Evil Genius at work

[Just Crazy Jim]

That last line sums it up. The emergent behaviour that can appear when you start messing around with RT3's industries is really quite something. It needs a lot of testing to make sure you have it right.

Re the Pen-y-Darren: I know you like it, but IRL it was never used for regular inter-city haulage. It only made one run, and was then converted to a stationary engine. IOW, ultimately you may have to accept that it's just too primitive to be a viable in-game locomotive.

Indeed.

And, yes, the Pen-y-Darren is starting to look like the worst of a bad lot of ideas. For now, I'll keep it as a place holder, but it very likely will be replaced by a horse or team of horses with all the Zero Era cars being some sort of horse-drawn wagon. I think I have found my source for ready-made models, although one has to take a rather "enlightened" view of copyright laws....

[Gumboots]

Best not to get too enlightened. You can always ask if in doubt. Most people are pretty good if you ask first. If they say no, there are oodles of other models online.

[Just Crazy Jim]

Oh, I know better than to bend copyright laws too much. And I know to go ask politely. I reckon in this instance, now that RT3 is under ol' Sid Meier's Umbrella Corp, there won't be too much trouble getting everyone to play along nicely with my dastardly schemes.

And, if all else fails, I am more than capable of making a metric tonne of simple models along those lines. To quote Patsy in Monty Python and the Holy grail: "It's just a model."

[RulerofRails]

On the run of the 1804 Go West Map in question, things were quite odd from the start. Firstly, I had no events manipulating furniture prices or production when the price on furniture went sky high. RNG had done things so that logging camps were rare on the map and quite distant from lumber mills (this later changed, but that lack of logs at the beginning became a key event), furniture factories (by my design) were all in the New England coastal cities. The cost of transporting logs to mills, then lumber to factory, then finished good to end market played some role in the high price. Also, the beastly slow rate of travel of the Pen-y-Darren loco (10 m.p.h. max) also must have played a role. By the time I was able to get furniture to New York from Boston the first time the price per unit was already over $300.

Ok. My thought at this point is that I hope there isn't a problem with the formulas pre-1830 which causes them to break in a similar way to "price breakout" bug when an event increases cargo price.

I would recommend some tests isolating one factor at a time to try to determine the cause. For example, one time try to start your map in 1830 to see if it behaves the same. For another test see if the speed of travel is a problem, try running the map with engine speeds around 40mph and see what happens. This might be a big hurdle for your map. Probably an important thing to get figured IMO, playability is at stake.

Questions:
You mention some cells being toxic for a cargo type. Is the price in those cells still higher than surrounding cells?
Where is the highest price on your map? Is it still at real demands (factories, warehouse, house, etc.) or do you see "breakouts": meaning highest prices in cells that have no demand for that cargo at all?

While all this is taking place, the cities are growing, each city gained a few more houses, some more than I anticipated. Clearly there is a direct relationship between supplying the demands of the Tycoonauts' little houses to city growth.

Can't recall reports of anyone doing a reliable test for this in the past. My feeling is that it's more the count of loads hauled in/out that causes this. In RT3, houses don't end up consuming that much cargo (more actually rots, see below). Houses are scattered amongst a fair few economic cells in the cities.

When cargo is dropped at the station, it will need to transit at the least one or two cells in most cases to reach most houses in that city. This takes some time. Only houses which have a supply will even start to consume the tiny 0.01-0.05 loads per year they consume. Within a year (sometimes less, depends on that city's size), another city on your network probably demands that cargo so it is essentially waiting for a train to a new destination. My view is that RT3 houses' most important role in the freight economy is as a "background" demand. They consume cargo too, just not that much.

Also, building industries in a city will increase it's "size" as indicated in "stars". Further speculation, but I think that if you build industries in a city, the game will view that city as less in need of a new building (house, etc.) to keep up with the projected growth figures that are controlled by sliders in the editor.

