Crazy Idea for Light Rail

[Just Crazy Jim]

So as I was sitting here being mightily frustrated with bugs in the user content tools, I smoked a pipe and watched a short video about America's lost light rail. Basically electric streetcars that went on interurban routes, almost all gone by 1945. I said to myself, that would be ever so clever to add to RRT3 for a period set metro map... but how to do it? After a few minutes, I came up with the idea of an invisible locomotive (which is crazy even by my standards) or a modified passenger car with cowcatcher at both ends and a trolley pole or bow collector on the roof hauling invisible passenger cars (also mad beyond my usual levels madness). Finally, I settled into a more familiar state of madness and accepted that there would be far too much work in making a locomotive (like an inspection locomotive) that also carries PAX with no consist. Then I saw two trolleys coupled in the video and came up with the idea of an electric loco clone using a B Era passenger car model and a B Era passenger car model as the tender.

Now, the $64,000 dollar question: Can a tender CAR file be hacked to carry cargo?
So you won't think me completely mad, a picture of an inspection locomotive.

[Gumboots]

No, a tender file cannot be hacked to carry cargo. You're plum out of luck there.

However...

I dd have one idea not long ago: McKeen railmotors.

These look cool, and would give the diesel nuts* something to play with in the early 20th century. They'd also be pretty easy to model and skin. Here's what I said to RoR via PM when I thought of this:

Incidentally here's an idea I had yesterday: 1910 McKeen railcar. Easy to model and skin, and would be a lot of fun to have running around. By playing with stats it can be made to go full speed with one load of express (which is what it actually carried of course) while being downright useless with even two loads of express. This requires using almost no weight for it, with almost all the weight being the actual consist, so fuel cost empty would be almost zero and fuel cost loaded wouldn't be a whole lot more. You'd basically just be paying for purchase and maintenance. Could be a nifty thing for scooting around the countryside picking up occasional express loads on lines that were basically freight.

The tricky bit here, which I didn't mention, would be to use LengthPoint trickery to suck the single passenger car up inside the McKeen. As long as both were scaled so the McKeen was slightly larger, the visible result would be a lone McKeen but it would still be hauling one car of express.

*I think diesel nuts would like them, even though the McKeen's were actually gasoline.

[Just Crazy Jim]

Brilliant! I had seen a few period images of the McKeens, but never knew a name for them. And I think the... let's call them "rail modernists"... would like any early power unit that's not slung from a wire or burning coal or fuel oil. I never got any thrills from diesel-electrics, there are plenty enough of them rolling by, usually dragging a mind-boggling number of coal hoppers, blocking the road for a half-hour when I need to be some place 20 minutes ago. At least steam has some stage presence.

Although, I saw an image once of an early acetylene burning sort of electric prototype.

And there's always Chesapeake & Ohio's monstrosity, the class M-1 steam turbine.

Norfolk & Western also tried their hand at a steam turbine.

One has to wonder what they were smoking.

[Gumboots]

Hypothetically, if you wanted to do a scenario with all freight handled by real locos and express handled by railcars, it would be possible to do it like this:

You'd customise the express cars so they had a transparent DDS for the skins. 16 px square would do. No point using up rendering on them. You could also substitute a smaller and simpler body mesh if you wanted to. Even a basic triangle would do, and obviously wouldn't need LOD's. No need for wheels or trucks either. The game is happy with just a body file, so nice and simple.

You'd leave the weights as normal, and use the aforementioned LengthPoint trickery to position all the consist cars inside the hauler. This would allow you to run normal express consists from 1 to 8 cars, including dining car if you wanted it, while having it look just like a railcar. Would also cut processing and rendering loads on the game engine, as a bonus.

[Just Crazy Jim]

Hypothetically, if you wanted to do a scenario with all freight handled by real locos and express handled by railcars, it would be possible to do it like this:

You'd customise the express cars so they had a transparent DDS for the skins. 16 px square would do. No point using up rendering on them. You could also substitute a smaller and simpler body mesh if you wanted to. Even a basic triangle would do, and obviously wouldn't need LOD's. No need for wheels or trucks either. The game is happy with just a body file, so nice and simple.

