How often should a train pass through your layout?

JonMyrlennBailey

Active member
...to be realistic

I have a two-track g-scale layout. 7 scale miles of mainline is on the visible layout and 5 scale miles is in the hidden staging yard room.

Tunnel portals separate the visible layout from the hidden room.

I have 3 d/e freights, 1 d/e passenger 1 excursion steam train and 1 Amtrak MOW truck on each of both tracks of the closed mainline circuit.

I have a modern mix of d/e hood unit freights: autorack automobile carrier unit train, TOFC unit train, wellcar unit train, tank car unit train and two manifest (mixed) freights.

The staging yard has sidings for each of the d/e trains but the steam trains and MOW trucks share a common siding so one vehicle parks behind the other in turn and waits.

There is an AI schedule for each of the 12 "trains".

The staging yard is a double ladder. Each way of the two-track mainline has its own yard so constructed.

All the trains have to pass through a common junction to get back onto the main line following their wait in the staging yard.

At each staging yard exit junction on the ladder where it merges on to the mainline, there is a track mark. I have all the d/e freights hold at this junction for 15 minutes before getting back on to the mainline. Each d/e freight stays on its respective siding for 15 minutes also. This lessens the chance of trains getting in each other's way at the common junction where they have to re-enter the mainline. The steam excursion trains are held on their sidings for two hours. The d/e passengers are held on their sidings for one hour. The MOW trucks are held for 30 minutes if it is the lead vehicle on the shared siding but if a steam train is holding in front of it it will have to hold for at least 2 hours and 30 minutes. In modern American railroading, d/e freight trains are more frequent and more common. Passenger trains are rarer and steam locomotives are special these days so they get even less frequent travel on the mainline. More d/e freight trains seen on the layout more often and fewer passenger and even fewer excursion trains seen less often.

Yes, MOW trucks should not be seen frequently too.

Using these AI schedule timings as said above, and observing one particular grade crossing on the layout for two hours straight railfan style, I counted 13 trains pass in both directions combined. That's about one train each and every 9.2 minutes. The number of trains on the route, the number of tracks, the length of the mainline circuit and the wait times dictate the frequency of how often trains pass.

I would have to increase the hold times of trains in the staging yard for so many minutes especially the d/e freights which are most numerous so that o average only one train apsses in either direction every 15 minutes. I would not what the advanced math is to make such a calculation so I will just have to add minutes to wait times and recheck.

The staging yard acts like a dispenser, admitting one train at a time to the mainline. Think of the roller coasters with multiple trains at amusement parks. One train comes in and lines up behind another and the front-most one loaded with passengers is sent down the line. I could line up trains one behind the other in tandem at a timed hold point on the hidden portion of the mainline AND without using a yard but that would take up too much space. The double-ladder yard with parallel sidings is much more compact for holding and releasing timed trains.


Right now, I'm running a new experiment. I added 15 minutes hold time to each of the six d/e freights to see how that now affects total train frequency on the line. Trainz is a fun little guinea pig of sorts to fiddle with. The math here is much trickier than those "train" questions you had on high school SAT tests. Complex computer algorithms must be used to plan modern train scheduling. I like the old image of passenger service whereby the conductor holds a Swiss pocket watch
to keep trains running on time. That was the old pride of American railroading: timeliness.
 
Last edited:
On my Gloucester Terminal electric, the d/e connecting mainline is a combination of freights and d/e commuter rail. I have a total of eight (8) consists emitting from a portal at each end at 15 minute intervals on a route of the same length. This means I get one of four (4) of these trains on the mainline every 15 minutes. Being double-track, which it is in real life, because part of this is based on the actual Gloucester branch of the former Boston and Maine, and originally the Gloucester branch of the Eastern Railroad, there are no conflicts. Given that in real life, there is no longer any freight on the line, these through freights add a bit of excitement too the otherwise commuter-only line, however.

Is this too much traffic? It doesn's seem that way. There are occasional meets in various places, but usually near Gloucester because the track is a bit slower there as the Rockport branch comes in at a tighter curve (in real life too), and because that's where Gloucester station is and the commuter trains stop there, and also this is just after the Annisquam River drawbridge, which requires a 20 mph crossing on what is otherwise a 60 mph line. This leads to a bit of excitement as a slowing commuter train is pulling into the 3-track terminal while a freight is running quickly through the station in other direction all the while both trains are blowing for the crossings at the same time a 2-car trolley train-set is pulling into the third platform!

