Finished: 2-6-4T Suburban Tank (revamped, and double header)

[Gumboots]

Ok, tis done.

One revamped-to-not-be-too-disgusting-to-use 2-6-4T Suburban Tank, in both normal and double header versions. This is still a bit beta, since I haven't yet got around to coding LOD's for the main loco body. That's currently running only one LOD, like most custom locomotives, which may cause a performance impact from the double header version if you use a lot of them. There are also a couple of minor details I'm not entirely happy with, and which I will take a look at over the next week. However, both versions work just fine and it's a nice old choofer, even if the actual graphics are only to default PopTop standard.

The single version has exactly the same stats as the old half-arsed-Class-500-running-backwards that nobody ever wanted to use because it was so ugly. I've used Lirio's update pack stats, since people generally think these are an improvement on the original 1.06 stats for overall game balance.

The double version has double pulling power, free weight, purchase price, and maintenance cost. Acceleration, top speed, and passenger appeal are the same as the single version. Fuel economy has been dropped from Above Average to Below Average, to simulate the extra fuel usage from running a double header. These stats can be revised, if necessary, after more testing.

---

Zips of final versions available in the archives: http://hawkdawg.com/rrt/rrt3_extras/rrt ... p#gumboots

---

Pix of the thing running live are below.

2-6-4T-1.jpg (101.75 KiB) Viewed 12714 times

2-6-4T-Double-2.jpg (83.91 KiB) Viewed 12713 times

[Gumboots]

And a few more pix, just for fun.

2-6-4T-night_front.jpg (88.7 KiB) Viewed 12713 times

2-6-4T-night_rear.jpg (77.9 KiB) Viewed 12713 times

2-6-4T-Double-1.jpg (88.97 KiB) Viewed 12713 times

[Gumboots]

Just took a quick look at how the default P8 LOD's are done. Should be easy enough to reproduce something similar for the 2-6-4T.

I won't use Blender's "Decimate" function because that makes a mess of the mesh. It's obviously good for modelling things like human bodies or whatever, but for locomotives and other basically geometric shapes it doesn't seem to be so good. Best way seems to be to just merge faces and delete stuff to get the next LOD down, which Blender will do while preserving the UV mapping for what's left.

Pix attached, showing first LOD with A skin, down to fifth LOD with E skin.

Class_P8_LOD_1_to_5.jpg (31.38 KiB) Viewed 12699 times

Class_P8_LOD_1_to_5_rear.jpg (25.6 KiB) Viewed 12699 times

[Gumboots]

I have the LOD's sorted now. Blender makes that quick and easy, even if doing them all manually. Now has five LOD's like most default models. Poly count isn't quite as low as the defaults on mid-range LOD's, but from what I've read about modern gfx cards this will make no difference to performance. The number of separate meshes seems to be far more important than the actual number of tris in those meshes, once you get down to the levels involved in these models. On that basis, I've just gone with what I think is reasonable.

I've given these some in-game testing last night, but not a lot. I actually found it hard to find a map which allowed the Suburban Tank anyway. The single version, since it shares the same stats as the old ugly one, should be fine. Double header stats seem ok on the limited testing I've done. I'll pack the things up again tonight with all LOD's in them, and post them up as "final" versions. If further in-game testing indicates minor stats tweaks would be good, those can always be done later.

Any feedback is welcome.

[Gumboots]

Ok, sorted. Both units now have all five LOD's, and have been tested to make sure they work.

Included in the attached zip are replacement PK4's for both the single and double versions of the 2-6-4T, along with the required .lco and .car files for both versions.

The files are already packaged in the correct folders, so you can install them by simply placing the zip in the main RT3 folder and extracting it right there.

Go and give them a flogging. Let me know if any improvements are desirable.

Zips of final versions attached to this post.

[Gumboots]

Oh yeah, here's the Blender file for this loco, so you can see how I've done it all.

