Verwood, Dorset - 3D Printed LSWR Type 1 Signal Box

[Steve Smith]

Verwood Station in Dorset has been a long-standing interest of mine, and I have been assembling information, photos, buildings and rolling stock ready to produce a model of it as it was from the mid-fifties until closure in May 1964.  It is a compact little station in a bucolic setting with a pub in the station yard, and a sparse though varied train service.  The line was opened in 1866, and the LSWR Type 1 signal box was probably provided between 1873  (when electric telegraph was in use) and 1875 when Absolute Block was recorded as being in use in the Board of Trade Returns.    Verwood also had a Ground Frame at the West Moors end of the loop to control access to the goods yard and this was released from the signal box.    One end of the box was built tight up to the canopy and there is a photo taken after closure that shows that this end was not weatherboarded; the diagonal bracing being almost entirely covered in poster boards.   A final piece of the puzzle was provided recently in the form of a photo of the back of the box - something that I never thought I'd see, and confirmed that the brick chimney at Verwood was quite different to the one at Instow that I'd used as the model for my laser cut and engraved box shown below:

This was a test build that utilised some 1mm grey card for key elements, as well as 1.5mm MDF and 300gsm card.  The grey cardboard was a most unsatisfactory material, lacking strength and very grainy in appearance.   This box was moreover a lot of work to build and partly due to that grey card not at all crisp in appearance.  The files that this was cut from had been amended (fixed) during the extended cutting session, and due to a laptop crash those files were lost forever.  Attempts have been made subsequently to get the road overbridge at Verwood laser cut, and the conclusion that I have come to is that the process has severe limitations, starting with a complete lack of standards for file formats and line colours.

 

The box was first drawn in pencil in 1992, and was then redrawn in CADRail, from where it was imported into QCAD and used as the basis for the current 3D Model.   Last week I bit the bullet and 3D modelled the chimney thinking that if nothing else it could be fitted to my laser cut box.  A view of the chimney attached to the rear wall is shown below.   QCAD was used to develop cross-sections that can be extruded and assembled into a 3D model in OpenSCAD.  The brickwork bonding is an extruded 2D drawing as is the weatherboarding and ogee guttering.  The mortar is made using OpenSCAD commands, notably 'cube' which doesn't just do cubes.  One of the issues that I have had with 3D printing is getting the bricks to look right with the Slaters brick sheets.  To this end, a 0.2mm gap between the bricks has been settled on, but the depth of mortar is parameterised and can be adjusted prior to exporting the .stl file for 3D printing.   In 3D printing, holes print too small and rods too large, so I would expect the finished mortar gap to decrease from the nominal size, but hopefully not disappear!  

The result from the chimney seemed pretty good and the STL file could be imported successfully into Lychee Slicer for supports to be added, so I pressed on with the canopy side and then the roof.  The roof used 11"x22" slates of which there were nine courses, and how making that fully parameterised worked out will be the subject of the next instalment.

 

 

 

 

 

 

 

Edited July 8, 2022 by Steve Smith
Have restored just the photos from the first and last few entries as I only have the final versions of everything. The photos of my mistakes on the way through are lost for ever! :-( Have replaced these photos for the second time....

[PaulRhB]

It certainly captures the charm of the small LSWR box. Are you 3D printing the windows too?

[Steve Smith]

17 minutes ago, PaulRhB said:

It certainly captures the charm of the small LSWR box. Are you 3D printing the windows too?

Thank you, and yes, a super little box, but probably no larger than it needed to be!   The glazing bars for single glazing are usually terrifyingly thin, and the one thing that laser etching did for me right from the start was to produce superb looking window sashes.  I use what I would describe as off-white cartridge paper.  The back of the strip of etched windows is sprayed with 3M adhesive, fastened down to the glazing material from where each sash can be cut out.  Rather than try to re-invent the wheel, I was planning to use my last remaining laser etch set.  It would be dead easy for me to order some more of these though, and I think they will prove easier to realistically glaze than in-situ 3D printed ones.  The intention is that the sashes will be capable of sliding on the canopy side, the inner sash of which seems always to have been open when the station was manned.   The 3D rendered sashes shown can be switched off for the production STL file.