You'd leave the weights as normal, and use the aforementioned LengthPoint trickery to position all the consist cars inside the hauler. This would allow you to run normal express consists from 1 to 8 cars, including dining car if you wanted it, while having it look just like a railcar. Would also cut processing and rendering loads on the game engine, as a bonus.

That's sort of where I was with the "invisible" stuff, but it never occurred to me to hide them inside the body of the locomotive (because I didn't know it was possible to do such a thing). That idea - in and of itself - is pure genius. The assorted benefits on the render load are just icing on the cake.

Mmm, cake. Fat man wants cake now...

In other games, I have made objects with no texture at all, playing with the lighting and reflection values to give things the appearance of being made of solid gold or glass or what have you. But the models in RRT3 aren't complex or complete enough for that sort of foolishness. It was the fact that the light bleeds through re-skins that I started me thinking about making a transparent texture on a locomotive, but my very next thought was "it will probably become a thing of pure light as soon as the sun sets". That realization is what sent me to thinking about tenders as cargo containers.

[Gumboots]

Hey that just gave me another idea. With the revised weight change scale I've been working on, the last era comes in at about the time cabooses were phased out in real life. They also look out of place in the game in the 21st century, especially with the futuristic locos. OTOH, the game needs to have one coded for all years, and it's handy if you want the reliability boost.

So I can just make the caboose for the last era invisible. It's a lot less work than modelling and skinning a visible one, and for the game's purposes will work just as well. It'd still have a profile icon in the train list, so you'd know when you were using it. It just wouldn't show up on the tracks.

Edit: Or I could just skin it as a basic EOTD and give it a red light at night.

[Gumboots]

That's sort of where I was with the "invisible" stuff, but it never occurred to me to hide them inside the body of the locomotive (because I didn't know it was possible to do such a thing).

viewtopic.php?f=67&t=3890#p39934

[Just Crazy Jim]

Hey that just gave me another idea. With the revised weight change scale I've been working on, the last era comes in at about the time cabooses were phased out in real life. They also look out of place in the game in the 21st century, especially with the futuristic locos. OTOH, the game needs to have one coded for all years, and it's handy if you want the reliability boost.

So I can just make the caboose for the last era invisible. It's a lot less work than modelling and skinning a visible one, and for the game's purposes will work just as well. It'd still have a profile icon in the train list, so you'd know when you were using it. It just wouldn't show up on the tracks.

Edit: Or I could just skin it as a basic EOTD and give it a red light at night.

I was living in Boston (Massachusetts) during the period that US rail was phasing out the caboose. I would occasionally see a loco pass with a consist of nothing but cabooses. I saw models from all periods, and shapes, and in every livery imaginable, but none of them were red.

The thing that replaced them was a snap-on flashing red light. I've never gotten close enough to see one of these lights in any detail, but it's attached above the coupling and a bit larger than a locomotive forward-facing lamp. I think I recall being told that the concept for the tail-end light replacing the caboose was Australian in origin. I imagine that might be true, because I can well imagine that on those long hauls through God awful nowhere, you'd want the relief crew in a position near the head of the train, not all the way at the back, to facilitate stop-free transport of freight.

There was probably some sort of similar arrangement for US rail during the peak of steam passenger transport. The end car was not a caboose on passenger service in the US, but a streamlined tail car with a prominent red lamp.

Thinking about those long distances for Australian rail reminds me of the problems of water for steam. One solution the US rail had was a 2nd water tender. Another was a scooping system that used a snorkel entering a through of water in the center of the track (or at the side of the track), allowing the water to be taken on at full throttle, never stopping. This scooping system was what allowed locos during the golden age of steam to set record speeds. That and projects like the Lackawanna Cut-Off that changed the routes. The iconic water tower of RRT3 was more a yard item than a feature of the landscape, and possibly something of a Hollywoodism.