So, 1.0 furniture arrives at New York at ~$550 per unit, the load was being used reduced fractionally before reloading, so the load that traveled from New York to Albany was 0.7 according to the mouse-over and has a unit price of $558.

The cargo loss you were seeing in transit is due to cargo rot. That's controlled by position #28 in the cty file for each cargo. A side-effect of very slow speeds.

In all, without consideration of terrain features, the Go West map is pretty small. Even with the Pen-y-Darren max speed of 10 m.p.h. it's fairly easy to achieve equilibrium or stagnation by game year 25.

I don't think that's a game bug. Price stagnation would indicate that there isn't enough demand on your map for the supply available. Something that you should be able to balance for with the lsiders for seeding and growth, or perhaps use an in-game events to reduce production as the game progresses to relieve a glut.

Adding fish and salt, seemed great ideas too. Now, I don't know. Adding oil early hasn't proved to be as brilliant in practice as it was in concept. I think to make oil "work" I'll need to add a slight {demand only} for oil to lumber mills, tool and dies, etc. But following that same logic, I'd then need to add a small output of "oil" to meat packing plants to represent the lard/tallow that was used as axle grease until the modern era. And taking that logic one step farther, I end up adding "oil" as an output to sheep farms to represent lanolin and, if I added olive farms (I really did consider that), oil there too... Where does it end?

For a cargo to work, there needs to be more demand on the map than supply (not just more nominal consumption, but more locations of demand). As long as you can figure that out, the fundamentals for s cargo in-game will be sound. For sure, some concessions to reality will likely always be necessary to establish the right balance of supply and demand.

[CeeBee]

Definitely strange things happen from time to time. I'm playing the Texas II map and am hauling logs to a lumber mill north of Houston a bit. There are no paper mills on the map but pulpwood is hitching a ride to the railhead and then going by wagon to Austin where it appears that one cell which happens to have a house on it is demanding pulpwood from all over the map. I think the guy who lives there has invented the pellet stove and is just waiting for a patent to go through before starting a new industry

[Cash on Wheels]

Definitely strange things happen from time to time. I'm playing the Texas II map and am hauling logs to a lumber mill north of Houston a bit. There are no paper mills on the map but pulpwood is hitching a ride to the railhead and then going by wagon to Austin where it appears that one cell which happens to have a house on it is demanding pulpwood from all over the map. I think the guy who lives there has invented the pellet stove and is just waiting for a patent to go through before starting a new industry

Naw its those ports. They seem to suck up everything. You can deliver anything you want there.

[CeeBee]

you mean one of those invisible desert oasis?

[Gumboots]

No, he's talking about the ports on his map. Not sure what's happening with your pulpwood.

Anyway thought for Jim: if you do decide to go ahead with a whaling ship map, with horse and cart transport on land, making the track vanish from view is really easy. The track graphics are taken from rt3_DDSW.PK4/TrackTexture.dds for the basic track, and Infrastructure.dds for the electric track catenary, so you can just alpha out the bits you don't want to see.

[Just Crazy Jim]

Well, the after a 40 m.p.h. test with anachronistic Baldwins, the wretched top speed of the Pen-y-Darren is clearly a factor in the miserable supply-and-demand economy of the 1804 mod. Prices stagnate quickly in the vicinity of the factory and within a short time, the region goes toxic* for whatever cargoes are being produced. Running the map with the same conditions in with a 1840 start gave rather different results in seeding, so that I am increasingly convinced that the early date I am using may be part of the problem. My email exchanges with PopTop in 2004 did indicate "there were problems with the engine with dates before 1829". I knew to expect weirdness when I started this, but I'm pretty sure some of the things I'm seeing relate to that vague statement by PopTop.

* In this instance, by /toxic/ I mean the price gradient is less than $2 difference between any station on my network.

My references to a "toxic cell" in my earlier posts I mean exactly what it sounds like, the specific economic cell goes red when surrounded by yellow or green, when logic dictates it should be the same or a near color of the price gradient of surrounding cells. Further testing is probably necessary, but I've tampered with the files and map settings so often trying to smooth things out that even though I have managed to remedy this (for the moment) I cannot guess which if any of my changes was the fix, if it was a fix at all.