What controls the frequency of trains, in addition to need are the following factors:

* track speed
* bunching
* signaling, spacing, etc.
* tracks available for traffic - like lanes on a highway.

The overall track speed can determine the number of trains that can run on a given line, but this is somewhat defeating as well and declines proportionally to the number of trains on a line. It's simple. A single train can run the route at the posted speed limits. A second train in the same direction, on double track, can travel at the same speed if timed a bit later to allow the first to pass a few blocks. If you start adding more traffic, well you'll start to slow things down as the additional trains need to slow down to half the posted speed as they encounter yellow signals. This is based on the NORAC rulebook, which is the basis for N3V AI signaling. Single track is another matter altogether because this can lead to longer delays as trains need to wait for clear track ahead while they sit at the end of the passing sidings (loops).

To control bunching, which is eluded to in the waiting for yellows, the trains are given what is called headway between them. This is to allow for the trains to pass signals and remain in a block before a following train is allowed on to the main. The purpose of this is to let the first train get ahead far enough so not to get the following train stuck. On freight-only lines this isn't too much of a problem, but dispatchers don't like clogged lines and will do what the can to prevent this, however, on a commuter route or transit system, this is a definite no-no, and will lead to really annoyed passengers. On a trolley system, for example, or even a subway system such as SF's Muni, Boston's MBTA green, blue, red, and orange lines, this is extremely important.

Signaling is very important in all this. Transit companies, especially, not only go through great lengths to ensure there that the signals are spaced far enough apart so as to keep a single track within a block, but also allow multiple trains on the tracks at the same time.
(The term block here is meant to show an occupied section of track regardless of the kind of signaling system used.) What this means is the signals have to be spaced apart a bit more than the longest train. If the system runs short trains in a territory with far-spaced signals, this can lead to dwell-time issues as the train will take too long to leave a section causing following trains to wait. Having too long a train in closer signals, will effectively tie up the line as it can occupy more than one block and cause multiple-red signals across the system.

Then finally, we have multiple tracks. In theory having more than one track in an area will increase the number of trains running, and it does to a certain point. The problem is all these additional tracks need to converge together at some point, and that's where there are problems. Having lots of multiple-track lines looks cool, but there are bottlenecks at junctions as the these multiple lines come together. In real life too these additional tracks are an additional expense due to maintenance and material costs, and railroads will avoid this as much as possible.

As you said it's a bit of complex math to get this worked out. Being a hobby, why would we want to make stuff more complex than it is. A simple does it feel right in this case usually does the trick.
 
On my Gloucester Terminal electric, the d/e connecting mainline is a combination of freights and d/e commuter rail. I have a total of eight (8) consists emitting from a portal at each end at 15 minute intervals on a route of the same length. This means I get one of four (4) of these trains on the mainline every 15 minutes. Being double-track, which it is in real life, because part of this is based on the actual Gloucester branch of the former Boston and Maine, and originally the Gloucester branch of the Eastern Railroad, there are no conflicts. Given that in real life, there is no longer any freight on the line, these through freights add a bit of excitement too the otherwise commuter-only line, however.

Is this too much traffic? It doesn's seem that way. There are occasional meets in various places, but usually near Gloucester because the track is a bit slower there as the Rockport branch comes in at a tighter curve (in real life too), and because that's where Gloucester station is and the commuter trains stop there, and also this is just after the Annisquam River drawbridge, which requires a 20 mph crossing on what is otherwise a 60 mph line. This leads to a bit of excitement as a slowing commuter train is pulling into the 3-track terminal while a freight is running quickly through the station in other direction all the while both trains are blowing for the crossings at the same time a 2-car trolley train-set is pulling into the third platform!

What controls the frequency of trains, in addition to need are the following factors:

* track speed
* bunching
* signaling, spacing, etc.
* tracks available for traffic - like lanes on a highway.

The overall track speed can determine the number of trains that can run on a given line, but this is somewhat defeating as well and declines proportionally to the number of trains on a line. It's simple. A single train can run the route at the posted speed limits. A second train in the same direction, on double track, can travel at the same speed if timed a bit later to allow the first to pass a few blocks. If you start adding more traffic, well you'll start to slow things down as the additional trains need to slow down to half the posted speed as they encounter yellow signals. This is based on the NORAC rulebook, which is the basis for N3V AI signaling. Single track is another matter altogether because this can lead to longer delays as trains need to wait for clear track ahead while they sit at the end of the passing sidings (loops).