TBH, I think to get the game to better express these things, the water usage of late steam would have to be adjusted down. Possibly to the same rate as oil. Maybe then the downtime on the Big Boy and Challenger would decrease enough to show the sort of profitability they had when in real world service. Diesel-electrics didn't become powerful enough or economical until the oil companies got into the business of making petroleum based diesel. Before WW2, diesel fuel production was almost exclusively bio-diesel and not abundant or all that economical. By then, we start to see the cargo weights creeping up and pushing the pulling power requirements creep past a threshold that single unit steam couldn't haul. The costs of adding a second steam loco changed the money game in favor of single unit diesel-electric.

Passenger service may or may not have been effected by the change in locomotives from steam to electric/diesel-electric. About this same time, the US embarked on a massive Interstate highway building project that made travel by automobiles much easier and more economical. This also made moving cargo by truck more than a from the railway station to final destination mode of transportation. There's a direct correlation between the progress of the Interstate highway project and the late 20th century decline of US rail. These days, I see mile-long coal trains with fair regularity, usually draughted by multiple diesel-electrics, as many as 4. I figure that diesel-electrics are approaching that same threshold that steam locos faced 50 or so years earlier.

[Gumboots]

Yup, an EOTD, or ETD, or FRED, and probably several other names, is the clip-on red light thingy. There are plenty of pix online if you look around. Comes in several different varieties, but for game modelling purposes it's basically just a small rectangular box with a handle on it. Very easy to model and skin. Would only need a 64x64 A skin and 2 LOD's, plus a light file.

The deal would be to call it a caboose, but not have any files for trucks or wheels (the game doesn't force you to have them). Just the basic light box. Set Front and rear length points to zero. Ditto for front and rear track points. This will make it sit smack behind the preceding car.

I'd offset the body mesh slightly forward of the zero point, because the length point of the preceding car will be at the end of its drawbar and you want the EOTD snuggling up against the car itself. I'd also offset it slightly to one side to clear the drawbar. Result would be a gizmo that looks just like the ones on real freight trains, but the game would think it was a caboose. So it'd chew up one car slot, but give you a 50% reduction in breakdown chance.

[Gumboots]

TBH, I think to get the game to better express these things, the water usage of late steam would have to be adjusted down. Possibly to the same rate as oil.

That's a common trap for scenario authors. It sounds clever, but it aint, for a couple of reasons.

1/ Reliability decreases as oil level decreases. The lower you run it, the more often you will break down.

2/ If you run the loco out of oil completely the breakdown chance skyrockets, and even when you refill the loco it will have permanent damage to its reliability. You do not want to run an RT3 loco out of oil, ever.

3/ Sand. On flat track you don't use any sand. On graded track of around 3-4%, all locos chew through sand as fast as a steamer chews through water. IOW, diesels and electrics in RT3 only have greater range on flat or lightly graded track. On moderate to heavy grades they will run out of sand ages before they run out of oil, and that will slow them right down (albeit with no reliability penalty).

If you want greater range, the way to do it is to reduce consumption of water and oil and sand if you really want a balanced result.

Maybe then the downtime on the Big Boy and Challenger would decrease enough to show the sort of profitability they had when in real world service.

Nope. The problem is that the game devs set the weight of the Challenger and Big Boy so high that they are basically never profitable to use. They can make some profit sometimes, if you're lucky, but another loco will always be better. Loco weight is part of the calculations for fuel consumption, so their fuel bills are always prohibitive. That's something we intend to fix soonish.

[Just Crazy Jim]

That's a common trap for scenario authors. It sounds clever, but it aint, for a couple of reasons.

1/ Reliability decreases as oil level decreases. The lower you run it, the more often you will break down.

2/ If you run the loco out of oil completely the breakdown chance skyrockets, and even when you refill the loco it will have permanent damage to its reliability. You do not want to run an RT3 loco out of oil, ever.

3/ Sand. On flat track you don't use any sand. On graded track of around 3-4%, all locos chew through sand as fast as a steamer chews through water. IOW, diesels and electrics in RT3 only have greater range on flat or lightly graded track. On moderate to heavy grades they will run out of sand ages before they run out of oil, and that will slow them right down (albeit with no reliability penalty).

If you want greater range, the way to do it is to reduce consumption of water and oil and sand if you really want a balanced result.

