(2014-03-19) Small Lexicon of Rail Jargon
North-American, British and French terminology.
The evolution of rail transport in the New World was largely independent from
its counterpart in the British Isles and the rest of Europe.
Some of the resulting differences in terminology are summarized by the following table:
(2014-03-16) A short history of the standard scales for model trains
The model railroad scales are best expressed in millimeters to the foot.
This strange convention involving two competing systems
of units came naturally to early train
modelists (in Britain) who started using
readily available metric rulers to build models of prototypes measured in feet...
The nominal scale of a minature train applies to all parts of the vehicles,
with the possible exception of the wheels and the paramount track gauge.
The reference to the foot as the prototypical length is sometimes omitted
and we may talk about the 7 mm scale (O scale, 1:43.5)
4 mm scale (OO scale, 1:76.2) 3.5 mm scale
Some such scales have branched out (the O-scale became
1:48 in the US and 1:45 in Germany) or they have been superseded.
For example, the 2 mm scale (1:152.4)
has been replaced by the very popular N-scale (1:160)
of exactly 1.905 mm to the foot.
To translate a millimeter specification into a scaling factor, recall simply that a
foot is exactly 304.8 mm.
This gives the following
for the HO (3.5 mm) and OO (4 mm) scales,
304.8 / 3.5 = 87.0857142
304.8 / 4 = 76.2
The HO scaling factor is 87.0857... (often rounded to 87.1 or 87).
The OO scaling factor is exactly 76.2 (often rounded to 76).
Upon standardization, it was decided that both scales would use the tracks
that are technically accurate for the HO scale only (they are thus 12.5%
undersized for OO models of actual trains).
The 8/7 ratio between the two linear scales means that the OO scale corresponds
to volumes that are larger than HO volumes by the cube of that ratio,
which translate into a 49.3% increase in bulk.
Increasingly, the HO scale is simply referred to as 1:87
and models are produced to this precise specification, which is
virtually indistinguishable from the original definition of 3.5 mm to the foot
(the difference is 0.1%).
Next up are the S and O scales:
The S scaling factor is exactly 64 (3/16'' to the foot).
The O scaling factor is exactly 48 (¼'' to the foot) in the US.
In the UK and in France however, the O scale is still defined as 7 mm to the foot, which is
equivalent to a scaling factor of about 43.543 (usually rounded to 43.5).
This "7 mm scale" is exactly twice the HO scale.
An intermediate ratio of 45 is used in Germany and Russia.
In Japan and Taiwan, a specific model gauge of 24 mm
is used to represent the Kyoki narrow gauge of 3.5 ft
(Cape gauge) at a nominal scale of exactly 1:44.45.
This model type is denoted Oj in the orient.
In the west, where 24 mm model track is virtually unheard of,
the same effect can be achieved with standard S-gauge track
at the American O-scale (1:48) since:
Historically, the O scale (7 mm to the foot) predated the HO scale.
This explains the latter acronym (it used to be called "half-zero").
The O scale can be denoted either by the letter "O" or the digit "0" (zero).
Likewise HO (two letters) and H0 (trailing digit) are synonyms. The arcana continues:
The scaling factor for "#1" is 32 (3/8'' to the foot).
The H scaling factor is 24 (½'' to the foot).
The G scaling factor is 22.5 (16/30'' to the foot).
Minor discrepancies in the scale of model trains have little conseqences,
as long as they're precisely engineered to operate on the same tracks.
Early on, manufacturers had to settle on a limited number of track gauges.
This process parallels the gauge standardization which had occurred earlier
in rail transportation, driven by the need for interoperabilty of rolling stock,
as summarized in the next section.
(2014-03-17) Prototypical railway gauges, then and now.
broad gauge of 7' ¼''
(GWR) lived from 1833 to 1892.
60% of the lines worldwide are in the standard gauge of 4 ft 8½ in.
By definition, the gauge of a railroad track is the inner
spacing between the two rails (normally, the term gauge
denotes the distance between the rails but it's sometimes used to refer to the space
between the rails, sometimes called the 4-foot way
in British railway jargon).
In metric terms, the standard gauge
is exactly 1435.1 m (often described as 1435 mm,
which messes up the exact arithmetic without any benefit at all).
On 2014-03-18, I took my late grandfather's vernier
calipers to an abandonned (standard)
branch line in Normandy
and found the width of those French rails to be about 56.5 mm
Thus, the center-to-center distance between the rails is 1.4916 m.
and the total width of the track is about 1548 mm
(between the outer edges of the two rails).
I can't explain the numerical
coincidence that gauge width is 56.5'' while rail thickness is 56.5 mm.
For this particular piece of French track, at least, the thickness-to-gauge ratio is thus
1 mm/in = 1:25.4.
The second most common gauge today is the Russian gauge
which was originally specified as 5 ft (1524 mm).
Finland still uses that original definition but the Russian railways have adopted
a rounded metric definition of 152 cm (1520 mm).