I'm approaching "toxic" on this myself, so I need to turn my head away from it a while.

But in recent edits of the map, I have seen two remarkable things.

1) even if I try to guess what will act as bait for the AI, how it goes about making the connection I think will happen is always complete and utter madness.

Example 1: AI company headed by George Train establishes first city pair between Albany and Lake George, no surprise there. I've seen it more than 50 times so far. Albany has a Tool & Die and a Furniture Factory. The Tool & Die is being fed iron by iron mines situated west of Albany on the Appalachian Plateau. What makes sense to me is a spur to Poughkeepsie, which is a 2.25 star city with a lumber mill and 2 nearby logging camps feeding the lumber mill. But no, after 7 years, the AI takes out 4 million dollars in bonds and runs a spur with 50+ mile tunnel to Oswego. 5 years later, the AI takes out an additional 2 million dollars in bonds to run a spur from Lake George to Poughkeepsie with 3 garbage bridges and the rails right over top of the Albany station's. Something that I tried in vain to replicate on a blank map, but the magnetic action of the rails would in no instance allow.

Example 2: AI company headed by Collis Huntington establishes first city pair with Allentown and Newton, 1 year passes before a spur to Raritan (Edison). 6 years pass, the AI takes out 2 million dollars in bonds and drops a spur to Bridgeport with the longest bridge I have seen in the game to date. All the while, New York is losing stars and houses, dropping from 3.25 (1804) to 1.5 (1818). In 1819, the port in New York self-deletes. 1820, the AI drops a spur from Newton to New York with tunnel and bridge. The same year, the AI takes out 6 million in bonds to run a nightmare connection from Bridgeport to Scranton (a 1 cheese town) with 2 tunnels and a bridge, bypassing Monticello. Literally nothing on the economic map substantiates such a link being built at that time. Meanwhile Hartford and Springfield have 3 stars each and are ignored.

Example 3: Back to George Train of the Albany & Lake George in 1827. Somehow, beyond all expectation, the AI has managed to turn a profit on this spaghetti map it's made and paid off its bonds. In rapid succession, a big ugly tunnel and ridiculous bridge appear on the map from Albany to Springfield, then a rail spur whips across the map to Newburyport then another back across the map from Newburyport with 9% grade tunnel to the 1-cheese town Keene.

2) Watching the little partial cargoes walking on the map, I can not with any confidence claim how much is being lost to "rot" and how much is being lost to "honest usage" by Tycoonaut housing. The cities on the 1804 Go West map are light on industry and heavy on housing by my design. I see very clear evidence that lack of cargoes stalls the growth of a city by merit of the fact that cities with no rail connection do not grow, whereas cities with a rail connection AND a steady supply of varied cargoes do grow. I am increasingly certain that the supplying of those fractional demands plays a critical role in the health and growth of a city. To this end, I made a non-city collection of six houses with a station midway between Boston and Newburyport. I made a route that ran Noname-Boston-Noname-Newburyport and let it run autoconsist for 10 years. 2 of the 6 houses upgraded to big houses. Express traffic from the Noname station was consistently profitable.

Lastly. I have spend a total of 6 hours 22 minutes in Blender getting done what took me 4 minutes in Milkshape 3D, then trying the same thing in AutoCAD, took me 6 minutes... I am old, and this dog doesn't have many new tricks left in him, so I am pretty sure Blender is not the tool of preference for me and never will be. I think I may have made a new locomotive model, but I can't be arsed to spend 4 hours figuring out how to get Blender to throw textures on the model. I am old... change is bad... mutter, mutter, mutter... why can't it ever be easy... whine, moan, belly-ache....

I need a beer.

[Gumboots]

I'm approaching "toxic" on this myself, so I need to turn my head away from it a while.

Give it up for a week and just go play trains. This is supposed to be fun.

