How far down the line can a human train crew see a RR signal?

500 feet, 1,000 feet?

The trouble with AI is it does not properly mimic human behavior. It can "see" a signal miles down the line and this is why AI runs trains in a non-authentic fashion. This is why they are running on yellow, half speed, much of the time when they shouldn't. A human train conductor will not begin to react to a signal unless it is clearly visible or he may otherwise act according to a dispatcher on the radio.


There is something odd on my line. My train is approaching a signal that is yellow over green but the signal immediately following is red over green. There is an unexplained yellow with the track clear ahead of it.

A day later.

I found a solution to my "yellow-itis" while driving in AI mode.


First of all, I replaced all the JR double-light signals with JR single-light signals. These are less complex signals unlike the double-light signals with one color light over another. A single light signal has 4 conditions. Red, solid yellow, flashing yellow and green. If any given signal is red, the signal behind it will be solid yellow followed by a flashing yellow followed by a green. Flashing yellow seems to slow trains in a 50 mph to half speed just like a solid yellow does. By putting the signals at such close intervals on the fast parts of the line, this limits AI's vision down the track ahead and reduces the sporadic speeding up and slowing down from intermittent yellow conditions.

In my 35+ MPH zones, I have signals about 1,000 feet apart which seems like overkill but it keeps the line green (clear) for up to 3,000 feet ahead. In my 25 mph zones, signals can be spaced farther apart becaused yellow condition does not adversely affect speed in these slow zones. A train in a 25 zone might travel 22 mph under 4 yellow. In a 50 mph zone, however, yellow will cut speed in half and that is lousy.

These setup is on my scale-model layout with 7.10 miles of closed loop on a double-track mainline system. I can have up to four trains following each other. Passenger trains have to stop every so often, every two to three laps, for 2 to five minutes. I make my MOW trucks stop for one minute at the station after each completion of the loop so it doesn't follow the slower, heavier passenger and freight trains in front of it too closely. If it follows a train too closely in a 35-to-50mph zone, it will sporadically incur a yellow condition because it is light and accelerates very fast. Making the faster trains stop on the loop periodically allows traffic to not get bunched up but keeps them more spread out to minimize yellow conditions on the line. The heavy freights circle the loop non-stop. They tend to be more stable speedwise due to their heavy mass. The lighter vehicles, following their scheduled periodic stops, will get ahead of freight trains fast due to their more rapid acceleration.

The signals use a basic NORAC signal rulebook standard.

JR has also broken down their signals into a simple code, which works well when planning which signals to place where.

Type 01 and Type 04 are Absolute signals.

An Absolute signal is used where a siding or side branch meets a mainline, and be considered a stop sign. Our own traffic signals are supposed to work like absolute signals, but most drivers think of them as permissive signals.

Type 05 are Permissive Signals. These are defined as permissive due to the small name board placed below the signal head. For AI drivers, it's a script, but in real life an engineer will call out the signal to the dispatcher and ask permission to pass should the signal aspect be anything but green or yellow. Yes, there can be red Type 05s, but they are not indicating stops, just the signal ahead is red.

These signals are used in conjunction with the Type 04 signals, and are used in between them. They allow trains to follow each other as they indicate the signal ahead.

With a Type 05 being permissive, the AI will pass by them at the indicated speed. A yellow-flashing signal indicates the following signal is a solid yellow. A solid yellow in turn indicates that the next signal is red, and to slow down to 1/2 the posted limit to prepare to stop.

Type 06 are Diverging signals and are used in conjunction with the beginnings of passing sidings when track goes from single to double. They have various indications such as Red over Green, which indicates a Medium Clear, and to take the right hand track.

There's another variant of these signals called Approach Signals, and these have a number board below the heads, just like the Type 05s. These are also called Distant Signals, and are used to indicate the upcoming signal indications of these so that the drivers can prepare to slow down if necessary.

Then finally there's the Type 08s or Interlocking Signals, which are used for wyes, and crossovers. These have 3-heads, which can display various aspects. Usually one head is always red on these signals.

With adequately spaced signals and by using the proper signals for the purpose, you can have relatively smooth driving. In general the spacing of signals is based on a few parameters.