For regular traffic, both definitions are compatible but high-speed trains have
On 12 December 2010, the
Allegro high-speed train
was inaugurated between Helsinki (Finland) and St. Petersburg (Russia) with Karelian Trains
(Class Sm6) of the Pendolino family manufactured by Alstom.
It's actually built for a nominal gauge of 1522 mm.
So are the new high-speed tracks compatible with the Russian gauge.
It's thus best to consider 1522 mm to be the one
and only modern definition of the Russian gauge (existing tracks are
well within manufacturing tolerances of this nominal definition).
Many narrow gauges
are primarily used for short hauls in industrial settings.
Some of them are much more widely used. Most notably:
The 3-foot narrow gauge (914.4 mm) in North America.
The metric narrow gauge (1000 mm) in Europe.
The Kyoki or
of 3 ft 6 in (1066.8 mm) in Japan.
The kyoki gauge of 1066.8 mm is dominant in Japan, except for its
Shinkansen high-speed lines
which are in standard gauge (1435.1 mm).
Australia also uses the Cape gauge of 1066.8 mm on most of its lines.
Unlike Japan and despite the lesser stability of such narrow track,
they're providing high-speed service on that same gauge.
The Tilt train of
is the fastest train in Autralia and the fastest train in the world on narrow gauge.
Yet, its record speed of 210 km/h
is far from what can be achieved on standard gauge (the current record
was achieved on 3rd April 2007 by a five-car TGV "V150" double-decker, specially prepared
for speeds beyond 150 m/s, or 540 km/h).
Yet, it's unlikely that the need for speed will ever resurrect Brunel's
broad gauge (2140 mm; slightly more than 7 ft) last used by
the GWR in 1892.
(The rare suffix "b7" is used by modelers to denote that gauge.)
(2014-03-17) Model gauges
The 16.5 mm model gauge is used by HO, OO, On30 or Gn15 models.
At the nominal HO scale of 3.5 mm per foot, standard gauge (1435.1 mm) would be
16.479666...mm, which is normally rounded to 16.48 mm or 16.5 mm.
That's the standard track gauge for both HO and OO model trains.
(1 ft) / (3.5 mm) = 3048 / 35 =
is the HO scaling factor.
1435.1 mm / (3048 / 35) =
is the HO standard gauge.
At the OO scale of 4 mm to the foot, standard gauge accurately corresponds
to the so-called EM-gauge of 18.333...mm, which is part of the
(P4) modeling standards at the 1/76.2 scale
(pioneered by Joe Brook Smith and Malcolm Cross in July 1964).
At HO scale, the 3-ft narrow gauge is exactly 10.5 mm,
which corresponds to the HOn3 gauge, commercially available
since 2010 from
(a division of Soundtraxx created in 2004).
The popular N-gauge is only 9 mm;
is advocating the use of this for narrow-track modeling (minework etc.)
at the OO scale (1:76.2) next to standard OO track
(16.5 mm) and in sharp contrast with it.
Dubbed "OO9", that use of a 9 mm gauge would correspond to the
unusual gauge of exactly 27 inches.
That rare Welsh gauge of
2 ft 3 in
was used in a slate quarry line opened in 1866 which is preserved since 1951 in the
which still runs the original #2 Dolgoch
locomotive and the engines
#3 (Sir Haydn) and
#4 (Edward Thomas)
rescued from the nearby
(1859-1948) which once inaugurated 27'' tracks with horse-drawn wagons.
The Corris Railway reopened to passengers in 2002 and they are now constructing another 27-inch gauge locomotive
based on the old design.
Besides Talyllyn and Corris,
only one other extant 27''
railway has been reported, namely the "funicular" of the
Yorkshire Mining Museum which opened in 1988 at the site of
adjacent Hope Shaft in Overton, near Wakefield
(that museum has been known as the "National Coal-Mining Museum for England",
abbreviated NCM or NCMME, since 1995).
narrow-gauge enthusiast had only one 30'' railway to report after his recent visit to
the site (either that or he misjudged the gauge by 3 inches).
The practical minimum gauge for useful
rail transport is around 15 inches (381 mm).
Anything below that is considered a miniature line.
The number 15 evokes very narrow gauges so strongly that the acronym
Gn15 (G-scaled very narrow
gauge) was coined as a generic term to describe their activity by the community of modelists who
run garden-sized model trains (at typical scales between 1:48 and 1:20) on HO tracks.
(2014-03-17) A Synthetic View of All Types of Model Trains
A type of model trains is the combination of a
scale and a gauge.
Although modelers might prefer to have tracks and trains built to the exact same scale,
they often settle for widely available miniature tracks whose actual gauge
(the inner space between rails) is only an approximation
of the scaled down dimension of the actual tracks
used by the ptototypes they are modeling at their chosen scale.
More than 80 different types of miniature trains are sufficiently
well-defined to have a recognizable standard designation
(not all of those are supported commercially or by associations).
To each such type corresponds a unique gauge and a unique scale.
