Jump to content

Cautionary Tales


MikeOxon

936 views

 Share

Accuracy of Drawings

 

In an earlier post, I wrote: “I used the same method that I described in my previous post to extrude the saddle tank from a drawing – this time a pencil sketch by F.J.Roche, reproduced in the ‘Broadsheet’ article. This drawing was useful for the front elevation but I feel the drawing in Mike Sharman’s compilation by the Oakwood Press is more dependable for the side elevation.

 

Some recent correspondence within the Broad Gauge Society (BGS) e-group suggested that the drawings in the Mike Sharman compilation may not in fact be that accurate, I quote: “The Sharman's book drawings were transcribed from originals published in the Loco Magazine. The one in question here* was published in 1903. The transcribers varied greatly. The originals are believed to be accurate but ... this particular transcription was one of the least accurate.

* This quote refers to a drawing of a Dean 2-4-0 convertible of the 3501 class, not my engine, but it sowed seeds of doubt in my mind.

 

On looking more closely, I noticed for example, that the spokes of the bogie wheels on ‘Aurora’ were not placed accurately on the drawing, which showed up clearly when I designed my own wheel, using ‘Fusion 360’s Pattern command to produce nine equally spaced spokes.


451112533_Bogieoverdwg.jpg.788e04fef5f2c03658b5f53e135b10d5.jpg

3D-printed wheel laid over Drawing

 

It’s a small point but a warning not to believe the correctness of all the details on the drawing. I have now measured the wheel base and other key dimensions on my 3D-printed model and have been relieved to find that they are all accurate. I shall pay more attention to the accuracy of any drawings I use as a basis for extruding models in the future.

 

Front Elevation

 

Looking through my own small collection of drawings, I found three showing the front elevation of one of these 4-4-0ST engines. Two of these, by F.J.Roche and by Ian Beattie are said to be of ‘Lance’ (both drawings are from the BGS magazine ‘Broadsheet’ no.17), whereas the one by Alan Prior (in his book ‘19th Century Railway Drawings') is of ‘Corsair’. Putting these three alongside one another shows that there are many significant differences between them:

 

 

370838085_4-4-0STFrontElevs800x600.jpg.442e46e5537b77b78c57fa5f83800373.jpg

Front elevation drawings compared

 

Faced with discrepancies such as these, I turned to photographs and, in particular, the one of ‘Aurora’ that I showed in an earlier post The tank seen in this photo appears to have the more rounded profile shown in the Roche drawing above, although the sand pipes are not so arched. Of course, those early engines often showed a iot of individual variation, so the dictum to work from photographs as far as possible is very sound but can be difficult to apply, when photos were so far and few between.

 

Following ‘rules’ may not be best

 

Another interesting point appeared when I started to create the trial prints of the wheels that I designed for my model of ‘Aurora

 

I made the initial prints by laying the inside of the wheel flat on my printer bed. This meant that the widest part of the wheel, with the flange, lay on the bed so that there were no overhangs as the printing progressed, which is the ‘preferred’ method.

 

In practice, the flanges came out thinner than expected and were damaged when I removed them from the printer. This may be a result of using a printer with an unheated bed but it was another case where disobeying the rule book yielded a better result!

 

My later prints were made with the outer face of the wheel on the bed, so that the flanges actually overhung the main part of the wheel, as it printed. Nevertheless, the flanges printed cleanly and, by printing in this orientation, I could include sleeves extending from the backs of the wheels, to guide the pin-point axles and ensure the correct back-to-back measurement between the wheels.


1503007563_3D-printedwheelsets.jpg.55303d5551e369bbc2105071b7ba6483.jpg

3D-printed wheel-sets with integral axle sleeves

 

Each sleeve contains a clearance hole for the 2mm axle, while the wheel itself is an interference fit onto the axle. I assembled each wheel-set by dropping a pin-point axle into the sleeve and then tapping it gently home into the wheel with a light hammer, as shown below. Once one wheel had been attached, I turned the part over and tapped the axle into the other wheel, until the two spacing sleeves meet at the centre line.