Heh, as usual, I made the post then ran a scenario with an event to test my pet theory. Two things became apparent:
1) Every thing you just stated is 100% true.
2) if water consumption is reduced (but not oil or sand), there is a high chance a locomotive needing water won't take on water or sand until it is far lower than 49%. As they approach the tower/maintenance shed, they seem to hesitate, but then keep on rolling. I watched a Big Boy do this over and over until it was bone dry on water AND sand before stopping to service. Changing the sand and not water did something similar. So it appears that they are linked in some manner inside the EXE. Probably using the same routine for calculations.

Maybe then the downtime on the Big Boy and Challenger would decrease enough to show the sort of profitability they had when in real world service.

Nope. The problem is that the game devs set the weight of the Challenger and Big Boy so high that they are basically never profitable to use. They can make some profit sometimes, if you're lucky, but another loco will always be better. Loco weight is part of the calculations for fuel consumption, so their fuel bills are always prohibitive. That's something we intend to fix soonish.

I'm just starting to get the weight part into my head. I'm stalling on this "free weight" business. One of the things I first noticed about the weights was that the CaboA is far more heavy than a square 2-axle flatcar with a glorified privy on top. I didn't change anything, because I didn't want to alter the fuel model until I understood it.

So what is this "free weight" and where is the data in the files?

EDIT: Stupid question, I've been reading about the 100 logos limit and other limits. Has anyone ever tried compiling one PK4 with 100+ logos to see if that overcomes the limit?

[Gumboots]

Ok, first thing is I haven't played around with logos, so don't know about any limits on them. You'll have to ask someone else about those.

So what is this "free weight" and where is the data in the files?

Free weight isn't really a good name for it. However, I can see why it was given that name, back before anyone reverse engineered the formula for speed vs grade vs load. Unlike locomotive and consist weight it has nothing to do with fuel consumption, and only affects speed.

It's basically a scaling factor. What it does is take the output of the speed calculation for any weight and grade, and then multiply that speed by a factor that depends on the "free weight" value. If this results in a speed greater than the loco's nominal top speed, the speed is capped at that.

The effect of the scaling is to push the speed on 0% grade to the loco's top speed if the "free weight" value is equal to or greater than the total weight of the consist. So, if you have a consist of one 40 ton freight car, and a "free weight" value of 40, your loco will do its full top speed on a 0% grade, regardless of any other values set in the .lco file. This gives people the impression that for every car weight in the "free weight" value, you get a "free car" in terms of speed on a 0% grade.

However, the scaling factor applies on all grades, not just on 0%, and on non-zero grades it does not give you top speed for every car weight in the "free weight". This is probably what led to the myth that "free weight" only boosts speed on flat terrain, and that "pulling power" is what gets you up hills. It's not true, but the boost on 0% was obvious to anyone, while the boost on grades wasn't.

The truth is that both factors, "free weight" and "pulling power", interact to give you your speed for any grade and consist. The formula has eight variables in it, and is not at all intuitive at a glance if you haven't done the analysis, even though it's perfectly logical once you know the maths behind it.

If you want to know, for 4 cars on a 4% grade (just as an example, in Windows spreadsheet syntax) the formula for your speed is:

MAX(1;(MIN(1;(1.2239^(Free_weight/10))*(0.9675^(SUM(Loco_plus_tender)))*(0.9481^(SUM(Consist_4)))))^(1/Pulling_power)*Top_speed)

The first part MAX(1; just sets a minimum speed of 1 mph, which the game uses as a failsafe value to prevent divide by zero errors, and to make sure no loco is ever completely stopped by any grade or load.

The next part MIN(1; makes sure the top speed value from the .lco file is never multiplied by more than 1.

SUM(Loco_plus_tender) is just the total weight of the two combined.

SUM(Consist_4) is the total weight of the four cars in the consist.

The way the values are arranged in the .lco file is shown in the attached screenshot. In this case, the hex 00 00 66 43 corresponds to a "free weight" value of 230.

[Just Crazy Jim]

I wouldn't say it's "simple", but you explained it perfectly well. I am starting to see the "big picture" now. Thank you.