Thus, we may talk unambiguously about the HO gauge
(nearly 16.5 mm between the rails)
or the HO scale (3 mm/ft = 1:87.1).
Likewise, the HOn3 gauge is 10.5 mm between rails.
A given gauge can be shared between many model types
(which can operate on the same track layout if their loading gauge
permits, although the scenery may be out of scale).
Conversely, models of the same scale can represent a consistent picture of
reality with vehicles of different gauges operating on different tracks in a single layout, if needed.
The big picture is summarized by the table below (a Numericana exclusive).
In our table, the model types in italics or between parentheses
have little or no support, for a variety of reasons.
For example, there are no prototypes with a gauge of 4' or 18'',
which makes types ending with n4 or n18 utterly useless for modeling purposes.
Bold numbers, for scales or model gauges,
indicate values that are exact by definition.
Likewise, bold model types are those whose nominal gauges
are perfectly true to scale.
For readability, some groups of prototypical gauges have been singled out with the
color-coding described below
(i.e., standard ,
Types of model trains (some extremely rare ones appear in italics).
Other gauges not included in the above:
7 mm gauge: HOn2 represents 2-foot narrow gauge at HO scale.
The 18.2 mm (EM) and 18.83 mm (P4) gauges are supported by the EMGS (see below).
12.7 mm gauge: On2
represents 2-ft gauge at the scale of 1:48 in the US (Maine "two footers").
24 mm gauge: In Japan, Oj represents Kyoki gauge (3.5 ft) at 1:44.45.
Would also do nicely for HOb7.
1:152.4 OOO (2 mm) superseded by N (1:160) survives in Japan as N (1:150).
1:96 E-scale (Eighth inch) was superseded by HO. It's also very close to TT3 (1:101.6).
The suffix "f" (Feldbahn)
can be replaced by "i" (industrial). HOj is called HOm in Europe.
The scale corresponding to the Stephenson gauge for HOn3 track
(10.5 mm) is called the track scale for the model gauge of 10.5 mm:
1435.1 / 10.5 = 136.6761904...
Normally, the track scale of a widely available model gauge will be used by many professional
and amateur modelers. The 1:137 is an exeption.
It's apparently all but unused by the modeling community.
It happens to be nearly one tenth of SE scale
(Seven Eighth = 7/8'' to the foot):
12 / (7/8) = 96 / 7 = 13.714285...
(**) The interesting combination of the British OO scale (4 mm to the foot, or 1:76.2)
with On3 gauge (19.05 mm = ¾'' ) would
result in an "OO19" modeling standard where the gauge is merely 1.15% out of scale.
This is currently unsupported as a whole.
Instead, the P4 standard
(OO scale with a prototypically correct standard gauge of
and the EM standard (OO scale with a -3.4% out-of-scale
18.2 mm gauge,
formerly dubbed Eighteen Millimeter gauge)
are both actively supported by the British
EM Gauge Society (EMGS)
whose members re-wheel OO models to run on custom-built tracks
using their preferred gauge
(instead of HO gauge or On3 gauge).
They use 28 mm track (28.084 mm) when modeling
Brunel's old broad gauge (2140 mm).
Color-coding and prototypical gauge suffixes :
is for entries which are used to model
the standard Stephenson gauge of 56½ in. (1435.1 mm).
When set in bold type, the correspondance is an exact one
which can be used to work out exactly its nominal gauge or its nominal scale
whenever the table provides only an approximation while giving the true value
(in bold) for the other quantity
(the gauge is 1435.1 multiplied by the scale and the scale is the gauge divided by 1435.1).
For example, the standard gauge at S-scale (1:64) is actually:
A more delicate case is the gauge shared by Sn3 (1:64) and the 3 mm scale (1:101.6) which is
boldly defined as 14.2 mm. This would entail prototypical
gauges respectively equal to:
64 (14.2 mm) = 908.8 mm
101.6 (14.2 mm) = 1442.72 mm
The relative error with respect to the target gauges
of 914.4 mm and 1435.1 mm are only -0.61% and +0.53%, respectively.
such tiny discrepancies are sufficient not to award bold listings to either type.
To put it another way, the perfect nominal gauges ought to be:
1435.1 / 101.6 = 14.125 mm
914.4 / 64 = 14.2875 mm
Whoever chose a 14.2 mm gauge (commonly called "14 mm track")
for both cases made a brilliant decision.
Neither type of models is noticeably out of scale!
As a bonus to British modelers, an OO scale model on
that gauge will represent the Japanese narrow gauge of 3½ ft
just 1.4% out of scale (OOj or OOn3½).
Likewise, the representation of 2-ft narrow gauge at
1:43.5 scale (O14) using the same track is just 1.3% out of scale.
signals three common narrow gauges of similar sizes:
American 3 ft
gauge (1 yd = 914.4 mm) denoted by an n3 suffix.
European meter gauge (1000 mm) indicated by an m suffix.
Japanese kyoki of 3'6'' (1066.8 mm).
