small wooden chest that appears to have a lift off lid that is flat or almost flat. This is rarely level front to back in photos, so I suspect this is just sat on a shelf. This one is positioned according to the photo of 294 and it was the resulting ugly gap between partition and box that first led me to think there must have been a shelf of some sort. I suspect that a shelf was both sides although a box was only visible on one side of the tender on 294. I've put some representative rivet detail on the tank top and into the coal hopper, and the flat filler lid has holes for a handle, there being no evidence that it was hinged.

There was a problem with the rendering process. This takes over an hour on my laptop, and about 25 minutes on our tower PC that has better graphics support. In order for the bead along the top of the flare to appear curved I'd set the facet number to 600 and the full render process didn't like this number to be above about 300 for this item. At 300 the very gentle curve was rendered as two straight rods. I changed the facet number to 20 and recoded the bead as a series of 10 straight lines that appear as a smooth curve. This allowed the render to complete without error, in less time, with reduced file size, and with a better fit to the flare - everyone's a winner! A major consideration in reducing file size and processing times is facet size, particularly for all those lines of rivets where the facet number used is 10. For comparison the facet number for the leaf springs is 120, and tank filler 60. There are facet size and facet angle parameters in OpenSCAD, but I've had no success with those with anything other than the most basic of shapes.


There is little information on the tender fronts, though I'm reasonably confident about the brake standard, and the sand hoppers with cranked pipes giving a straight drop to the rails. The GA doesn't show any kind of shelf below the coal hole, although there is a suggestion of one in a photo. It looked silly without one, so I've done the minimum and added a door and some rivets to give the whole a bit of life. The front handrails will be minused out for the print. This hole will be a nominal 0.65mm, but will print at about 0.45mm as represented here. The tender handrail brackets will also be minused out ready to receive the half etched brackets in the chassis etch. My plan for the brake standard is a handrail knob and bent wire to represent the handle. The wooden tool boxes and shelf components will be separate prints.



The original 3D model was based on the drawing in MRN from 1965, cross checked to photos. There is a GA of one of these tenders (as designed) in the relevent Bradley and this gave me the layout for the lamp brackets. I already had my doubts about the flare and was not too surprised when I overlaid the layout from the GA onto my 3D model that the top lamp bracket was only half way up the flare. Anyway a written measurement on the GA confirmed that the top of the tank in the MRN drawing was 1mm too high and the top of the flare a further 0.5mm too high. Photos could now be seen to support the GA dimensions. After a bit of rework, the top bracket was where it should be and the beads and rivet lines all tidied up again. The shunters handrail/water gauge is just a length of wire mounted on two long handrail knobs with an extra knob at the base to better represent the prototype detail. Visible underneath are the etched chassis frames positioned just to check fit.

Something I haven't considered yet is the drawbar. It should go through the dragbeam, but that's not usually very convenient in model form, so I'll run it just underneath the dragbeam, which I might fillet out a bit so the drawbar can be nearer to where it should be. Next step is to start separating out the different parts and complete all the minusing out that's needed for a 3D print.

 

[Stevers]

Some fairly serious upgrades have followed a lot of help from a friend. I'd traced the leading axlebox and hadn't realised that it had a 3' leaf spring, and the others 3'6". Quite clearly so on all my drawings and a howler that's now been fixed.

A photo of the front of the tender has now been seen, and I have been able to make mine look more like at least one example of these tenders. So definitely no shelf, big rounded sand hoppers (as also glimpsed on one of the Beattie Well Tank conversions in Bradley) and just room for a single step. The lift up door is so large that corners needed to be trimmed so they would be less likely to cause injury on any ascents of the coal mountain to reach the back of the tender. I've already enlarged the coal opening and am wondering if logically it shouldn't be a little larger at the expense of the door. I've only suggested the drag beam and buffer detailing as it will be below a fall plate. There are at least three sorts of sanding hoppers on these tenders that I would describe as 'narrow tapered' as per K's kit, 'wide tapered' and 'rounded' as this one is.


Models of the Ilfracombe Goods tend to show the coal hopper filled in below the forward tool boxes and it doesn't seem impossible that this somewhat dead area of the hopper might have been boarded in, giving somewhere for the tool boxes to sit on at least some of these tenders.

 

[Mudmagnet]

Test printed the tender today and works really well. Will post to Steve to evaluate.