Anyway, the stupid behaviour of RT3's AI is legendary. There's no way of predicting what they will do, except that you know it will be really dumb.

I see very clear evidence that lack of cargoes stalls the growth of a city by merit of the fact that cities with no rail connection do not grow, whereas cities with a rail connection AND a steady supply of varied cargoes do grow. I am increasingly certain that the supplying of those fractional demands plays a critical role in the health and growth of a city.

Hadn't noticed it myself, but then I hadn't been on the lookout for it. It's quite plausible that cities would be coded to operate like industry to some extent, and therefore supplying their demands would result in improved performance. I can easily see good arguments for including something like that in the game engine, so I suspect you are right.

Lastly. I have spend a total of 6 hours 22 minutes in Blender getting done what took me 4 minutes in Milkshape 3D, then trying the same thing in AutoCAD, took me 6 minutes... I am old, and this dog doesn't have many new tricks left in him, so I am pretty sure Blender is not the tool of preference for me and never will be. I think I may have made a new locomotive model, but I can't be arsed to spend 4 hours figuring out how to get Blender to throw textures on the model. I am old... change is bad... mutter, mutter, mutter... why can't it ever be easy... whine, moan, belly-ache....

I need a beer.

No problem. Use whatever you want. It can still be imported and exported via Blender without much drama. If you get the thing modeled and skinned and into a file format that Blender will import, I can do the rest for you.

I am 55 myself and didn't find Blender hard to pick up BUT I did spend that many hours (6:22 + 4) reading the manual, doing tutorials, and familiarising myself with the interface BEFORE trying to do any actual work. I did try it the other way (ie: just jump in and go nuts) and that simply didn't work.

Not that you have to use Blender. I'm just reiterating that point for anyone who does want to learn to use it.

I can't be arsed to spend 4 hours figuring out how to get Blender to throw textures on the model.

1/ Select object (right click)
2/ Set a material (right side panel, small globe icon)
3/ Scroll down to Transparency. Set Alpha and Specular to 0.
4/ Apply a texture (next checkered icon, to the right).
5/ Set Type to "Image or Movie".
6/ Set Mapping > Coordinates to UV.
7/ Scroll down to Influence. Set Alpha to 1.

There ya go. Model skinned.

[Just Crazy Jim]

Give it up for a week and just go play trains. This is supposed to be fun.

Anyway, the stupid behaviour of RT3's AI is legendary. There's no way of predicting what they will do, except that you know it will be really dumb.

Insane and really, really stupid. *nods*

No problem. Use whatever you want. It can still be imported and exported via Blender without much drama. If you get the thing modeled and skinned and into a file format that Blender will import, I can do the rest for you.

I am 55 myself and didn't find Blender hard to pick up BUT I did spend that many hours (6:22 + 4) reading the manual, doing tutorials, and familiarising myself with the interface BEFORE trying to do any actual work. I did try it the other way (ie: just jump in and go nuts) and that simply didn't work.

Not that you have to use Blender. I'm just reiterating that point for anyone who does want to learn to use it.

I can't be arsed to spend 4 hours figuring out how to get Blender to throw textures on the model.

1/ Select object (right click)
2/ Set a material (right side panel, small globe icon)
3/ Scroll down to Transparency. Set Alpha and Specular to 0.
4/ Apply a texture (next checkered icon, to the right).
5/ Set Type to "Image or Movie".
6/ Set Mapping > Coordinates to UV.
7/ Scroll down to Influence. Set Alpha to 1.

There ya go. Model skinned.

Aye, Sir Gumboots, 'tis true, it's supposed to be that simple. I read the instructions, watched a video, but things on my screen did not line up with what I read or saw. I had to resort to calling the college kid over to the house to see if I had gone daft. It took him about 20 minutes to figure out that the panel at the right side of the interface had to be resized to show any of the parts I couldn't find... In the process of finding that the panel could be resized, he cussed in a fashion worthy of his ancestors and only managed to set off the animation function twice using AutoCAD hotkey combinations. So, if the college kid whose stock and trade is GIS and CAD has an outbreak of Tourette Syndrome when faced with this open source darling, I don't feel completely stupid.