Suffix is j or n3.5 or n3½.
When such an entry appears in bold, (which occurs for HOn3, OOn3, On3 and Fn3)
it corresponds exactly to the relevant prototypical gauge.
Thus, Fn3 entails the following exact scaling ratio:
914.4 / 45 = 20.32
Incidentally, the table allows you to retrieve indirectly the fact that HO scale is exactly
3.5 mm to the foot (HINT: start with the fact that
HOn3 gauge is exactly 10.5 mm). Then, you can obtain the true nominal HO gauge:
1435.1 / (304.8 / 3.5) = 16.47916 mm
(usually called "16.5 mm")
is for extremely narrow gauges approaching the practical
minimum of 15 inches (381 mm).
The absolute minimum for modeling purposes is 14 inches
(e.g., 45 mm track at 1:8 or 22.42 mm at 1:16).
At the opposite end of the spectrum are a few broad gauges
which can be fairly well represented at popular scales using commercially available tracks
(the Russian gauge, which is 6% above standard gauge, isn't one of them).
Since most modelists prefer compact layouts for a given size of locomotives,
the broad gauges are not nearly as popular as standard or narrow ones
(they're identified by a "b" infix, in the upper triangle of the above table).
The Irish gauge of 5 ft 3 in (i.e., 63''
or 1600.2 mm) is represented
perfectly at a 2 mm scale (1:152.3) on standard 10.5 mm track (Nb63).
The Indian gauge of 5 ft 6 in (i.e., 66'' or 1676.4 mm)
can be well represented at various scale on available model tracks.
Including Nb66 which is only 0.2% out of scale at 1:160 scale on 10.5 mm track.
Among larger scales, "F scale" is favored by those who are concerned
with matching the scale and gauge of a model train to those of a prototype.
In many cases, the dominant factor for outdoor train modelling is the gauge
of the tracks, which is rarely changed on a given property,
because of the investment involved.
Manufacturers will accomadate the installed tracks.
"E-scale" is what one manufacturer (RMI Railworks of Fresno, CA) calls
their preferred scale (1:3.2) for large ridable models.
They are heralding this scale as "Estate", "Exceptional", or "Extreme"
(it's exactly 30 times larger than the obsolete "E-scale" of 1:96,
or one eighth of an inch per foot, which has been superseded by HO).
They offer rolling stock for a variety of gauges at that scale
but only up to 12 inches, which still corresponds to a prototypical
narrow gauge of 38.4''. To match the standard gauge
of 56.5'', a model at scale would have to use a gauge
nearly 50% larger (448.47 mm) which isn't supported by RMI.
(2014-03-17) The main sequence of commercial scales and gauges
They are roughly geometric progressions with inverse common ratios.
The system below is based on the numbers 32 and 45 whose
product (1440) is close to the standard gauge expressed
in millimeters (1435.1) and whose ratio (1.40625)
is close to the square root of 2 (i.e., 1.41421...).
A regular approximation to the actual system of model train gauges
This regular system is clearly not very different from what's actually used in the industry and it shows
the result of natural market selection mascarading as engineering design over a century or so:
From one gauge to the next, both the gauge and the scale are multiplied
by the square root of 2 (for two steps, the gauge and scale are thus doubled).
We're simply dealing with a straightforward geometric progression here!
Never mind the lack of regularity of the historical naming scheme:
ZZ is more exreme than Z (1440 mm / 300 = 4.8 mm).
Z (last letter) was heralded as the ultimate in miniaturization in 1972.
N is for Nine milimeter gauge (1440 mm / 160 = 9 mm) superseding
OOO (triple O) considered the ultimate in miniaturization in 1964.
TT is for Table Top (its 12 mm gauge at scale 1:120 is 1440 mm).
HO was Half the original O scale
(now S7: 1:43.5 on 33 mm track).
O scale was called "zero" because larger scales had positive numbers.
H scale stands for Half-inch (namely, 1:24 scale).
F scale stands for Fifteen-millimeter scale (exactly 1:20.32).
(2014-03-17) The height of the rails in miniature tracks.
The code is the rail height in thousands of an inch
(Code 100 = 0.1'' ).
For HO scale:
C110 = 2.8 mm (old Jouef track).
C100 = 2.54 mm (common track).
C83 = 2.10 mm (American scale, Kato).
C75 = 1.90 mm (British fine scale).
C80 and C55 are commonly used for N-gauge.
(2014-03-23) Premium miniature rails are made out of maillechort.
This alloy is also misleadingly called nickel-silver (it contains no silver).
Maillechort is actually an
alloy of copper, zinc and nickel.
It is to brass (i.e., copper-zinc alloy) what stainless steel is to iron, as the addition of nickel improves
the resistance to corrosion (lower-grade miniature rails are also available which are made from stainless steel).
Cu62Ni18Zn20 (NS106) is the most popular variant.
It's sold as Awa®, Nickeloid®, Silmet® or Spedex®.
That's the "nickel-silver" used for premium miniature railroad track (stainless steel is considered lower grade).