 

[Stevers]

An excellent quality print - very pleased with it! All those rivets, the flare, the steps, the axleboxes and the spring detail have come out really well. There should be a second large tool box, and the axlebox retainers look a little awry. The latter doesn't seem unusual for the parts of a print nearest the plate. I'm thinking I need to add sacrificial rails along there so that the print can get going properly before it gets to the retainers.

 

[James Spooner]

View attachment 227945View attachment 227946

An excellent quality print - very pleased with it! All those rivets, the flare, the steps, the axleboxes and the spring detail have come out really well. There should be a second large tool box, and the axlebox retainers look a little awry. The latter doesn't seem unusual for the parts of a print nearest the plate. I'm thinking I need to add sacrificial rails along there so that the print can get going properly before it gets to the retainers. I'm thinking that needs a sacrificial rail that can be cut off to allow the print to get going properly.

That’s looking very impressive Steve! Can’t wait to see the rest of it!

cheers

Nigel

 

[Stevers]

Can’t wait to see the rest of it!

Some of you may have guessed by now that the 'rest of it' was likely to be a model of the Lion Class 0-6-0 that piloted the train on that fateful day. So over the Christmas Holiday I thought I'd explore some options for modelling Stour. This process turned out to be slippery slope. There were over forty examples of the class, and there are photos of a subset of these engines in Bradley, but none of Stour. The only drawing that I have access to is from the Locomotive Magazine from 1933 published in Mike Sharman's collection of drawings. This is only a side view, but I scanned, scaled and traced it in QCAD and then started to develop plan, front and rear elevations from the very few known dimensions.

Tipping over the crest of that slippery slope, I modelled the boiler unit in OpenSCAD with a view to 3D printing it.


The boiler needed a firebox, and it needed to be a double one as the Lion Class received the latest development of one. The only drawing I had was of 7' 2-4-0 Clyde of similar vintage, but once traced it was found to fit inside the Lion Class cab like a glove. This is also the backhead that the crude example in the K's kit was clearly supposed to represent so this will be an upgrade for Vesuvius Class 294.


How to make the rest of this tiny somewhat humpbacked loco though? Chemical Etching seemed the way forward for cab, splashers, running plate etc., so I started to mock up the etched shapes I would need and started to assemble them around the boiler and smokebox unit. This allowed me to pose my 3D model against photographs allowing me to adjust my jigsaw puzzle of etched parts to improve the proportions, and confirm fit of the different components. I now had a boiler unit, sitting on a reasonable looking running plate with the filigreed splashers and temporarily with an Adam's cab. To robustly model the 'face' I'd elected to half etch a smokebox face surround so that I could represent the handrail brackets and lamp irons, with the fastenings for those being pushed through from the back. The smokebox door and hinges, would be 3D printed as would the sandboxes and buffer beam. There's a gap behind the buffer beam for access, and the cylinder block is angled to align with the crank axle.


Without chimney or dome it looked a bit very silly. Adams stove pipe chimneys are available commercially so I didn't think that would be a problem, but I'm not aware of any source for those distinctive Beattie domes. As has been mentioned earlier in the thread OpenSCAD has no tools to flare one cylinder into another, but I knew conceptually that rolling a ball round the junction of the two cylinders must just be a matter of some convoluted trigonometry. I wanted this to be a generic capability so I parameterised the boiler and dome size and the required flare. I managed to hang segments of flare round the dome and calculated the correct height of each segment using my somewhat rusty trigonometry so that the corner sat at the junction of the two cylinders as shown above. As can be seen this wasn't the whole story and doesn't reproduce the 'ball rolling round the junction' effect that was required.


Having an engineering background rather than being a mathematician I resorted to a more empirical approach to resolve this problem by adding a 'fudge factor' (in this case 1.55) for the vertical axis that can be adjusted to further drop the curve of the flare in the correct position for the two cylinders. This left a bit sticking out each side of the 'saddle' and this was removed by an intersection operation with an elipse controlled by a second fudge factor (in this case 0.95). No doubt a better man than I could come up with a suitable formula for these two crucial values, but I'm more than happy with where I am with this approach. For Stour and 294 these domes would have been painted, so they will be printed in positon on the boiler, though hollowed out for resin drainage and for weight to be added.