[Steve Smith]

So what of this roof then?   Using the ideas and code I'd developed for the rear of the box I fairly quickly developed the canopy elevation, the only novel elements of which were the windows.  Lacking reality (and the laws of physics), the two sides now really needed a roof - if only for photographic purposes.  

The slates were 22x11 or  'Marchioness'.  Other size slates existed on buildings at Verwood, so I didn't want to be tied to that size only.   It therefore seemed essential that the slate size, gap, 'step', lap and generated roof area would all be input parameters to the process.  This enabled me to do a calculation for the number of slates in each row, and for this hipped roof how to centre them.  (For gable ended roofs there is the additional complication of finishing on a full or half slate at each end, but that's not relevant for this building.)    Also, the number of rows could be calculated allowing for the top row of slates also being shorter.  Two additional rows were printed at the start to avoid placing any half slates at the beginning, these would be trimmed at the end.   Each row of slates was printed flat, and each new row stepped up from the last.  At the end I did a bit of trig. using going and step to level the finished sheet ready for further processing.  An afterthought was to add a slight random rotation round the nail end of each slate, to render the finished slates just a little uneven as you see in real slate roofs including the two that I've worked on!  This worked much better than raising and lowering the whole slate.  I should add that real Welsh slates are very thin, even thinner than glazing bars!   

Each pitch was generated individually with boarding and rafters applied underneath, then the whole raised to the correct angle, lead flashing added (as a 2D extrusion) on hip and ridge, then hip and ridge boards applied underneath, all using a lot of so nearly forgotten trigonometry.  Then each pitch was trimmed (effectively a mitre) all around, ready to be printed.   Oh yes, and on the chimney side a lead valley was added behind the chimney and a then a rebate for the chimney with its leadwork to butt up to it from. 

 

Since finishing the roof, I have modelled the front elevation complete with that deep set locking room window.  A 3D printed vent and name board have also been added and these can be seen in the flesh in my first post as they were prepared for the laser cut box. 

 

The whole roof took just shy of 250 lines of procedural OpenSCAD code and five .dxf drawings, two of which probably weren't needed the way it turned out.    No 3D Fusion 360 style wrangling is required for this model, so an easy way into 3D modelling for anyone experienced in 2D CAD.  What it can't do sensibly do is flare a chimney or dome into a boiler, but that's not a problem with most buildings - or of course yard cranes - see Verwood (and Semley) Yard Crane.

 

 

   

Edited April 3, 2022 by Steve Smith
Restore photos - some were lost forever.

[BrushType4]

Interesting. A good idea to mix 3d printing with laser cutting. The best of both.

 

My version of LSWR was fully laser cut though.

[Nick C]

That looks great Steve - will you be offering prints for sale at all?

 

If you need any photos of Type 1 boxes (including interior), I have loads of contemporary images of Alresford on the Mid-Hants, as I'm a signalman there. 

[Steve Smith]

Thanks for your kind offer Nick, I shall take you up on that!    The elephant in the room is the huge question of whether it can be printed successfully.  The Verwood Crane is a very different beast, and strangely the most successful prints of that were not commercial, but made by a very talented YMRG'er who's moving away from Somerset and giving up his share in a 3D printer.   Once I have a 3D printer and somewhere to put it, prints could certainly be made available to those interested, not least to recoup the cost of the 3D printer plus the essential wash and cure station.   It would be a shame if there was only ever one of these models, so expressions of interest would be most welcome!

 

As many of you will be aware, Verwood should have much in common with Alresford, and I have used the photos that are on the MHR website as inspiration for the fireplace and the inside of the roof.  Yesterday Martin Finney was kind enough to send me a drawing of Semley box that had been prepared for his model, and that was based on surveys of the surviving Type 1 boxes at Alresford and Instow.   I had counted the bricks on the front (successfully as it turns out), but there was no chance that I could do the same on the sides, so for that I plumped on a round number of bricks probably based on the BR plan from the Waterloo Arch, and to me the result looked credible.  Perusing Mr Finney's drawing and looking at the proportions of the end windows I realised last night that I was two whole bricks out on my estimated depth - doh!    Anyway, after a quite a lot of intense rework yesterday evening, this is the albeit still slightly rough result:

The last few test renders have been done without cladding fitted, and I do need to track down the relatively few photos that exist of these boxes in this condition if I am going to take such a version any further.  The triangular cill profile is most unusual, and I suspect just a capping for the cladding.   Resizing the roof went surprisingly well, but the trimming of the roof has not been completely successful, and I need to look into that.