The NS106 alloy was selected for its durability and its mechanical properties, not for its
conductivity (it's 16 times worse than copper but still twice as good as steel).
However, the lack of surface oxidation of nickel-silver rails tend to reduce or eliminate sparking.
One potential advantage of the poor conductivity of NS106
is that miniature rails could be welded electrically
flash butt welding)
like real rails are, to form a strong continuous welded rail
or "ribbon rail".
Modelers have shunned that technique in favor of soldering
(which is adequate for electrical contact).
There's no silver at all in maillechort,
but its early use as a substrate for silver-plated silverware had lead to several
misleading commercial names, including "nickel-silver", "new silver" and "German silver".
Variants of the alloy are may be given several other names in different applications, including
argentan, alpacca, ruolz and EPNS (electro-plated nickel-silver).
Adding nickel to a copper alloy (brass) decreases its conductivity.
Maillechort was perfected in 1819 by
Maillet and Chorier, two Frenchmen from
Lyon who combined their own surnames to name the invention.
The one-euro coin consists of a rim of yellow maillechort surrounding a white
center of cupronickel on a nickel core. For the two-euro coin, it's the other way around
(the yellow maillechort is at the center).
Since 1728, maillechort (nickel silver) has been a popular choice for the
manufacture of musical instruments,
although it's now less prestigious than it once was.
Since 1970 or so, the top instrument makers have been returning to solid silver or gold
exclusively for the metal parts of their best bows.
The spacing of railroad ties (railway sleepers).
The French called travelage the number of cross ties
per unit of track length; the French standard calls for 1666 ties per kilometer.
(2014-04-16) Scaling Light
(cd for headlights, lm for car interiors)
at a given scaling factor s (e.g., s = 87 for HO).
In a finely scaled model, lighting should be properly scaled as well.
The basic physics of scaling light is very simple but it's unfortunately all but
ignored by most modelers, which may lead to gross misrepresentations...
In the US, the Federal Railroad Administration (FRA) sets
(we've corrected their utter disregard for the plural form of "candelas").
The FRA requires that the locomotive headlight (steady burn) used for road service have a luminous intensity
of at least 200 000 candelas.
The headlight light focus angle in the horizontal plane in relation to the centerline of the locomotive
must illuminate the track so that the locomotive engineer can identify moving or stationary objects or
conditions at a distance of 244 m (800 ft) in front and ahead of the locomotive.
The reduced luminous intensity (60 000 cd) and distance requirements (91.5 m [300 ft])
for railroad yard headlight operation is required to reduce excessive glare for railroad employees.
On Union Pacific locomotives, the headlights are 200 watts each and the ditch lights are 350 watts each.
The candela rating depends on the efficiency of the conversion from radiant to luminous power
(watts to lumens) and the focusing of light by the headlight optics (lumens to candelas).
On the other hand,
the lighting of passenger cars comes from nondirectional sources
(ordinary lightbulbs) whose total luminous power is measured in
To model such a light source, you must determine the lumen rating L
for the prototype and work out a scaled equivalent with the
technology you're planning to use in the model
(LED or incandescent light bulb).
The following examples might be helpful for guessing that data:
60 W incandescent light bulb
At a distance d, the luminous flux received by the retina of the observer is inversely
proportional to the square of d. When looking at the corresponding light source
on the model, we should have the impression that its distance is s.d.
Therefore, its lumen rating should be :
L / s2 = L / 7600 for an HO model
(L/5800 for OO, L/4000 for S, L/2000 for O, etc.)
To provide properly scaled lighting inside a passenger car, you should first
determine the total lumen rating inside the prototype.
Divide that by the above reduction factor and find a way to match that rating with
whatever lighting technology you choose (LED is probably best,
especially if you use the newer white LEDs without the blue tint of the previous generation).
I just bought (on e-Bay, for $20) two vintage trois-pattes
SNCF passenger cars by Fleischmann ("probably" #U371 1442)
with electrical contacts on the axles, but no working lights.
A perfect opportunity to do the job right...
One of my favorite electronic components is the LM317 voltage regulator which comes in
3-pin packages similar to transistors. Standard devices can supply up to 1.5 A
but Fairchild makes a low-power version
in a small TO92 package (the LM317LZ) which can deliver up to 100 mA.
Mouser sells the LM317LZ
for 51¢ a piece, $2.98/10, $14.10/100 or $94/1000.
Bridge rectifiers are
41¢, $3.69/10, $33.90/100, etc.
(2014-03-16) Layout Scenicing & Scenery Scales
HO and OO model trains may share the same layout (at different times).
Permanent layouts with detailed scenery are an essential part of the hobby.
entered the model train field by providing injection molded plastic models of buildings,
under the brand Plasticville® U.S.A.
Running HO and OO models at the same time is not recommended at all
because of the blatant scale difference when the trains are side-by-side.
