How Do I Train? (adapted from a talk)

[jtlan]

I've wanted to write this since last summer, when I picked out a model as a "demonstrator". I don't expect that all the material will fit in one post. This post covers part of my process for building scale models. I previously presented some of this material in a talk at Bricks By the Bay 2016, titled "How Do I Train?".

 

Introduction

I built Lego trains prior to heading off to college, but didn't take my bricks with me when I started school. I started building again when I returned. Seeing the high-quality work of the early train builders inspired me. In particular, Ben's works served as inspiration both before I left for college and after I returned. Like many builders, I base my models on real trains. I got started with my current building process when I wondered why my models didn't really look like the things they were baed on. Clearly, building a model while looking at references helps. But continually checking against known dimensions of the real thing will yield even better results.

 

Scale Models

The models I build now are scale models of real trains. A scale model is

"a proportional replica of a physical object"

(Wikipedia)

The original object the model is based on is called the "prototype". The model reproduces the features of the prototype at a smaller size and also maintains the correct positioning of those features relative to each other. The amount of reduction is called the scale of the model. For example, a 1/6 scale model of a 6-foot tall person would be 1 foot tall.

Here are two images from a pamphlet that illustrate the idea of "scale". This image shows the same plane modeled at different scales:

The planes have the same proportions as each other and the original plane, even though they are all different sizes.

This second photo shows models of different planes built at the same scale. 

As the planes are all scaled down from their prototypes by the same amount, the models accurately depict the difference in sizes between the real aircraft.

Widths Are a Distraction

Many train builders describe their models as 6-wide, 8-wide, etc, corresponding roughly to the width of the primary portion of the model. These are not scales. They are *sizes*.

Widths are NOT scales!

The width of a model is useful for explaining roughly how big it is, but the same width may reflect different scales depending on the size of the prototype. A Big Boy built at the same width as Stephenson's Rocket would be built at a smaller scale, because it is wider to begin with and has to fit in the same amount of space.

Conversely, building at a fixed scale can result in models of different widths, reflecting the difference in sizes of the prototypes.

 

Picking a Scale

The first instinct when deciding to build at a fixed scale is to try to build at "minifig" scale. That approach is doomed to failure, or at least inconsistency. Minifigs have very different proportions than humans:

A minifig is about twice as wide as a human the same height would be. Because of this fact, a minifig will seem either short or wide relative to a model of a real vehicle designed for real humans.

The scale I choose to build at is 15 inches per stud (381mm / stud). This works out to about 1:48 scale. At this scale a minifig represents someone about 6 feet (183cm) tall. American and most continental European rolling stock is about 8 studs wide; British rolling stock clocks in at 7 or 8, depending on the size of the prototype.

 

Constraints

Generally, I avoid modifying parts or using third-party parts in my models. I make an exception for wheels from Big Ben Bricks. Ben offers a variety of wheel sizes which are helpful when building steam locomotives. His small wheels are slightly thinner than the official Lego ones and have no webbing between the spokes. On the other hand, the official Lego wheels feature grooves traction bands, which is important for making powered locomotives (more on this later).

I also try to make sure that my models are able to run smoothly on standard Lego track. This means all arrangements of R40 curves and switches, or at least the ones I am likely to encounter at shows. Ideally the models can also handle some unevenness in the track.

Planning Process

Generally the first thing I do is pick a prototype to base my model on. Once I've done so, I locate references using search engines, Wikipedia, and more dedicated sites like RailPictures.net. If I find an interesting image I'll look at the site it comes from, which often turns up relevant information. Searching in other languages can yield additional information on foreign prototypes. I try to get photos of the prototype from a variety of angles, or at least pictures of other models of the prototype. Both of these can be tricky if the prototype is rare, exotic, or unique.

The most important thing is to find an engineering drawing or blueprint. These images show the prototype from a few different angles, with critical dimensions labeled. They are helpful for constructing accurate models.

Scaling

The next thing I do is scale the technical drawing. To do so, I choose a labeled length, convert it to inches, then scale by the chosen scale. For example:

The scaling equation yields the size of the chosen length in studs. I then overlay the drawing on Lego graph paper. The paper has vertical lines separated by the width of a brick and horizontal lines separated by the height of a plate. It's useful for building models that are primarily studs-up. The paper was previously available on Lego's website but has since disappeared. I've uploaded a pdf here.

Here's what the drawing looks like overlaid:

I usually colorize the drawing to make it stand out against the grid.

Adjusting Numbers and Selective Compression

From the earlier equation, you might remember that the distance between wheels scaled to 4.72 studs, which is not a whole number. In cases like this, I round to the nearest whole number (in this case 5). This process introduces some distortion in the model, but it's usually small and hard to detect.