The back of the box had a narrower weatherboard than the front and porch side, and that suggests to me that it was boarded at a different time, the use of narrower boards might suggest later rather than earlier, and as that is the most exposed elevation these could be a replacement for the original weatherboarding.  As for what the cross bracing looked like underneath, two sets of panels each side of the chimney seems far too narrow, so one set each side will be tried.  I had wondered if the porch pre-dated the cladding, but it gradually dawned on me that the horizontal boarding visible on the inside would almost certainly have been that visible on the outside.  Similar TG boarded panels received weather boarding on the down platform shelter at Verwood, so it doesn't seem unreasonable that this is what it would have looked like.  

 

The Semley drawing had the fancy valancing drawn on it, so that's been added to my shortening to-do list!

 

Edited April 3, 2022 by Steve Smith
Due to the disaster recovery - the photos show the valancing that wasn't there at this point!

[Nick C]

I've sent you a PM with a few photos on

[Steve Smith]

Yesterday was spent on a bit of bug hunting and lots of little jobs.  The rear gained the option of cross bracing, the canopy side the option of cladding, the cabin a door and door stop, the porch steps got the correct shape round topped handrail.  The size of the porch was adjusted so that both sashes are the same size.  The roof and cladding was added to the porch.   Today I generated some rainwater goods ready to fit the downpipes as seen above.  Printing them in-situ eases assembly, but makes painting more difficult, but then the real thing is a devil to paint too.   The general arrangement follows the prototype, with minor adjustments.   The pipe across from the porch into the hopper should be skinnier, but I think the fatter pipe is more likely to survive printing.  There was no dink in the pipe where it meets the brick string, so they probably just knocked it through.  

The porch now has guttering all round, but I need to decide whether to print the pipe connecting the two with the porch, or the side of the cabin.  Printing those two parts separately will certainly help with the painting and probably printing.  As one of my pals is so keen to point out, there wasn't much paint left at Verwood in my timeframe, but if there had been it might have looked a bit like the above!

 

Edited April 3, 2022 by Steve Smith
Restore photos as best I could!

[Steve Smith]

After quite a lot of time looking at the 3D model from the same angle as various photos, and fully appreciating that there is no single correct answer for the roof pitch of LSWR Type 1 boxes, I had another go at the conumdrum.  Originally the narrower roof was drawn at 36°, and after a brief flirtation with 32°, I finally plumped for a slope of 34°.  In the process, I reworked the roof logic to make it even more parameter driven and make the arrangement of the rafters more to my liking.  Alresford box has timber bosses  where the hips meet the ridge, and there is no reason to think that those weren't the base of finials that probably didn't last all that long.   Another result of looking at those photos was to lift the whole chimney two brick courses, requiring a little more rework to slope the roof flashing to match the brick.   The result of all this nonsense is Verwood box as it might have looked just before the turn of the century.

 

The valancing that is often said to have been removed when the cladding was applied has been added, although Verwood is known to have had both valancing and cladding for a while.  The brickwork under the porch was also 'modern' and for the cross-braced version I have substituted the posts that would have supported it.  My valancing was lifted from a drawing of Semley, and fits Verwood without fudging in a sort of vote of confidence in the general arrangement of things.   The relationship of downpipes to valancing was interesting, and would have needed the unusually extended swan neck seen on the real thing for full clearance. 

 

Verwood was bang up against the station building where there would have been a convenient downpipe at the corner, and I am going to assume that Verwood box, being an add-on, would have been connected to that at the canopy end.

And lastly the box in its final form as so often photographed.  More staring at photos resulted in me shortening the newel posts.  The treads of the steps were on blocks, and not rebated in as I have shown them, evidence of more ad-hoc repairs.  Next job is to draw up a new set of sashes for laser etching and to make sure that there are grooves top and bottom to take them together with those little pegs along the front edges so that they drain properly.   The Verwood nameboard lettering was traced from a photo, but that doesn't mean that they're the correct scale yet, so to match photos I will be reducing the size of the lettering slightly.   Rendering these models was pretty quick until round objects started to appear, and I've had to decamp from my 2011 laptop to a more powerful gaming PC running Windows 10 and the latest version of OpenSCAD.