However, the same layout could accomodate both HO and OO model trains
at different times. If that's the intention,
then it's best to minimize the discrepancy between the trains and the
scenery by choosing for the latter an intermediate scale
(sometimes dubbed HO/OO) equal
to the geometric mean of the HO and OO scale, namely:
304.8 / sqrt (3.5 x 4) = 81.46...
The resulting 7% scale mismatch so entailed is hardly noticeable.
However, because HO is so dominant, precise scenery is usually designed at the 1:87 scale,
which is 12.5% undersized for OO models.
It's also possible to use OO figurines in the foreground and HO decorations in the
background to create the illusion of a greater depth of field
The eye is especially sensitive to the size of human figurines;
if background figurines are smaller, distances appear larger than they are.
(2014-04-16) Mirrors and Shadows
Never cast a shadow on a flat backdrop. Magic of mirrors and lighting.
The backdrop is a key element of a scenery.
It represent scenery elements so distant that they can be painted on a
relatively nearby wall while preserving the illusion of true perspective.
That illusion of perspective, however, is immediately and utterly desroyed
if a scenery element from the foreground (a tree or a building, say)
is allowed to cast a shadow on the wall!
This is the single most widespread mistake among modelers
and most of them are not even bothered by it on the final layout
wherever it happens. Neither are the spectators, because
they don't believe in the illusion in the first place
(they see the room and the celling and know that the sky is not real on the first place).
However, when you make a video of the layout, an otherwise perfect illusion can be
completely destroyed by such shadows which are then seen
for what they truly are: mistakes.
Viewing your railroad World through a window...
For a confined layout which must be located along a "main" wall, you may want to simulate
a window frame between two long horizontal pieces of wood
(vertically aligned and decorated with the same trim).
The lower one should extend slightly above the highest piece of
terrain occuring at the edge of the bench.
The upper one will help create the illusion of the layout's "sky"
beyond the window frame so created.
This upper frame can be extended into a narrow shelf to accomodate
lighting fixtures for the layout and hide the "end" of its fake
sky (a bluish curved surface starting vertically at track level and ending
somewhat horizontally above that shelf, hidden from the eyes of little children
at bench level).
This opens up the possibility of artistic upgrades,
like the projection of moving clouds, simulated sunrise/sunset or even nightsky.
How to avoid unrealistic shadows on the painted backdrop :
Distracting shadows don't occur when the light rays are parallel to the back wall, or nearly so.
If a shadow hits the wall, it should be hidden from the spectator by some foreground element.
Another technique is to have scenery elements touch the wall (possibly with some part
of the element painted on the wall itself).
Finally, if all else fails, you can "erase" the shadow by illumating it "just right"
with localized lighting (possibly using a few bright LEDs on the backside of
the offending scenery element). Such a technique is so unexpected by the brain
of the spectator that is will create a strong illusion that the wall is not even there.
It's very delicate to do right, though. An artform in itself.
The magic of mirrors :
One powerful technique to create an illusion of space at midrange
is through the use of mirrors. We have all seen how small restaurants may appear
much larger by having the upper part of at least one of their walls mirrored up to the
ceiling. This makes the ceiling look at least twice as large
(possibly infinitely larger if two opposite walls are mirrored).
There are several ways mirrors can be successfully applied in a model train layout,
using either large or small mirrors.
Large mirrors will always remain fairly obvious but small mirrors can
be used in ways that fool the spectator completely...
Both techniques require considerable planning to be effective.
If a large mirror perpendicular to the main wall is used,
the tracks should be planned never to go too close to the mirror
or be hidden from it by scenery elements.
If at all possible, arrange things so that an object and its image
are never seen together (especially for trains and other moving
For example, you may install a sunk track next to the mirror
which is never seen directly but whose reflection in the mirror is
This would give the impression of a double track with one track hidden from view.
If that's the desired effect, watch the direction of circulation
of the mirror image
(if your layout is meant to be consistent with,
say, trains running on the left of a double track).
Any writing visible in the mirror should be painted backwards!
One great way to use small mirrors is to place them
at nearly 45° from the main wall.
This can give the illusion of something beyond
that wall, obtained as the reflection of some hidden perspective,
more or less parallel to the wall.
A lot of planning must bo into this
(it's safer to plan the surrounding scenery for the mirror,
not the other way around).
Another very specialized use of mirrors which I find interesting is to simulate depth for
the "subway" stairways of platforms (when actual holes are ruled out).
One example is the Butterfly Station Platform Shelters
in the Walthers Cornerstone Series (HO-scale,
are sold separately from the matching
Union Station). The roof over the platform will typically prevents the
spectator from looking directly into the stairwell and discovering the trick
(a first-surface mirror
perpendicular to the stairs),
The illusion only works if the slope of the stairs is exactly 45° with identical vertical
and horizontal surfaces (which is the case in the aforementioned kit,
probably to prevent the assembly mistakes that a more realistic asymmetrical design would
allow). A tiny lightsource in the plane of the mirror can complete the illusion.
In most illusions involving mirrors, first-surface mirrors are mandatory.