Here's another example where the dimensions didn't quite work out:

Here, the distance between the center wheels and the two outside ones is ~5.5 studs. It would be inconvenient to place the middle wheel in that position if I wanted to implement working drive rods. For this model, I used a technique called selective compression. Selective compression is a modeling technique where certain features of the prototype may be reduced or omitted to reduce the size of the model. For example, a model-maker might omit some windows on a building while retaining their size and spacing, resulting in a smaller model. For the above model, I shortened the distance between the first and last driving axle by 1 stud:

This yielded a more usable spacing of 5 studs between axles.

Conclusion

I hope you've enjoyed this look into my planning process for train models. Let me know your thoughts. If there's interest, I'll continue this series with some posts on building and motorizing models.

Cheers!

[Phil B]

Absolutely interested in more posts like this! Keep it up!

[supertruper1988]

I really like this topic and I have been building in the same scale! 

 

[LEGO Train 12 Volts]

Very interesting topic ...perfectly made!

I use a similar technique and some illusions to fulfil my works

I hope to see other posts from you because I think you're a PRO!

[monai]

Very interesting, I was involved in exact model scaling and I wonder how do you scale at 1:48 if the Lego gauge is in 1:38 scale, without using non-Lego tracks. Let us forget the absurd minifig scale that, you said, is not proportionally correct; the railway modelling world is scaled by the gauge, and eventually I came to the conclusion that simplty not every prototype is available to Lego downscaling. The fact is that we are working with a "quantized" material (bricks) so we cannot model anything and the "selective compression" could be useful but not resolutive. Anyway your tecniques are very interesting and I'm very curious to see your brick-workss.

 

best regards

Sergio and Eros

[supertruper1988]

24 minutes ago, monai said:

Very interesting, I was involved in exact model scaling and I wonder how do you scale at 1:48 if the Lego gauge is in 1:38 scale, without using non-Lego tracks. Let us forget the absurd minifig scale that, you said, is not proportionally correct; the railway modelling world is scaled by the gauge, and eventually I came to the conclusion that simplty not every prototype is available to Lego downscaling. The fact is that we are working with a "quantized" material (bricks) so we cannot model anything and the "selective compression" could be useful but not resolutive. Anyway your tecniques are very interesting and I'm very curious to see your brick-workss.

 

best regards

Sergio and Eros

Part of the selective compression comes into play for the rail gauge. AFOL's know its wrong but its what we have to work with so we deal with it. if we used the rail gauge as a scale then trains would be 10 studs wide and trucks and cars would be 7-9 studs.

[Lego Dino 500]

Thanks for posting this! I'll have to try this method, I just eyeball stuff right now, which has a somewhat detrimental effect on the design of my models.

It looks like Sariel's scaling program could work for this.

[jtlan]

5 hours ago, monai said:

Very interesting, I was involved in exact model scaling and I wonder how do you scale at 1:48 if the Lego gauge is in 1:38 scale, without using non-Lego tracks.

 

4 hours ago, cameronmiller1988 said:

Part of the selective compression comes into play for the rail gauge. AFOL's know its wrong but its what we have to work with so we deal with it. if we used the rail gauge as a scale then trains would be 10 studs wide and trucks and cars would be 7-9 studs.

Pretty much. I accept that wheels, running gear, etc, may be a bit wide for the scale. I've only built standard gauge locomotives on the Lego track, with the exception of a narrow-gauge locomotive using the narrower tracks. Even then, the track is still too wide -- the prototype is 2-foot gauge, which would require the rails to be placed 1 stud apart! In practice I've found that the track distortion less obvious than might be expected because a viewer observes the model from a very different vantage point than they observe a real train.

 

1 hour ago, Lego Dino 500 said:

It looks like Sariel's scaling program could work for this.

I've seen Sariel's program, which is a good start for scaling the whole model (and also doesn't require starting another program). I find it helpful to have the full grid, which allows me to see the effect of scaling on the whole model at once. Having an actual image editor also enables me to do selective compression, take notes, and perform some other tricks I'll get to in the next article. I use Pixelmator to make my scaling drawings, because I have a Mac and wanted an inexpensive and basic image editor. Any image editor with layering capabilities should be sufficient.

[dr_spock]

Interesting. Has anyone tried scaling to LEGO Friends minidolls? 

 

[Duq]

That's pretty much how I work as well. I wrote a little tool a long time ago that draws a grid over an image. I use a scale somewhere between 1:40 and 1:45 for most models, based on wheel size mostly. That means I can't build some German steam engines because even the Big Ben XL wheels aren't big enough...

In model railroading shortening was more common in the past, when carriages for h0 (1:87) were reduced to 1:100 for length. These days they tend to be full length.

For my Umbau carriages I used narrower windows and I took out 2 windows either side. That reduced them to a length that's easier to manage around layouts.