If you check it out you'll see how I've cheated in several places to get the thing running and looking ok, without redoing all the graphics. A flasher job could be done with a complete new skin, but it's quite reasonable the way it is. Since this was an engine I didn't have a huge interest in and just wanted to de-bug a bit, I went for the easy option.

If you compare it with an imported P8 model the changes to the mesh should make sense too. Basically, the P8's tender has been merged into the loco body to form the coal bunker behind the cab, then a few other tweaks have been applied to get it all playing nicely as a 2-6-4T (ie: some mesh changes, and some remapping of existing graphics).

The reason the main bogies are rotated 45 degrees relative to the drivewheels is that for some reason this loco wants to randomise the rotation of the bogies when it loads in the game. I remember WP&P grumbling about this when he was working on his Royal Hudson model. I never had the same problem with my Berkshire, but for some weird reason it's happening with this one. At the moment I don't know why, since the drivetrain coding is standard stuff. What's more, the angle it borks them to isn't constant. It'll sometimes change after a station or maintenance shed.

I have thought of one possible fix, which is to change the coding to declare the four bogies as drivewheels. Since the current drivewheels are locked into correct behaviour, changing the bogies to drivewheels may sort them out too. OTOH it's not really noticeable once the wheels are turning anyway.

---

Edit by Hawk: File removed. It's now in the archives. - http://hawkdawg.com/rrt/rrt3/Xtras/NewE ... m#Gumboots

[DTL98]

Any feedback is welcome.

Slap on a pair of Witte smoke deflectors and shorten the stack a bit and you've got yourself a sharp looking BR 66! I'd think that you should make the BR 66 a separate locomotive, since it was introduced in 1955.
(I seem to recall you mentioning it not looking like any actual locomotive in another forum topic, but you sure came close!)

[Gumboots]

Yeah I did think of doing it as a BR66, but the introduction date was out by decades and it would have required more re-modelling and more graphics work. Going for a generic-but-unspecified 2-6-4T means nobody can tell me I got bits wrong.

[Gumboots]

Has anyone done some testing of this? Any further thoughts on it?

I've been between RT3 binges myself, but from the testing I did I think the single version is ready for the archives. It all works and, given the limitations of the default PopTop skin, it looks pretty good. Stats are the same as before, so it will be fine for game balance.

The only thing I did think of is that the double-header version may be better with the reliability bumped up one notch. Due to its greater hauling power it will probably be worked harder, and the game's hard coding will reduce reliability faster due to the heavier workload, which means that to get similar service life to the single version the double-header could probably do with a slight increase in the reliability rating.

[OddBox]

Is there a way of using this as a separate loco to the previous version?
I just like having as many different trains as possible running on my railroads!

[Gumboots]

Sure, you can do that if you want to. You'll have to do it yourself though, because it's not something I have an interest in doing. There's honestly not much point. All this version does is replace the old ones for looks, since the old one looked like crap. Specs and performance are the same.

[RulerofRails]

Included in the attached zip are replacement PK4's for both the single and double versions of the 2-6-4T. These can be dropped into PopTopExtraContent, and the earlier files removed. No changes are required to the EngineTypes folder.

Wanted to give this a try, but after removing the 1.06 "original" 2-6-4T I get a "Pack element < 264T_Profile.imb > not found" message when I try to load a map. I found only 2-6-4T.pk4 to remove, is there something else I need to do to make this work?

[Gumboots]

Was this a map that had been saved with the old 2-6-4 already in use? I admit I didn't test that.

I only tested on fresh starts, on maps that had been coded to allow the 2-6-4T already. It may have broken something if applied to a map that was saved with the old 2-6-4T already running on some routes. From memory I was pissed about the file naming inconsistencies of the old version and tidied them up some so they made more sense.

[RulerofRails]

What I am doing is starting a new scenario from scratch I was trying Canyonlands (doesn't have the original 2-6-4T ticked in loco selection) and another blank test map (created in 1.05) for kicks. Upon clicking "start scenario" virtually instantly the game crashes to desktop after showing the loading screen for a flash. Then I see the message I posted above.