This is the current state of play with the loco attached to the tender. The rear dragbeam (and running plate) is currently a scale 6'6" wide and that seems very narrow. The running plate is visibly 6" narrower than the bufferbeam, that in itself is about the same width as the Beattie tender buffer beam. Photographs confirm that the loco was narrower than its tender but by this much? The Stour of 1884 would have had a sort of pressed steel looking Beattie cab, and the CAD for that will be cardboard based to try to work out what an etch for that should look like.

 

[James Spooner]

Steve

I‘m very impressed by your mathematical skills and the loco looks to be coming on very well!

cheers

Nigel

 

[AJC]

Looking good thus far, an interesting mix of phases in the one machine. I’m certain that I have a Gibson Adams stovepipe somewhere. If I find it, I shall pass it on via dad (I have a notion that it’s in the box my O2 came in, which would be logical, and that’s in a chest of drawers in Yeovil (so if you’re watching, dad…)).

Adam

 

[Stevers]

very impressed by your mathematical skills

All smoke and mirrors - trust me!

 

[Stevers]

To try to confirm (or otherwise) the accuracy of the Lion drawing, the domes on the Lion version were compared with the ones on the larger scale and more detailed Clyde drawing. On the Clyde the rear dome was shorter with a bigger bell and that seemed to match photos better, so I modified mine to match, incorporating more detail in the open bell. This dome will later gain those strange Salter safety valves The front dome was a good match other than the shoulder radius which was greater, so that little change was made and whatever is sticking up in the middle was added to better match photos. As shown here, the domes are now integral with the boiler 3D print and hollowed out for resin drainage and for weight to be added. Rendering that little lot nearly bust a blood vessel on a high end laptop from 2015.

Next the mainframes and firebox grate were added - lots of scope to add weight down there when driving the centre axle from above. I then knocked up sets of EM and P4 wheels and 'fitted' them. It was immediately clear that even EM 'fine' (e.g. Gibson) wheels did not have adequate clearance under the near scale splashers. The idea of 19mm wheels was toyed with, but neither ScaleLink (crank pin between spokes) or Markits (15 spokes if available) seemed that good a match so I opted to increase the splashers by 5% to gain 0.5mm clearance all round to fit 20mm wheels. Unusually 'generic' 20mm Romfords are actually a perfect match in respect of size, number of spokes, crank pin position and a normal crank throw! This change required me to raise the sandboxes and change the shape of the bit of filler behind them, but this put them where they should be in relation to the smokebox. The boiler is quoted as 4', so that's what I made it, but if the cladding wasn't included in that dimension then my boiler is a tad small. The topline of the modelled boiler matches the Lion Drawing, so if the boiler was larger it would push the cylinder covers even more out of sight behind the buffer beam, and that wouldn't be a good thing - oh the uncertainty!

This loco needs to run well and to go round 3' curves. At scale positioning there will be insufficient room in the splashers each side of the centre drivers for the necessary side play, so the plan is to add 0.5mm width to the front and centre splashers. At the same time the running plate might be widened to match the tender and perhaps 0.5mm added to the cab. Although the tender tank is wider than the cab, there doesn't look to be as big a difference in photos as has been modelled to date. Looking at the buffer positions it is possible to see just how narrow the Beattie tender is too. The tender is on P4 wheels for contrast - well that's my excuse.

In adding the mainframes it was noted that the raised strengthening sections over the hornblock cutouts clash with the splashers inside the cab in their current position, and that there's no room for equalising beams to act on hornblocks. My preference is to drive the centre axle, so I will fit the twin compensating beams on the front axles and put a single pivot at the rear. This (one likes to think) will give three point stability 'Morgan' than 'Reliant' style. This approach would permit a fixed rear driven axle and single or double beam at the front for anyone else having a go at one of these. A friend will build one in 00 Gauge driving the rear axle, so his needs are being considered too. Currently there is the option of printing the boiler/firebox or backhead with P4, EM or 00 Wheel clearances, P4 having no cutouts because the wheels are where they should be. Cab splashers and footplate will have half etched lines for the three gauges as appropriate, so that the cab footplate is the maximum possible width for each gauge.