[monai]

On 11/11/2016 at 7:39 PM, cameronmiller1988 said:

Part of the selective compression comes into play for the rail gauge. AFOL's know its wrong but its what we have to work with so we deal with it. if we used the rail gauge as a scale then trains would be 10 studs wide and trucks and cars would be 7-9 studs.

well it is what we are facing with . I mean, if the tracks system (including switches) is dimensioned to 38mm this is the right scale, other scales simply don't work, as our friend Tenderlok teaches us, the way to go to is to adopt other tracks system, if you need a 1:48 scale you should look for a 30mm gauge (O-scale has 32mm, acceptable). One thing is to play (like children do) other is to model in scale. Your technique is very interesting but I guess if the final products will be in scale.

Sergio

[ScotNick]

5 hours ago, monai said:

well it is what we are facing with . I mean, if the tracks system (including switches) is dimensioned to 38mm this is the right scale, other scales simply don't work, as our friend Tenderlok teaches us, the way to go to is to adopt other tracks system, if you need a 1:48 scale you should look for a 30mm gauge (O-scale has 32mm, acceptable). One thing is to play (like children do) other is to model in scale. Your technique is very interesting but I guess if the final products will be in scale.

Sergio

Well, you are not quite right about modelling to scale - there are compromises you have to make, especially when modelling (or should I better say - trying to model) with Lego.

If you compare the Model Railway scales H0 (used in Europe, US, Asia,...) and 00 (used in the UK) you'll see that they use the same track system / gauge but HO engines are scaled to 1:87 and 00 1:72.

If you put engines from both scales next to each other it looks like the following

66's - comparison between HO & OO by Brian Considine, on Flickr

66's - comparison between HO & OO by Brian Considine, on Flickr

They use the same gauge but a different scale - but I wouldn't want to say one is a toy and the other one is a model, they are both models.

 

What I want to say - using Lego's track doesn't necessarily mean you have to use a specific scale.

Cheers, ScotNick

[monai]

17 hours ago, ScotNick said:

If you compare the Model Railway scales H0 (used in Europe, US, Asia,...) and 00 (used in the UK) you'll see that they use the same track system / gauge but HO engines are scaled to 1:87 and 00 1:72.

---

They use the same gauge but a different scale - but I wouldn't want to say one is a toy and the other one is a model, they are both models.

What I want to say - using Lego's track doesn't necessarily mean you have to use a specific scale.

Cheers, ScotNick

Ok, first of all let's say that we are playing and everyone is free to do everything, said that, the percentual change in 1:87 (almost correct scale 1435:16=89) and 1:72 (uk wrong! scale) is smaller than your 1:38 to 1:48, so your difference will be more pronounced. When I speak about toys or models I'm referring to the usual LEGO train sets that are really too tight (6 studs) to "model" the reality. Since you have started the topic about exact scaling these are simply my concerns about it, probably your models will be fantastic. Take into account that a large model allows you to define better the small details of the lokos.

regards

Sergio

[Paperballpark]

My general view is that if you want perfect scale models, LEGO isn't what you should be using.

[Sven J]

An interesting discussion about different design philosophies.

Personally, I very much agree with Sergio’s (monai) view. While I’m actually using a scaling technique very similar to that described by jtlan, I’m fully aware that it’s impossible to build  100% accurate models with Lego bricks (I like the expression "quantized material“  - so we’re all quantum scientists here…).
But that’s just why I think that it’s all the more important to get at least the major proportions right if you want your model to look as realistic as possible; and, especially for steam locomotives, this means that the relation between wheel size, track gauge and locomotive width should be close to the prototype (as jtlan and Duq mentioned, overall length or wheelbase is not as critical). Otherwise, cylinders and valve gear tend to look too pronounced.

Nevertheless, I acknowledge that there are many fantastic and true models, explicitely not toys, using Lego track despite of being built to roughly 1:45 scale. It’s just a different approach to our common hobby, but can also yield great results.

So I think the best summary is to just quote Sergio again: "We are playing“ – and shouldn’t take things too seriously…

As for the H0/00 comparison, I’d like to point out that the seemingly odd proportions of 00 gauge are simply due to practical historical reasons (as described here). But track systems for accurate 1:76 scale do exist, although not for the mass market.

Best wishes,
Sven

Edited November 16, 2016 by Tenderlok

[Aaron]

On 11/11/2016 at 1:12 PM, monai said:

Very interesting, I was involved in exact model scaling and I wonder how do you scale at 1:48 if the Lego gauge is in 1:38 scale, without using non-Lego tracks.

I've always built in 1:38 scale (10-wide) and never had any issues.

[kurigan]

Just a note on an otherwise thorough presentation:. Minifigs do share proportions with humans, just human infants instead of human adults. It’s nice to see someone else considering the "fat or short" paradox thought. 