That's especially true for this deep stairway illusion
(you may also want experiment with lines parallel to the edge of the mirror,
which will help hide its exact location, even to people who suspect its existence).
Knuckle couplers. Loop and hook.
(2014-04-07) Loading gauge and "six-foot way".
The maximum size of rolling stock.
Because railways were developped in Britain ahead of the rest of the World,
the need for very large rolling stock was not yet anticipated at the time.
As a result, the British loading gauge is very restrictive.
This is where the locution "six-foot way" comes from.
British outline. German outline.
The Berne gauge defines a clearance enveloppe on a curve of 250 m radius.
(2014-03-17) Sectional Curved Track
First, second, third and fourth radius (R1, R2, R3, R4).
Various manufacturers of sectional track sell elements of curved track
which are pecified to cover a fraction of a full circle (usually
expresses in degrees; if it takes n elements to build a full circle,
then the bend of each element is 360°/n).
The radius of that circle is measured with respect to the center of the track;
each manufacturer proposes their own set of radii, as tabulated below
Note, however, that manufacturers occasionally advertise the footprint
of circles made with their curved elements (this is especially
so in the specifications of starter set "ovals" with or without
In such cases, what's given is the outer diameter of the circular track,
which is equal to twice the central radius of curvature plus one
track width (including crossties and roadbed, if applicable).
For example, starter sets using Bachmann EZ-track curves with an
18'' central radius of curvature have an advertised
diameters of 38'' (since EZ-track is manufactured to
a width of 48.9 mm, which is nearly 2'' ).
Central track radius of a full circle (in mm) for various sectional systems :
Kato's Unitrack® is based on a 24-inch basic radius of curvature
(609.6 mm, rounded to exactly 610 mm) and a 60 mm center-to-center
spacing between parallel tracks. Half a dozen curvatures are commercially available
in that system:
Kato's Unitrack curvatures
Radius / mm
The standard Unitrack siding (featuring parallel track 60 apart, center-to-center)
entails a #6 turnout with two compensator elements: An R867-10 corrector curve
and an S149 (149 mm straight section).
The straight part of the #6 turnout is equivalent to two S174
The combined length of this assembly is
(2014-03-17) Double Track Spacing
Distance between tracks to allow safe crossings of all trains.
6-foot way, bumper locking, overhang, big-boy.
In British railway jargon, the space between the rails of a track is called
the four-foot way and the space separating the outer rails
of parallel tracks is dubbed the six-foot way.
Those traditional terms are a poor indication of the width of those ways
(roughly 5 ft and 9 ft, respectively) but they can ne useful
in sorting out the function of each rail in a busy yard.
In modern times, the quantity of interest is the center-to-center distance
between parallel tracks It's equal to the width of the the afforementioned "ways"
plus the width of 2 rails.
French tracks are ordinarily spaced 4.2 m (4.5 m for high-speed trains)
center-to-center, for straight segments.
At HO scale, 4.2 m becomes 48.23 mm. However, for the layout to accomodate
OO models, the distance must be at least 55.12 mm.
The 4.5 m French high-speed standard becomes 59 mm at OO scale,
which is probably a good design standard for straight tracks in a
layout meant to accomodate HO and OO models.
On a curved track, rolling stock can protrude significantly inward and outward...
(2014-03-30) The frog number is the cotangent of the frog angle.
At the heel of a #N turnout of length N, the tracks are one unit apart.
In other words, the number rating of a turnout is the length it takes to achieve a unit offset.
In US railroading, the rated speed of a switch (in mph) depends
on its number: It's about twice the number for moderately long turnouts (#15 or #20)
it's less than that for the sharper turnouts used in yards (low numbers).
The rated speed of very long turnouts (#22 or more) is more that twice their number.
The Russians rate their turnouts by giving the tangent of the crossing angle, which is simply 1/N.
A few problems with modular track systems :
The lengths of the straight sections in the #5 and #6 turnouts of Bachmann's EZ-track system
are respectively 11½'' and 15½''.
Both are packaged with a small 2¼'' additional piece of
straight track which have an essential feature which isn't documented.
The two tracks at the heel of those long turnouts are too close to mate with
regular pieces. However, the underside of short pieces feature two slanted groves
on the side meant to meant with the turnout.
You need two such pieces to use the turnout (one ot them has to be purchased
separately). Soften the roadbed by bending back and forth two opposite groves
on short pieces, then fold the roadbed inward (you could cut it or break it bit you
don't have to). So modified, the pieces can mate with the two tracks at the
heel of the turnout.
Unless there's a matching turnout somewhere else in the layout,
the length of its straight path must be compensated by short straight pieces
from the EZ-track system
(4½'' and 2¼'' pieces are sold in packs of four, and an assorted
pack of 10 pieces is sold as #44592 which contains 5 sizes:
¾'', 1'', 1¼'', 1½'' and 2'').
(2014-03-21) Systems of Track Sections
What manufacturers ought to supply to allow interesting compact layouts.