 

Edited July 8, 2022 by Steve Smith
Restore photos, and again...

[Steve Smith]

More staring at photos, and the roof still didn't seem quite right.  The ridge was definitely too short and this could only be because the box was narrower, or more likely the hip and main pitches were at slightly different angles.  Since the hip had one less row of slates, the latter seemed very likely.   I resumed my flirtation with 32° for front and back, and tried 33° for the hips.  I tried to come up with an accurate way of deriving the angle of the hips in plan using trig, but 33/32*45°, and its inverse was as good as I could get it, which just seems too easy!    I then amended the swan necks so that they cleared the valances.  As part of this, a slightly thinner and shallower downpipe run was fitted to take rainwater from the porch.  Next I tweaked the steps, slightly fattening the newel posts and moving them back on the concrete step to better match those photos.   Lastly I added the missing blocks to support the treads.

 
Attention then turned to the interior.  Some bits and pieces are already available, and I hope to make use of those, but I don't believe anyone has tackled a Stevens frame.  So using an end on view of the frame at Verwood, a drawing from an old railway magazine and a couple of measurements of the Stevens frame from Branksome now at Midsomer Norton, I have concocted something that looks a lot like a Stevens frame, but currently lacks the old style LSWR lever plates that all but lever 11 need.  Happily, the Verwood ones were in the same style as those at Alresford, so I know what I'm aiming for - thanks to Nick C!

 

In the Openscad program that generates the frame, three arrays are used to specify  colour, lever angle and number of stops for each lever.  The latter is to cater for lever 11 that had three positions and was used for switching out the box.   The number of levers can be varied, and cut down handles could be catered for easily enough, though Verwood didn't have any of those.    At around 4' long, the frame takes up much less room in the cabin than I expected. 

 

Securely fastened to what is clearly a convenient piece of MDF is a colour render.  The brown levers are for releasing the ground frame and switching out the box, and should probably be blue and brown, but that's further than I'm prepared to go for a mere render!

 

Edited April 3, 2022 by Steve Smith
Restore photos

[Nick C]

No problem - I also took a bunch of measurements during one quiet turn last year, so if you need any, let me know... I didn't measure the pitch of the roof though, I couldn't reach that!

[Steve Smith]

3 hours ago, Nick C said:

No problem - I also took a bunch of measurements during one quiet turn last year, so if you need any, let me know... I didn't measure the pitch of the roof though, I couldn't reach that!

Hi Nick,

Since I couldn't reconcile the overall measurements I made at MSN with the tiny drawing of a Stevens frame in Pryer's 'Southern Signals', there has been an element of guesswork over the exact size of the frame and the angle of the levers at their extremes of travel.  Confirmation of the distance between the centres of the bolts would be much appreciated, as would the lever angles if you have them.  The levers themselves are pretty sound as I had a good dimensioned drawing of those.   The angles for the levers that I used came from an end on view of the frame at Verwood, and they do seem a tad extreme. 

[Nick C]

Just checked my notes and I'm afraid I don't have either of those measurements, sorry! I can have a look next time I'm down there, but that won't be until April at the earliest.

[Steve Smith]

Hi Nick,

At my rate of progress, and heading into the gardening season, that should be absolutely fine!  There's a ground frame to knock up as well, although we don't know if it was a Stevens'. RSC or something else entirely.

Steve

[Steve Smith]

After the disaster recovery, most of the photos that showed all my mistakes and resulting improvements were lost.  I have restored as many as possible but have had to replace some interim photos with the latest versions which means that the text now talks about faults that aren't visible - which is confusing, but can't realistically be helped.  

 

As it doesn't seem to have featured before, I've added my 'Airfix' exploded signal box view:

Edited April 3, 2022 by Steve Smith

[Steve Smith]

Two days ago I replaced the photos, and today most of them seem to be missing again.   Not sure if I can be bothered to do it again. 

As a sometime Oracle DBA and UNIX SysAdmin, this is not how it is supposed to be!  😞

 

 

[SouthernRegionSteam]

7 minutes ago, Steve Smith said:

Two days ago I replaced the photos, and today most of them seem to be missing again.   Not sure if I can be bothered to do it again. 