[Gumboots]

Hang on, did you also upload the new .lco and .car files for the EngineTypes folder? Won't work if you don't do that. That's why I included them in the zip.

[RulerofRails]

I didn't see any .lco or .car files in the zip, just two pk4s. One single and the other dual. The post with the attachment says this:

No changes are required to the EngineTypes folder.

Which made me think that all was ok without them. That's probably the issue though.

[Gumboots]

Well that was dopey of me. Must have been short of sleep or something. The double version is a brand new loco, so obviously needs its own .lco and .car files to run at all. The single version would need its ones replaced too, which would be what is causing that crash to desktop.

I had the whole lot sitting in a folder named "Uploads", with EngineTypes and PopExtraContent folders inside that all ready to go, but for some reason I didn't zip the EngineTypes stuff with the PK4's, and I didn't put the PK4's inside a PopExtraContent folder (which is where they really should be, so you can just extract the zip straight into the RT3 main folder).

Bit funny that I posted the zip about 7 weeks ago and nobody has picked up on it until now, even though it has been downloaded 18 times. Goes to show how much people have actually bothered to try it.

Anyway, moving right along, I have repackaged the whole lot and replaced the old zip with a new one: download/file.php?id=3867.

Have also edited the post to give the right message. Try the new ones.

[RulerofRails]

Is there a way of using this as a separate loco to the previous version?
I just like having as many different trains as possible running on my railroads!

Just follow the new install directions and you will be able to use the original + the two new versions.

Try the new ones.

Thanks! Working perfectly now. Looks good as always from the Gumbootz Lokomotivfabrik und Bierkeller! That's half the reason that I didn't check it out, I know the modelling will be top notch quality. Only thing I could do is run it around for a few minutes to do a basic check (should have done that). I could try to throw some stat ideas around given the question on the fuel cost seting. I know testing methods are really discretionary, but here goes anyway.

---

Fuel Costs
The following stats were made on Expert difficulty WITHOUT Lirio's loco balance pack. (I uninstalled it awhile ago fearing it sometimes skewed the creator's intention.) In this test I setup a route on the Canyonlands map after DISABLING the start-up engine modifier events from that map. I setup a run from Panguitch to Glendale. I wasn't really careful with track-laying and had some smaller sections of 5% so that engines with a true mountain climbing ability should come out ahead. All tests were run with 8 cars before 1950 using 1.06's ability to keep the train always full with 2 Cheese and 6 Milk cars (PopTop version cars).

Test 1
Further details: The route was using 118% of water capacity. I was using only one water stop per trip between the two, with an Oil stop on the up leg. Effectively trains were without water for 18% of the journey so half of maximum speed during that time. This has no consequence on the second set of figures.

Averages over 3 years for fuel cost and speed attained in one year including in-line maintenance stops:

• New Tank: 29MPH, 42k fuel.
  New Tank (Double): 34MPH, 98k fuel.
  Northern: 33MPH, 143k fuel.
  
  One-year figure for the following:
  A1: 34MPH, 92k fuel.
  P2: 31MPH, 69k fuel.
  Mikado: 26MPH, 58k fuel.

Color guide: Fuel Costs: Normal = Average; Blue = Above Average; Red = Below Average.

---

Test 2
Details: A test without maintenance stops, replacing the loco instead of water stops. Simply a complete trip, up and back, stopping the loco upon completion and noting the stats. This was an attempt to get a better idea of fuel costs in a hilly enviroment over a fixed distance with the same load.

---

I realize that Lirio's stats for the Old Tank are what the new one has. After this test I took a quick look at some more of Lirio's figures, I am sure they have a narrower gap in price of performance. I took a quick look at the comparison and many of the other locos I tested have boosted stats. Will have to do the second test again with her stats.

The comparison between the single and double Tanks is unaffected. The Double Tank costs double in fuel and the maximum advantage in speed (second test) it could get on this up a hill where it should have an advantage was 30%! Go figure!