It might be time to update PJay's notes for this - the download version says:

Code: Select all

?1

for the free weight section of the LCO file.

Then:

Code: Select all

?2 (?1 , ?2 : sth with grade climbing)

for the Pulling power section of the LCO file.

[RulerofRails]

If you have found a repeatable bug where a train is skipping service stops please document it. We know of one such bug: if the service facility is placed within 6 "spaces" of a track junction. The game may miss it in one direction or perhaps even both.

I tested water and oil consumption by distance one time:

This is a little rough, but basically water drops 1% every 4.5 track miles.

Oil consumption is even rougher at 1% every 15 track miles.

So Water consumption is roughly 30% of Oil. When considering that a train will refill at a service facility whenever the level is below 60%, this makes sense. A train can run right out of water (not many miles without water) before it will normally stop for Oil. This provides some strategy in placement of service facilities. Spur servicing, or forcing the service stops (Ctrl+click to add them to the trains roster) can be used, but you can be clever without those measures (a normal game for me). Now bear in mind that this is normally a factor of distance, if consumption level is reduced via event there will be less opportunity to do this as the normal route is only barely long enough as it is. (If steamers are balanced against Diesels/Electrics, most of the strategy available may disappear, see below.)

A slight drop in water consumption would be good because then the train wouldn't actually run out of water before allowing Oil to be taken on.

Of course, if maintenance facilities are placed at random throughout your network, there's potential to have twice as many water stops as Oil stops. I think the designers were going for this.

Lastly, I don't think there's a wrong and a right in this, just a strategy situation for the player.

Steam vs. Diesels/Electrics.
Without placement trickery (at longer distances thanks to lower consumption figures, the typical routes of the average map are too short IMO), a steam loco will tend to do 3 stops (2 Water, 1 Oil) in the same time that a Diesel/Electric needs only one. Consideration must be given for sand, but most players chose to go round mountains rather than over them. In other words, it depends on the map and play-style as to how this figure plays out.

The stopped time for an Oil refill is similar to that for a Water refill (about 1/2 a month). However, placing the Water tower nearby (warning on the 1mph bug if they are butted hard up against each other, see first link in this post for extra details) the Maintenance shed belays some of the acceleration cost of a stop when both the Water and Oil are needed, but that is loco specific since top speed and acceleration rate as factors.

So, without an actual test, I hypothesize that the figure to equalize Steamers and Diesels/Electrics is somewhere in the range of 50-66% reduction in Water and Oil consumption on flat ground.

PS. Yes they probably should. Pjay's notes are ~10 years old. Nobody understood the loco figures too well (free weight and pulling power especially) until Gumboots deciphered it last year. There might be a little other misinformation as well.

[Gumboots]

I had thought a while back that one possibility for a scenario would be to give the player the option to eliminate water stops, at considerable capital cost. IOW, a pop-up dialogue that would let you build "water troughs" and fit your locos with "water scoops". Cost would be fairly substantial, since IRL these things were not cheap to set up for an entire network.

If water consumption was reduced to zero (or -100% in event terms) the actual water consumption would still be around 2% of default (based on some long term testing I did). So in terms of water alone, the range of the loco would increase by a factor of around 50.

[sbaros]

The tricky bit here, which I didn't mention, would be to use LengthPoint trickery to suck the single passenger car up inside the McKeen. As long as both were scaled so the McKeen was slightly larger, the visible result would be a lone McKeen but it would still be hauling one car of express.

I was especially delighted to dig out this post! For several months now I had been thinking of a variation of the LengthPoint concept for the simulation of self-propelled cars as follows:
Not to construct the actual railcar around the hauled car in order to hide the latter inside (with a negative coordinates trick) , but define a locomotive comprising only the windshield, the cab doors, a vertical or horizontal mini-boiler with smokestack for steam railcars, a pantograph for electric ones or an exhaust for DMU's, which would protrude over the carbody and create the desired impression. Non-steam options would become available after the turn of the century.
Now, as to the sort of the railcar, almost every conceivable passenger, mail and freight railmotor has existed at some time. My own employer still uses a couple of 1904 vintage ex-passenger EMUs converted to self-propelled low-side gondolas for night-time ballasting duties. Until the '40's we also had a couple of self-propelled tank cars for sprinkling (digging out the photos won't be easy). Even container-carrying ones are proposed!
I suppose acceleration and speed limits should be high on all gradients for single cars, more moderate for another trailer, and forbiddingly low for more than 2 trailers.
This could help us among other things in enhancing our periurban areas with local interurbans around large metropolitan conglomerations.
Well, if you set its consist to 0 cars with no caboose, you will see a phantom pantograph running around, but who cares after all...