I won't share what he had to say about Blender and Gimp... Suffice to say, even I blushed at that bit of prose.

But after better than 30 minutes' miserable work, we got a texture to hang on the mesh...

[Gumboots]

Yer college kid is too hung upon expecting every GUI to be identical to the ones he usually uses. Why the @&@# would he expect AutoCAD hotkeys to work in Blender? If they don't work in Blender, that is defo a case of PIBCAK. If you use Blender hotkeys in Blender, they work perfectly.

It took him 20 minutes of swearing to figure out how to resize something? Tell him to RTFM.

https://en.wikibooks.org/wiki/Blender_3 ... ng_Windows

IOW, resizing can be done with any window, and works exactly the same way that it does on any desktop. You just click and hold on the border, and drag the flipping thing.

Anyway, although I haven't used it there is an option under User Preferences > Input to select options for 3dsMax or Maya. Presumably these options change the inputs to be the same as those applications. You could try those options if you are interested.

I've never used them because, not having used Maya or 3dsMax myself anyway, learning one interface would have been much the same as learning another, so I thought I might as well learn Blender's. And, for anyone who does have the time and inclination to learn how Blender does things, it really is very well thought out. It doesn't take long for things to become second nature and fast.

Regarding the notes on your screenshot:

1/ Getting the skin onto that model is about a one-minute-with-brain-in-neutral job, once you know how to do it. That includes starting from a bare default file and importing the model. It is definitely not the sort of job that will always take half an hour of misery.

2/ Speaking of which, that little useless box that is your mortal enemy is simply the default cube model that is in every bare starting file. You can see it labelled as "Cube" in the overview panel at upper right, just above where it says "Elephant".

You want it gone? Easy. Select it (right click) then go find the "Delete" button on your keyboard. Click that there button. Guess what happens?

Oh, you deleted Elephant by mistake? No worries. Hit Ctrl+Z. Elephant will come back.

You can also delete the default camera. It's of no real use for what we are trying to do.

3/ Rotating the model. It's pretty straightforward. The 5 button on your numpad, over the the right of your keyboard, toggles you from ortho to perspective and back again. The other numpad buttons are all hotkeys for various views.

The 1 button gives you the front view. Ctrl+1 gives you the back view. 3 gives you left side view. Ctrl+3 gives you right side view. 7 gives you top view. Guess what Ctrl+7 does?

2 and 8 do rotation up or down around the horizontal axis. 4 and 6 do rotation around the vertical axis.

9 spins the model 180 degrees around the vertical axis.

Or if you want total freeform, hold down Alt+LMB (Left Mouse Button) then drag your mouse to spin the model anyway you like. You can also use Shift+Alt+LMB to pan in any direction.

If the model is incoveniently placed, you can use the period . button on the numpad to centre the view on the selected object and zoom in on it.

3/ The Alpha setting for texture set to 1, in combination with the Alpha for material being set to zero, will allow transparent sections of the skin (like the gaps between wheel spokes) to be transparent in the viewport. If you are just looking at a loco body skin that has no transparent areas anyway, it makes no difference. Sooner or later though, most likely sooner, you will be grumbling that your wheels look solid and why is Blender doing this? I gave you the settings to avoid that.

The other settings are ones you don't have to worry about unless you want fancy renders. Ignore them. The only thing I usually bother with is turning down the specular on material. I find knocking it back to 0.1 or 0.05 works well with the basic lighting I use.

4/ Importing all the widgety parts at once (mass import) would be cool, but unfortunately the bloke who wrote the import script for us didn't include that option. That's not a Blender fault. It's down to the custom script. I do not currently have the Python skills to code a mass import for the custom script.

[Gumboots]

Oh and change your bottom window (the one that is almost collapsed) from Timeline to UV/IMage Editor. How? Click on the clock icon in the lowest toolbar to bring up a list of options. Select the one you want.