Additionally, in general I would say that fuel cost setting is very important. I haven't realized the seemingly major effect the pulling power setting has on fuel cost. There is a point of overworking the engines which was definitely the case with the Old Tank, Mallard, and Pacific. But even considering this, the fact that the Pacific is dead last gives an idea of how expensive an engine it is to fuel.

[Gumboots]

I must admit I gave the double version minimal testing. It was basically a case of just get it finished before going back to some other things I had to do.

I suppose it comes down to deciding how it's going to be used. Lirio was fanatical about getting a "smooth progression" of loco characteristics over the entire roster, whereas I tend to not care much about that. My preference is more towards making the relative performance of each loco a pretty realistic comparison to the others. IOW, the difference in hauling power between a Berkshire and a Mikado in the game should more or less match how it was IRL, and if that means no "smooth progression" then so be it.

So personally I'd only be worried about whether the single version had a relative performance that made sense compared to what an average 2-6-4T could do IRL, and the double version would obviously have double the grunt. I suppose you could not make the fuel rating worse, on the basis that each unit of a double would have an easier time of it than a single unit trying to haul the same load. That should help a bit. Could maybe boost the reliability one notch on the same basis. I suppose acceleration could go up one notch too.

I think maintenance cost and purchase cost should still be twice that of the single, so then you have to figure out where the double is going to be useful enough to justify the extra cost. It's only ever really going to make much sense on fairly heavy grades. Express on heavy grades would be one situation, since the extra speed would be handy there.

It's also probably only going to make sense in scenarios where you are restricted to using the 2-6-4T, or just want to use them because you think they fit the feel of the map. I don't have the patience to try and revamp every default and custom engine to give the 2-6-4T Double some hypothetical perfect slot to fit in compared to absolutely everything else.

[RulerofRails]

Test 3
Details: Same as Test 2, but with Lirio's stats. It also includes the Berkshire and the Schools class.

The Schools class joins the Red Devil in the exclusive rank of steamers with a Good fuel rating. Another note: I have kept the economy at normal throughout these tests. When an economic change newspaper comes up, I build a small piece of track which triggers an event to put it back to normal. Also, this route isn't a very hilly route, it has a moderate climb and then some gentler stuff. I don't want to give the impression that it has massive mountains.

---

Here's a little comparison of the difference between Lirio and 1.06 stats:

Note that the Daylight has been given a -2 level fuel cost bump. It's now among the top performers.

Comparing some of this to the free weight and pulling power in Lirio's spreadsheet is quite interesting. The Class Y6, has higher free weight and pulling power than the Challenger or Big Boy, but lower speed. Still, the extra speed is costing a lot with high pulling power and it ends up being more efficient to go slower at those large pulling power numbers, so the Class Y6 wins. Are 30% higher fuel costs worth 10% higher average speed?

---

Whoever made the double GP7 left the fuel cost and acceleration alone, instead changing the free weight, pulling power and maintenance (10k became 18k). I know these things are hard to model, but an increase in fuel cost and pulling power is really going to hurt an engine's score. Pulling power must be increased to give some advantage for forking out a lot more for a double. I don't see a sensible way that fuel costs can be increased in a double.

If it was up to me I would knock the levels of the single Tank down to a little above what it had in 1.06. Tank locos seem a passenger, light freight loco to me. With a double I would see it being a medium duty loco. I think it is meant to be a European design?

The American G5s also costs the same and in real had 41,000lbs vs the Tank with around 25,000lbs. If I am right, the engine cost is easily adjustable on a per loco basis vs. speed and pulling power which increase for all locos in use?

Having the G5s much cheaper than a similar performance double tank is probably the best that can be done. A map maker would need to bump up the price of the G5s if he wants to use both together in a sensible manner.

Acceleration is a close one. Very Fast acceleration seems a bit much, but it's likely necessary to entice purchase of the double.