 

 

711656333_Axlefitting.jpg.4f3c25ba7f1fee2712a23002158ef65f.jpg

Assembling a wheel-set

 

Once I had produced a set of wheels, I could place them under my model, so that it began to look like a real engine!  It has a purposeful look, well matched to its task of hauling important passenger trains over the South Devon banks. The famous Gooch singles may have stolen the limelight but it was these tank engines that maintained services across the more difficult routes of the South Western peninsular – they must be lauded for that capability

 

 

Aurora_4-4-0ST.jpg.df1d71a4a01a4eb0428a38a830133bc6.jpg

‘Aurora’ on her Wheels

 

A few years ago, I could not have contemplated making a model like this and I am pleased that, during its design and construction, I have gained an appreciation of the prototype’s remarkable qualities. No longer shall I call it an ‘ugly duckling’ :)

 

As usual, there’s a lot of finishing still to be done. One day, I must get together my collection of unfinished Broad Gauge locomotives and have a session of handrail fixing,-plus all those additional fittings and polished brass-work that make make them into beautiful swans.

 

Mike

Edited by MikeOxon

  • Like 4
  • Interesting/Thought-provoking 1
  • Craftsmanship/clever 10
 Share

16 Comments


Recommended Comments

  • RMweb Gold

It does look good, nothing particularly ugly about it.

 

I'm impressed that you managed to do the outside flanges on the wheels. Might they need some sanding etc to go through pointwork?

  • Like 1
  • Agree 1
Link to comment

I too have found anomalies when putting drawings into Fusion, which of course can measure down to an amazing number of decimal places! I think some of the issues are down to the thickness of the pen/pencil lines abs others due to the impossibility of the human hand matching CAD precision.  Then of course you can throw in parallax errors from photos and photos of drawings! My work in progress of a Hawthorn class saddle tank looks different to the drawing, but matches the measurements - all due to parallax issues when the drawing were photographed.  Mitigation? Measurements and pictures combined with drawings; and remember absolute fidelity is impossible no matter what sone might think. There is always a compromise!

Duncan 

  • Like 1
  • Informative/Useful 2
Link to comment
4 hours ago, Mikkel said:

It does look good, nothing particularly ugly about it.

 

I'm impressed that you managed to do the outside flanges on the wheels. Might they need some sanding etc to go through pointwork?

Thank you Mikkel - making the model has changed my mind about these engines and I now think they are rather handsome.

 

I'm still simply enjoying exploring what I can do with 3D modelling.  I haven't given any thought to actually running any of these models and they may well remain as 'display' items.  All the photos were taken without any significant cleaning up of the prints, apart from removing obvious loose strands and the like.  I think the wheels would need proper metal rims to make working models.

  • Like 2
Link to comment
2 hours ago, drduncan said:

I too have found anomalies when putting drawings into Fusion, .....

I agree with all the points you make, Duncan.   I always check horizontal and vertical dimensions separately and make any necessary adjustments.  The 'calibrate' function in 'Fusion 360' is very useful. 

 

I'm more concerned about actual drawing errors, where details are wrongly placed or wrongly sized, and so on.  There's a lot we shall never know about these old engines but I don't like to perpetrate errors, if I find them.

Link to comment

On the subject of inaccurate drawings, I found that the drawings I have of Firefly do not include the mini-splashers over the carrying wheels, although they can be seen in photographs.  I only realised after I built my model and found that true scale wheels would not fit.  I've now re-designed the footplate to include these splashers:

 

3D-assembly-new.jpg.514d603bc6272773d069cf67fff2def2.jpg

 

 

  • Like 1
Link to comment

I noticed that the trailing wheel splashers seems to missing on the GA of the 3521 0-4-2st but they can be seen in photos!

D

Edited by drduncan
  • Thanks 1
Link to comment

On the other hand, with such relatively poor levels of info out there (photos drawings etc) for BG at least it’s hard for us to be proved wrong!

D

  • Agree 1
Link to comment

I completely agree with everyone. I've been lucky with the GNR locos I've done that I have the Bird book on GNR locos, which contain profile drawings and some measurements, Groves, who also did a (much) more recent survey of GNR locos with no drawings, but a lot more measurements and detail changes, and on occasion, Isinglass drawings. The measurements are mostly consistent, so I tend to start by boxing in the big shapes using measurements only and superimposing those on the Bird or Isinglass drawing. Once the bigger blocks are filled in, I start on the details using measurements, if I have them for those parts, and then photos to fine tune/fill in gaps. As you say, it can be more of an art than a science. 