1) The length of the train
2) The speed of the train
3) The number of trains

The length and number of trains generally factor into No. 2. With lots of trains following each other, there's a chance of having a follow the yellows scenario because each following consist will move no faster than the indicated signal. With trains following each other, there's a chance of running into a red signal due to a block occupied by another train, and with a block occupied ahead, each following train will be under restriction and running under yellows. If there are lots of shorter trains, this can mean a lot faster trains running within the route because each train will quickly clear the block ahead.

And finally, the longest train run will determine the signal spacing throughout the route. The problem is having lots of short trains running on a route with massively spaced signals can cause poor running times because the following trains will end up running up to a red signal while the train ahead trundles down the track and finally passes the signal ahead. The longer spacing of signals is also found on tracks with faster speeds because generally there are fewer trains running, which will run into this problem.

Working out signal spacing takes experimentation and experience. For Trainz I use the 2000T Coal consist, which is built in as my ruler to determine the signal spacing. For longer faster tracks, I use doubles, and for general commuter lines, I use the single consist.

Hi John, you're assumption is both correct and incorrect. In real life, each class 1 railroad (And even class 2's,), have their own rules regarding signals, when to react to them, and what to do.
Seeing that you are here in the Boise, ID area, you should go over to Nampa, and check out the real life UP mainline there, and watch the trains depart and the signal indications, and what the crews do.
So, say you have a train departing eastbound out of Nampa, but over at Kuna, a train coming west is getting ready to roll into the yard for a crew change. The first interlocking signal up near the terminal will display green over 2 reds, indicating completely clear for the next block, the 2 reds on the bottom have no meaning in this case, but say, that the signal on top is solid yellow over 2 reds instead, this means approach, entering the next block prepared to stop at the signal and no faster than 35MPH regardless of track speed. This means the train will crawl into the next block, and the bigger and heavier the train, the longer the time it is going to take to stop and react, meaning that most of the time, for trains such as MHKNP that are over 10,000ft and 13,000 tons+ will not enter this block at more than 15MPH, regardless of the 35, and then 50MPH track speeds.
Now let's get back to our scenario on the main. The first 2 blocks east out of Nampa, will display greens over reds, meaning all clear. The next intermediate at Robinson road will be a flashing yellow, known as advanced approach. This means proceed into the next block preparing to slow to a speed of no greater than 35MPH, as the next signal will either be yellow over yellow (This is if he is going into the siding at Kuna), or solid yellow if he is preparing for the meet at CP Fox. Track speed out of Nampa is 65 MPH, but with the signal indication, he can go no faster than 35MPH. So he hits the next intermediate at less than 35 MPH preparing to stop at the red over red at CP Fox, meaning stop and waits out the other train. But let's talk about if he was going into the siding. CP Fox is an interesting Control point, as 2 main territory goes down to 1, but in about .5 miles, there will be CP Kuna, which is the sidings, creating a unique back-to-back CP situation which gets a unique signal indication. So the next intermediate before CP fox will be Green, but then CP Fox itself will be red over flashing yellow, meaning advanced diverging approach, proceed at a speed of no great than 35MPH being prepared for the next signal, at CP Kuna, which will display Solild yellow over red, meaning diverging approach, proceed at prescribed speed through the turn out (in this case, 25MPH) and take the siding, prepared to stop at the next signal. Through here, he will be rolling through the crossing at Black Cat rd., at about 35MPH, even though track speed is 70MPH. So basically, check your own railroad your portraying rules, and see what those signal indications would mean, because surprisingly enough, the trainz AI is actually almost true to life in how they react the signal indications.

Now for your other part, dual headed signals, are mostly for your turn outs, such as sidings, and divergences from X-mains down to x number of mains. Three headed lights are for interlockings, and yard entrances and stuff like that. Those one headed signals are the intermediates to give indication of what is happening as you proceed down the line. For example, the siding at Owyhee, is a 25MPH turnout, but track speed is 70, say that there are no intermediates between the siding at Kuna, and the siding at Owyhee, how would the engineer know other than DS telling him, that he would need to slow down to 25 MPH to enter the siding at Owyhee from Kuna, and wait out another train? With these signals, being able to run multiple trains on a single track is now much easier than it was in the past with dark territory. This is why accidents related to this are less common now, unless somebody is not paying attention, which is almost why nearly all accidents involving this was either a signal failure, or human error.