[sbaros]

I heard somewhere that the early railways invented amusement parks and, in fact, seaside resorts generally, to attract passenger traffic. So the amusement parks were owned by the rail company and operated at a loss so the trains could make a profit.

I am not sure about conventional railways, certainly North American interurbans did that.

[Gumboots]

I suppose acceleration and speed limits should be high on all gradients for single cars, more moderate for another trailer, and forbiddingly low for more than 2 trailers.

I set up trial stats for a railcar, so that it went really well with one car load but was useless with two. It's quite easy. You have to make the "railcar" weight almost nothing (.car file) and then set pulling power and free weight to be just enough to haul one car at reasonable speeds on chosen grades. I probably still have those stats saved somewhere (will take a look). Due to weighing almost nothing it ends up being very cheap to run, but since usable consist is only one car it won't make stacks of revenue anyway.

Edit: I can't find the old stats. They may have been on a surplus spreadsheet that I deleted. I should be able to recreate them pretty easily.

[sbaros]

LengthPoint Issues

I was especially delighted to dig out this post! For several months now I had been thinking of a variation of the LengthPoint concept for the simulation of self-propelled cars

In conjunction with the steam railcar project, I took some time off this weekend to experiment and explore the LengthPoint concepts discussed in various posts using the "Planet" as a test bed. My initial target was to manipulate its LengthPoint so as to submerge the whole locomotive into its tender as in the picture.

Afterwards, I would eliminate the latter altogether, so that I'd have the "Planet" incorporated within the first car of its consist, before further alterations would be made. Nonetheless, this is failing so far. Actually, I hex-edited specific portions of the LengthPoint.3DP by guessing. More specifically, I edited those parts after the "INST" references that seemed to concern the X-coordinates. One of them is marked below.

I didn't mess at all with adjacent "cells", assuming that these refer to Y and Z coordinates, which would be better left alone.Playing with these values a bit with the aid of an online Little Endian Converter, I was able to deliberately separate the engine from the tender as far as I wanted.

Bear in mind, this may become of use for a separate project of mine, supplementing or replacing the military escort adaptation of the caboose concept. I attempted to apply negative values into these "cells" as I had successfully done with the negative weight of the "banking loco" concept. However, this could only engulf slightly the loco into the tender, no matter what values I tried, no further than the following picture shows.

Now, I am not sure what is going wrong. Am I putting the wrong hex codes into the 3DP "cells"? Does the program have some control routine "trapping" negative inputs for LengthPoints? Am I missing something about the LengthPoint.3DP syntax? Has anyone else tried such superimposing using negative LengthPoints? Are any other parameters playing any role, such as the TrackPoint? I wouldn't touch it either before some advice is at hand, so that I don't screw things further up.

[Gumboots]

You can't use negative values for the rear length point. See: viewtopic.php?p=39934#p39934

At the moment I can't recall if you can use negative values for the front length point. I don't think I ever tested that, but my guess is they wouldn't work either. I think zero is the minimum value the game will recognise.

Oh, and phpBB strips out one line break from posts in the editor. It's a known bug. To get a clear line between images and text, you have to enter two line breaks.

Edit: Also, your link closing tags are wrong. If you are typing the tags manually (which you seem to be) the closing tag needs a slash directly after the first square bracket.

In other words, this will be broken:

Code: Select all

[url=http://hawkdawg.com/forums/viewtopic.php?p=39934#p39934]No slash in the closing tag[url]

This will display properly:

Code: Select all

[url=http://hawkdawg.com/forums/viewtopic.php?p=39934#p39934]With slash in the closing tag[/url]