To control bunching, which is eluded to in the waiting for yellows, the trains are given what is called headway between them. This is to allow for the trains to pass signals and remain in a block before a following train is allowed on to the main. The purpose of this is to let the first train get ahead far enough so not to get the following train stuck. On freight-only lines this isn't too much of a problem, but dispatchers don't like clogged lines and will do what the can to prevent this, however, on a commuter route or transit system, this is a definite no-no, and will lead to really annoyed passengers. On a trolley system, for example, or even a subway system such as SF's Muni, Boston's MBTA green, blue, red, and orange lines, this is extremely important.

Signaling is very important in all this. Transit companies, especially, not only go through great lengths to ensure there that the signals are spaced far enough apart so as to keep a single track within a block, but also allow multiple trains on the tracks at the same time.
(The term block here is meant to show an occupied section of track regardless of the kind of signaling system used.) What this means is the signals have to be spaced apart a bit more than the longest train. If the system runs short trains in a territory with far-spaced signals, this can lead to dwell-time issues as the train will take too long to leave a section causing following trains to wait. Having too long a train in closer signals, will effectively tie up the line as it can occupy more than one block and cause multiple-red signals across the system.

Then finally, we have multiple tracks. In theory having more than one track in an area will increase the number of trains running, and it does to a certain point. The problem is all these additional tracks need to converge together at some point, and that's where there are problems. Having lots of multiple-track lines looks cool, but there are bottlenecks at junctions as the these multiple lines come together. In real life too these additional tracks are an additional expense due to maintenance and material costs, and railroads will avoid this as much as possible.

As you said it's a bit of complex math to get this worked out. Being a hobby, why would we want to make stuff more complex than it is. A simple does it feel right in this case usually does the trick.

Thanks, John. I don't really aim to be a logistics manager for a railroad or an industrial or traffic engineer for one but I find the mad science behind all this quite fascinating. The simple way for me is to fiddle with the wait times in the staging yard where the trains hold for so long and are admitted to the the mainline over time. Yes, limiting the frequency of trains passing over the route at a time adds realism and eliminates yellow conditions at least almost all the time. Logically, it seems, if one increases the wait times in holding, the overall frequency of line traffic will decrease. Each and every train has to do a wait in the staging yard for each and every pass or lap of the route it makes. Six trains all have to funnel through one common junction and yard ladder for each side of the line to re-enter the mainline from the yards so only one train can be admitted at a time to the main tracks.
 
Last edited:
I agree the mad science behind this is very fascinating. Being a technician with some engineering background in electronics, I find complex setups and the timing involved with them very fascinating as well.

The thing is do what looks good. Keep in mind we don't have a real dispatcher and rely on timing to ensure that the AI don't tie things up and bump heads along the way. I learned in the past that all it takes is a minute or two of separation in opposite directions to prevent a cornfield meet in 99.99% of the cases. By offestting the consists just a bit, I'm able to run multiple trains in both directions and have them timed so that they will be on the passing siding to allow the other train to pass.

Getting this to work is a challenge, but worth it in the end, and I do it by eye and a lot of analysis and thinking offline. If we had a dispatcher module that we could connect to a session, this would be very helpful. The dispatcher, or dispatchers, then could control the junctions for us and the schedule so that the AI don't get stuck, and the operation would be more realistic because in real-life, the engineer and crew rarely stop to flip their own switches unless it's in an industrial park, yard, or on a transit line where the trolley conductor needs to aling the switch for his route.

With the dispatcher module, this would require permission for even a human driver to enter a block if stopped at an absolute signal, and this would prevent the AI from flipping switches on us while we drive. A call into the dispatcher could be as simple as a menu with call dispatcher to request permission. An answer such as permission granted and then the switch changes for us would be perfect. On automatic mainlines, an operation similar to the Interlocking Towers could be implemented, but make their setup so much easier, and allow for interaction with them, which is not there now. This can get quite complicaed, but it's a whole entity in its self. With multi-player sessions, their could be a human dispatcher controlling the route. Thinking about this off the cuff right now, that might even be possible where we're a dispatcher instead of a driver, and we're now in charge of controlling all the AI instead of the AI running around and doing their own thing.
 
Back
Top