Honestly, I would space your signals out a little further, as this would actually space your trains out more as well, causing better performance for the AI as well. On my 70+ mile route, I have signals spaced out between 1.5 miles and 2.5 miles, and the trains operate very smooth, and react to their indications accordingly. So this is really a test and play kinda deal. But understanding of actual signal rules, and operations will also help you out a lot, as you can understand what exactly is occurring so that you can clear it up, and see where you need more capacity and such.

Very good treatment of basic signaling as used by N3V.

Invisible signals help AI with this. Placed between the train and the real signal.

gd5150 said:

Invisible signals help AI with this. Placed between the train and the real signal.

Click to expand...

And in particular, there is a certain situation where this is especially important. By default Trainz signals can only see one set of points ahead of them regardless if they're absolute or permissive - any others beyond the first are literally invisible to that signal, hence it's impossible for a single signal to control or guard all the points in a yard for instance (as the old official Auran Trainz signalling guide said, this is not exactly the same as real life railway signalling practice).

The only way for a signal to function with multiple sets of points is to either use a third party signal that is uses scripting to see those multiple junctions ahead of it and guard them accordingly, or to otherwise place invisible signals in front of every set of points past the first. That way the signal will see the next signal after the points it guards, and an AI driver should control their train accordingly.

Of course though, these are AI drivers we're talking about...

I my experience the distance a signal can be clearly seen is highly variable. It does depend on the line of sight (the point that the signal can be physically seen from with out something physically blocking the view), the time of day (lit signals are much more visible at night, and forget trying to read a semaphore with out a light in the dark), type of signal and signal lighting (the newer LED ones are really bright, but try reading a dwarf signal in a late season big snowstorm), the color (you bet, try it on a traffic light if you can find on you can get far enough away from), weather (fog, rain, snow all degrade visibility), the crew's situational awareness and visual acuity (knowing when and where to look, as well as what to look for, and how good are their eyes).

Depending on conditions the distance a crew can see a signal can be as little as a hundred feet (or less) or upwards of a couple of miles. For example a rookie crew dealing with bad weather such as heavy precipitation or fog might be nearly on top of a signal before they can correctly interpret it. On the other hand an experienced crew under optimal conditions may be able to see multiple signals ahead of them all miles distant. What you had for the previous signal governs what you can get for the next signal, unless there are special circumstances (most of which are bad).

To elaborate on JCitron's previous comments it seems the US style signaling in Trainz loosely follows NORAC practice and not other operating rules (GCOR for example). For those who might not know, NORAC (Northeast Operating Rules Advisory Committee) is the ruleset commonly used in the northeastern US including the famed Northeast Corridor (currently using it's eleventh edition). This is a goodly bit of studying if you are inclined to learn it, mostly memorization. It is mostly a speed based system, and the signals give no indication of which way you are going at a switch/interlocking just how fast you should be going when you get to the switch/interlocking. Of course experience and reasoning will tell you what's going to happen; that "approach medium" my technically mean to approach the next signal at 30 mph, but it does generally mean your headed into a siding or through a crossover.

Most of the signal problems can be worked out if you understand the logic and rules. However, as JCitron said, many short trains in long blocks are an issue. I would argue that one should not base your signal blocks on the longest train you plan on running but on the longest regular train you plan on running (fifty tank cars of clay slurry is a lot shorter, and heavier, than fifty auto racks). [EDIT] largest not longest train...


It will take a bit of trial and error to get the signals right on a route... but a least in Trainz you aren't looking at serious consequences for running a signal.

All of which could be an academic discussion in a decade or two (or possibly sooner).

I was reading an interesting report on the future of UK signalling (can't recall where or what it was) but it seems that UK's Network Rail and probably many others as well are looking at abolishing track signalling entirely. Each train would be enclosed in an electronic "bubble" using GPS and other technologies. The bubble will move with the train and these bubbles will keep all other trains at set distances from each other and moving at set speeds, depending on factors such as track condition, traffic levels, weather, etc.