[monai]

6 hours ago, kurigan said:

Just a note on an otherwise thorough presentation:. Minifigs do share proportions with humans, just human infants instead of human adults. It’s nice to see someone else considering the "fat or short" paradox thought. 

well, very young babies indeed, they have no knees and a 6 or 7 years old boy don't share proportions with minifigs. Let's say they are studied to "move" us as puppies do..

The discussion has been proposed to investigate the possibilities to model in scale with LEGO, for me the answer is yes but only for selected prototypes, the main problem is actually the lack of choice in diameter wheels.

[jtlan]

5 hours ago, monai said:

The discussion has been proposed to investigate the possibilities to model in scale with LEGO, for me the answer is yes but only for selected prototypes, the main problem is actually the lack of choice in diameter wheels.

A particular set of constraints necessarily eliminates certain prototypes, but relaxing those constraints or choosing different ones may allow for those same prototype to be built. For example, the constraint "runs smoothly on R40 curves" eliminates most long-wheelbase 2-axle rolling stock, but if the builder chooses to relax that constraint (perhaps by using ME Models large-radius curves), those prototypes may be available to model again.

The "correct" choice for wheels is not always clear, as at most scales the flange on Lego train wheels is proportionally a lot larger than the flange on real wheels. At the scale I build at, the flange would be 6" (152.4mm) tall -- comically large compared to the real thing! Given that, is it "correct" to scale to the diameter of the hub of the wheel (which is the figure quoted by Big Ben Bricks)? Or is it "correct" to scale to the diameter at the flange? Compounding the issue is that the apparent size of the wheel might not correspond to either of those:

It varies from model to model, but generally I consider a wheel "close enough" if the scaled diameter lands somewhere between the hub and flange diameters.

 

[monai]

On 11/19/2016 at 7:39 PM, jtlan said:

It varies from model to model, but generally I consider a wheel "close enough" if the scaled diameter lands somewhere between the hub and flange diameters.

 

I agree with you about the comically large LEGO flanges, and therefore to consider it as the "real" diameter to take into account for scale modelling. Nevertheless also doing so you are facing with a lack of choices: the largest BBB wheel is 36mm which is (at 1:38) equal to a real 1368mm diameter wheel. Ok you have a lot of lokos to model but not the large beautiful steam engines everybody likes, for instance the "flying scotsman" which inspired the Emerald night (LNER Class A3 4472 with a 2033mm wheel diameter!). The Emerald is nice to play with and the new wheels are a succes, but the scale is simply ridiculous. Probably the real problem are the tracks, I guess a very large wheel simply doesn't work on standard lego tracks, not only for the curve radius but also for the profile of rails (they are not vignoles). Always speaking of "scale modelling", you should consider the "static" solution, which Tenderlok has so well investigated (see the main page of BBB for confirmation). 

[Sven J]

On 23.11.2016 at 0:20 PM, monai said:

On 19.11.2016 at 7:39 PM, jtlan said:

It varies from model to model, but generally I consider a wheel "close enough" if the scaled diameter lands somewhere between the hub and flange diameters.

I agree with you about the comically large LEGO flanges, and therefore to consider it as the "real" diameter to take into account for scale modelling.

That sounds reasonable at first – but in my opinion, this approach would mean that you can’t use both flanged and blind drivers on one model.

Another way to "gain“ a few centimetres is to claim that your model represents a locomotive with worn-down tires – at least according to former German regulations, maximum tire wear is 45mm, so wheels can actually be 90mm smaller than their nominal size.

By the way: Of course, LEGO flanges are quite high, approx. 3mm. But when you take into account that even H0 (1:87) wheels may have flanges up to 1,2mm according to NEM standard, they do not seem that much oversized in comparison to "real“ railway models.

Edited November 26, 2016 by Tenderlok

[monai]

On 11/26/2016 at 8:33 PM, Tenderlok said:

 

By the way: Of course, LEGO flanges are quite high, approx. 3mm. But when you take into account that even H0 (1:87) wheels may have flanges up to 1,2mm according to NEM standard, they do not seem that much oversized in comparison to "real“ railway models.

well, from the table the 1,2 mm is the maximum depth of flanges not the standard, LEGO are all 3mm tall, but, as I said, the main problem are the tracks, the 1,2mm H0 still runs on their tracks, 3mm LEGO are not running on "scaled" (1:38) tracks (code 200 Gauge1), as you well know.

Not to speak about the switches..

cheers

Sergio

[Sven J]

@monai Of course, you're absolutely right with regard to the rails (although conventional NEM-compatible H0 track is oversized by about 40%, too). What I was trying to express is that visually, the high flanges don't matter as much as one might expect.