Besides the long curved elements described above,
there should be long straight elements of roughly the
Turnout components are indispensable which allow a choice between a straight
path and a curved one (in a left turnout, the curved path is to the
left of the straight one; it's the opposite in a right turnout ).
For compactness and flexibility, the straight and curved sections in standard
turnouts are shorter than the standalone straight and curved long sections.
Shorter elements must be supplied to accomodate the following layout requirement:
(2014-03-29) Block Detection and Transponding
Locating DCC equipment by monitoring changes in current.
In a track layout divided into electrically isolated blocks, the current supplied to each block
can be monitored. In analog (DC) mode, this merely tells whether something
is on a given detection block (and/or if something has moved from one block to the next).
With digital control (DCC) the command center can rapidly switch on and off a device
(typically, the headlight of a locomotive) and determine precisely on what block
that device is located by sensing which block experiences a change in current.
This technique gives the appearance of two-way feedback communications
(it looks as though a locomotive is telling its location to the command center).
It's known as transponding.
For a computer to avoid collisions it's also important that all rolling stock draw at least
some current to be detectable
(at the very least, the first and last element of a train should do so).
This is most easily accomplished by bypassing the isolation ring(s) in a metal wheelset
with a resistor of no more than 15 k (to draw a current on the order of
1 mA or more). This is done with either
a small surface-mount resistor or some resistive varnish (containing
tiny particles of graphite which come together and form a stable resistor as the varnish dries).
(2014-05-31) Controller Area Network
Open Local Control Bus (OpenLCB).
Input "producers" & output "consumers".
The most popular cables for local area networks (LAN)
consist of four twisted pairs (8 conductors)
connected to two RJ45 jacks with molded plastic hooks
("RJ" stands for
RJ45 jacks are also dubbed "8P8C" as they feature 8 positions and 8 conductors).
Those cables are commonly known as Ethernet cables
(they are used for
Ethernet over twisted pair,
which has replaced the
thin coaxial versions of Ethernet).
They come in several mutually-compatible grades (known as "categories")
according to the maximum data transfer rate they can support.
The evolution of the various grades of those cables parallels the evolution
of Ethernet (at times Ethernet innovations were advertised as working
with existing cables, possibly by using more pairs).
An oversimplified history of Ethernet cables (4 twisted pairs, baseband communications)
For the undemanding specialized use discussed here, any grade will do,
including old Cat3 or Cat5 cable (which is very cheap).
Pictured at left is the
TIA/EIA T568-B color-coding of such cables.
Pair 1 is blue (pins 4 & 5) Pair 2 is orange (pins 1 & 2)
Pair 3 is green (pins 3 & 6) Pair 4 is brown (pins 7 & 8).
The electrically equivalent T568-A cables use conventions prevalent in the
telephone industry, with pin assignments of the orange and green pairs swapped
compared to the above. Yet,
the orange pair is still called #2 and the green pair is still called #3.
The surrounding of the center pair by another pair is inherited from traditional
2-line telephone RJ14 telephone jacks (6P4C) compatible with single-line
RJ11 telephone jacks (6P2C). The RJ25 three-line jack (6P6C) are mechanically and electrically
compatible with RJ11 and RJ14 but do not obey the same logic as RJ45 or larger jacks,
(the third pair is formed by the two outer conductors and has no equivalent in RJ45).
NMRAnet Physical Layer (RJ45 connector and Cat-3 or Cat-5 cable)
Connected to 6 and/or 7, as needed
Reserved pair (must withstand 100VAC)
Connected to 3 and/or 7, as needed
7.5 VDC to 15 VDC, 500 mA Max (must withstand 27 VAC)
To a telecommunication engineer, this standard may look rather crude.
However, it can serve its undemanding purpose very well and more cheaply than more sophisticated alternatives
(for example, Ethernet can use the full data-transfer capability of the 4 twisted-pairs and still carry
"phantom" DC power via two different common modes on two different pairs).
Nevertheless, the OpenLCB/NMRAnet network is a bus just like Ethernet is.
As such, it should be properly terminated and grounded
(without ground loops) for trouble-free operation.
Reproductions of famous locomotives.
Midland Pullman, blue livery.
Bachmann 51810 Alco 2-6-0 Mogul,
Union Pacific #39 (HO scale, DCC Sound Value).
Hornby R3100 Class A3 (103)
"Flying Scotsman" (OO scale, black livery).
Hornby R2991XS Class A4 (60018)
"Sparrow Hawk" (OO scale, garter blue livery, DCC sound).
Jouef TGV 150 574.8 km/h
world record in 2007 (HO scale, garter blue livery).
Jouef has made two
different sets based one the record-breaking train, both are discontinued:
One is a DC starter set with tracks, transformer and throttle
The collector's edition (ref. HJ2058)
includes only a DCC-ready locomotive,
three cars and a non-powered locomotive (with fake pantograph).
(2014-04-11) Dream Trains
Humble or prestigious passenger services.