 

[Stevers]

As mentioned previously the boiler looked a tad malnourished, and in addition the border round the smokebox door seemed too narrow. The Vesuvius boiler is stated to be 3'10" same as Falcon (so at least the K's boiler should be OK). The Vesuvius GA is detailed enough to show that the quoted diameter excludes cladding which looks to be between 1" and 1.5" There is a line on the Lion Class drawing that could be the bottom of the boiler at 17mm diameter, which would align with the 4' boiler diameter plus 1.5" cladding. The bottom line was that the boiler was too small and a bit too high. Attached is the view of it with the wider boiler - 0.5mm all round,with the 'Clyde' firebox widened to match. It probably should be pitched a litte lower, but then the cylinder covers would sink even further and that would look wrong. The smokebox door doesn't seem overlarge, so it's unclear quite where the problem lies - if there is one. There are compromises with the splasher height that can't be ducked, and this is probably the least worst compromise for proportions, and a start can be made on adding detail again, particularly the front buffer beam.

View attachment 231355

This is a view of the frames at 00 spacing arranged for the rear axle to be driven, although what sort of gearbox will fit under the cab floor is moot. The builder in 00 Gauge has asked for the front splashers to be at scale spacing, and that is one advantage of 00! I'll also provide slots in the frames so that the small spacer could also go in front of the backhead, allowing the centre axle to be driven from behind - thus minimising intrusion into the boiler.


And here the frames at EM spacing with spacers arranged so that I can drive the centre axle from the upper firebox with lots of weight added in the bottom of the firebox. The HLK RoadRunner Compact gearbox just clears that spacer that's in front that's intended to mount pickups. There will be an etched cosmetic representation of the frames going forward to the buffer beam tabbed into the running plate. No doubt there were angle irons securing the buffer beam and I will be able to 3D print them onto the beam. The cosmetic bufferbeam can also be impaled on the frame extensions for a good solid fit. The etched underpinnings from this point on will be straightforward and won't require 3D modelling for fit.

 

[Stevers]

For a change some slightly messy modelling.

Stour was freshly outshopped at the time of the accident (and reportedly in fine fettle), but it had yet to receive an Adams' style cab as shown in the renders. This meant that I needed to think about what a Joseph Beattie 'pressed steel' cab would look like flattened in an etch ready to be folded up, and my best guess was arrived at working out what it should have been at the ends, and then eliminating all the things it couldn't have been at the corners if the two edges that needed to meet were the same length. Not the sort of thing I could leave to chance, so I printed the thing out and UHU'd it to a degreased piece of 0.36mm nickel silver sheet and fretted the proposed shape out.


Attempt one was an abject failure in that the bends were too tight since it was impossible to regulate them even with appropriate sized steel bar, so for attempt two I made an aluminium former that would exactly fit inside the finished cab, with a second piece the same width that would hold the front of the piece flat while doing the folds - all without damaging the piece which was proving easy to bruise. Having rounded off the corners of the former an appropriate amount, the end result was quite pleasing with the resulting gaps at the corner well within the scope of solder filling.
Not obvious in the flesh is that my former needs slight correction to the top curve. The windows are a little large, but there are no plans to use this test piece for an actual model. The problem here is that we're trying to represent material which is maybe a scale 1/8" thick with something that scales at over an inch thick, hence the bulging gaps at the corners - well that's my excuse! This style of cab requires an alternative cab side/splasher, and this will be included in the etch with the parts for the Adams' style cab pictured in the renders.

In the meantime work has proceeded on detailing the 'face' including putting some handy flanges on the smokebox ready to receive some pipes. The buffer beam has been detailed with the (etched) loops for the safety chains and the lamp iron brackets that photos show must have been on the rear of the big wooden buffer beam. It will be printed without the buffer heads ready to receive steel sprung heads, and with slots to take the safety chain loop and frame inserts. Just three enormous coach bolts each side holding that buffer on as far as I could see! I've added a short section of cosmetic frame slotted and tabbed into the front of the running plate that I can impale the buffer beam on, and that will carry the guard irons at the correct spacing, this conceit hidden by the brake hangers. The darts exactly match one of my reference photos, and the front handrail knobs really are just nice plain cylinders. The glow of red inside is the envelope of an HLK Road Runner Compact gear box driving the centre axle from above.

Having decided that the underpinnings didn't need to be 3D modelled in CAD, I had another look at photos, and the arrangements for the retro- fitted steam brake on the Lion Class are distinctly unusual, need careful thought and I reckon a most unusual solution.