I'm currently looking at GNR L1 No. 125. I've only found 1 photo, and that doesn't indicate which of two possible smokebox fronts and doors it has, so I'll try to make informed guesses and using a photo of a different numbered loco as a guide for that area. Although the loco below is right per the drawings and initial measurements, it's wrong per the photo I have (for a start, the cab on this one was increased in height by 10"), so will have to be changed.

 

581284321_R1progress.jpg.00d4790bd11b3641a399014a962af6ff.jpg

 

cheers

 

Jason

 

 

Edited by JCL
  • Like 2
Link to comment
  • RMweb Gold

I suspect that this is one of the photos of this class that you have - but just in case you haven't....

 

Link to comment
4 hours ago, JCL said:

 ...I tend to start by boxing in the big shapes using measurements only and superimposing those on the Bird or Isinglass drawing. Once the bigger blocks are filled in, I start on the details using measurements, if I have them for those parts, and then photos to fine tune/fill in gaps. As you say, it can be more of an art than a science...

i suppose the difference is that I aim to create the main box and cylindrical shapes by extruding directly from drawings. 

 

It's not too difficult to make scaling adjustments in the computer, to compensate for errors in the drawings.  The major difficulties arise when the drawings are incorrect in the placement or dimensions of specific features. 

 

in the end it's a personal matter how far we wish to take things but no-one wants to make things incorrectly, if better information is out there somewhere.

  • Like 1
Link to comment
2 hours ago, phil_sutters said:

I suspect that this is one of the photos of this class that you have - but just in case you haven't....

 

Nothing beats a good photo ... when you have one! 

 

Those of us attempting to model the earlier part of the 19th century very rarely have that opportunity!  In addition, many of the drawings we have were created long after the event, so are often extrapolations from limited information.  In such cases, it is necessary to find as many independent sources of information as one can and after that it's down to personal judgements.

Link to comment
9 hours ago, phil_sutters said:

I suspect that this is one of the photos of this class that you have - but just in case you haven't....

Hi Phil, There aren't many photos around (although I've probably for a treasure trove compared to Mike), but I've certainly not seen that one.

Link to comment
6 hours ago, MikeOxon said:

Nothing beats a good photo ... when you have one! 

 

Those of us attempting to model the earlier part of the 19th century very rarely have that opportunity!  In addition, many of the drawings we have were created long after the event, so are often extrapolations from limited information.  In such cases, it is necessary to find as many independent sources of information as one can and after that it's down to personal judgements.

Certainly Mike, it's definitely down to personal preference and the type of information you have available.

Link to comment

Drawing accuracy... Here's an example of how things can go wrong... This drawing of a 94 is from CJ Freezer's "Locomotives in Outline", and presumably the 'Railway Modeler' before that. In fact I recall it from RM as being the first thing I ever attempted to scratchbuild. (got stuck on the boiler). anyway, nice enough drawing. 

060-9400-cjf.jpg.cfab82140f903021eb36e5c368767fb9.jpg

But lets look at my 94 drawing I did for my book:

060-9400-jc.jpg.4747811f1923f212974d8d18dd231de4.jpg

Spot the difference?
What seems to have happened is that CJF somehow got muddled between the 94xx and the 57xx, and his 94xx drawing has the same rear overhang as the 57xx, not the extra length of the 94. The bunker is the right size, but he shortened the cab to make it all fit. I imagine he worked from the front back to the front of the cab, and from the rear to the back of the cab and the doorway. Its all a lot easier when you can just electronically lay one over the other...
Incidentally, don't go thinking I believe my drawing to be perfect: Its based mainly on the GWR weight diagram, which is far from a reliable source, so I'm not claiming full accuracy. I'm not quite sure about the height of the firebox, and I think I had better revisit the brake pull rods. Like models, electronic drawings are never finished...

Edited by JimC
Link to comment
2 hours ago, JimC said:

Drawing accuracy... Here's an example of how things can go wrong...

... and that's what I would call a 'recent' engine!  It's not just small-scale modellers that go astray.  There are some substantial differences between the Firefly replica at Didcot and photographs of the original Firefly.  I know compromises have to be made if a working engine is to meet modern standards but the firebox cladding on the replica is almost wholly cosmetic, with the true firebox hidden inside.  Yet the replica has a 'Haycock' style firebox where original photos show the 'Gothic' type, with arched sides.  Why?

 

974022256_Fire-flyorigvsreplica.jpg.bf71f8833e4664988c338e388e0fbf52.jpg

  • Informative/Useful 1
Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...