See the red banner at the top of every page. Or this post here.
By the way, being my local line (well... if it still existed, that is!), I'm watching this thread with interest.
Whilst I haven't modelled Verwood, I did start a model of Breamore (being the closest station to me). That said, it's not a patch on your efforts, not least because I never finished it - although to be fair I built it about 2011 (gosh, was it really that long ago?); my skills have improved somewhat since then!
 

[Steve Smith]

Thanks for the kind words.  I didn't see the red banner, as I was at the bottom of the page, and I scrolled up to find missing photos on both threads I'd updated. 

 

That's an interesting thread about the outage, and given the setup, it was an accident waiting to happen.   Definitely getting a little twitchy about the whole thing, but it does look like my restored photos are all back again - phew!    

[Steve Smith]

A Quick update:  Test prints of the rear roof pitch, rear elevation (weatherboarded no valance) and chimney are on their way to me together with two 7mm Verwood (and Semley) Crane kits - one spoken for!

[RailWest]

On 20/02/2022 at 21:26, Steve Smith said:

 

 

 

In the Openscad program that generates the frame, three arrays are used to specify  colour, lever angle and number of stops for each lever.  The latter is to cater for lever 11 that had three positions and was used for switching out the box.   The number of levers can be varied, and cut down handles could be catered for easily enough, though Verwood didn't have any of those.    At around 4' long, the frame takes up much less room in the cabin than I expected. 

 

Securely fastened to what is clearly a convenient piece of MDF is a colour render.  The brown levers are for releasing the ground frame and switching out the box, and should probably be blue and brown, but that's further than I'm prepared to go for a mere render!

 

Certainly it is clear from an interior photo that lever 6 was two colours, so blue/brown probably is correct. The colour(s) of 11 are a bit more speculative, as it might have been classed as a 'King' lever or similar, although with only the bottom part visible in a photo it seems not to be 'striped'. One could speculate also whether 11 might have had a short handle - although it did not work any power equipment, neither did it do any mechanical work other than shifting bits of interlocking about, so no real physical effort was required to work it.

 

Edited May 1, 2022 by RailWest

[Steve Smith]

Over the last couple of days I've splashed some paint on my test prints to give an impression of what we have.  Before painting I could already see that the quality of the 3D printed parts was beyond anything I could have hoped for, and my somewhat rushed painting hasn't spoilt that impression! 

 

There are some issues with my CAD work - The ogee guttering is too delicate at 0.2mm, and there was a troublesome void in the fireplace surround causing the distortion that can be seen.   The brickwork printed extremely well, and the pointing with diluted paint looked fantastic until it dried, so needs more paint and less thinner next time, or perhaps just a second coat. 

 

The roof also came out well and looked better before I picked out some slates in a darker shade.  On the Mid-Hants the rafters and roof boards are white, but the tin of cream enamel paint was open, so I used that - not Humbrol's finest in that particular (new) tin.  The painting and preparation don't do it justice, but there is more than enough potential for me to revise my CADs and order a complete box.   

 

If all goes well, the painted test prints will be with me at ExpoEM on Sunday.  If anyone is interested, Muz if on the SWC stall, or the Sidmouth crew will be able to point me out.  

[Nick C]

18 hours ago, Steve Smith said:

If all goes well, the painted test prints will be with me at ExpoEM on Sunday.  If anyone is interested, Muz if on the SWC stall, or the Sidmouth crew will be able to point me out.  

I was going to say hello, but I think I just missed you - I asked Mike King and he said you'd been there 5 minutes earlier, but couldn't see you around...

[Steve Smith]

On 15/05/2022 at 16:44, Nick C said:

I was going to say hello, but I think I just missed you - I asked Mike King and he said you'd been there 5 minutes earlier, but couldn't see you around...

Sorry I missed you - I was there, right to the point when the layouts started crashing down.  It was a very enjoyable and useful exhibition (and expedition), with useful contacts made for laser cutting/engraving and 3D printing.  Quite a lot of people got to see the test prints, and will try to enhance the pointing of the brickwork ready for another go at photographing it.

[Blandford1969]

Daft question, would it be possible to purchase a frame ?