Edit: And while I think of it, the standard Windows delete button does deletion, but often you just want to hide something. That's what the H key does. H for Hide. Clever, hey?

Don't want it all hidey anymore, and want it to come back where you can thump it? Use Alt+H. Hey presto! Bunnies jump out of hatz.

[Just Crazy Jim]

Let me stress the problem here is not that Blender is bad and evil and wrong. It is that the college kid and I both have spent years using AutoCad. Almost 20 years for me, about 11 for him. In effect, using Blender for me is not "learning a new interface" so much as it is "unlearning everything I know about using another program in which I learned everything I know about 3D modelling (part 2: electric boogaloo)". I already went through this once stepping from AutoCAD to Milkshape 3D.

Blender is not evil for having a trillion new hot key combinations to learn, it probably needs every single one of them and more. Lord knows, there are an untold number of AutoCAD functions I have never used, not even once, and likely never will. My problem is (more or less) being confronted with a dazzling array of information about Blender functions that are useless to my endeavor and only get in the way of understanding only what I need to know to do what I want to do. AutoCAD was very much the same way. Too much program for what I wanted to do.

All that being said, until I can set aside the time to get a bit less stupid, I will make my models in software I know, then use Blender to make the work in RT3 because I have no other choice

[Gumboots]

That's fair enough. I understand the muscle memory thing could be problematic when trying to use something different. I've also done 3d modelling in Delftship, which uses different conventions to Blender, and sometimes find myself getting them mixed up. Then I remember "Oh yeah, that's for t'other one."

I have deliberately not tried to learn everything about Blender. That way lies madness. I have only learned stuff that will be directly useful to me for RT3 modelling. This cuts the load on the brain quite substantially. And every so often I stumble upon something else useful, which is even better.

But sure, use whatever you're comfortable using. It if works, it's all good. If you get the mesh sorted and the UV's mapped in Milkshape and/or AutoCAD then there's no need to screw around with 99% of Blender. All you'd need to know is the import/export options, and how to select an object for export.

[Just Crazy Jim]

A few days of hammering on my pet project in play-mode and I've encountered a few of the bugs that I've read about, but never personally experienced. The game seems robust and stable... until it isn't.

Bugs and more... Then there's events in the as-shipped-from-the-factory Go West map that increases and decreases the prices of munitions and weapons... before the cargoes are even available... traces of rivers and lakes that obviously weren't cleaned up well enough and act as impassables to track-laying. Strangeness in the GMP's height map causing a 20-mile-tall wall at the north edge of the map during some saves. CTDs when bulldozing things on certain (seemingly) random pixels of the map. Random CTDs just any time for any reason, reload and run from the last autosave, pass by the date/time events of the last CTD with no repeat. Just more seemingly random weirdness.

After giving the in-game economy a good hard shaking. I've decided to "economized" my project to a less than "whole game" mod. The game economy is just to simple and fragile to go monkeying about with changing established patterns beyond the most fractional adjustments, which can mostly be achieved with game events. Still, I wouldn't hate adding a couple of buildings to dress out the early 1800s look and feel, but they'd be cosmetic features if I am honest with myself. Adding a Cooper's Shop (Barrel Maker) to make Barrels that are used by other industries as containers for finished cargoes could more easily be achieved by adding a demand for lumber to existing transformations. Adding a Fish cargo could more easily achieved by adding livestock to a port's output.

And then there's my whale oil fascination/fixation... adding oil early was easy and did no harm, but the slight demand for oil as set in the game is so small that even a 5-star city barely needs one full unit of oil per year. This points to the very "we didn't live back then so we have no idea how much people were willing to pay for it" mindset of the developers. I mean, come on, guys did not get on a ship to sail from Nantucket around the Cape of Good Hope to spend five or more years gathering whale oil because it was "oh so much fun". In modern terms, a quart of whale oil sold for about 6-months wages for a working class stiff. There is a reason that once kerosene became available at 10-cents a gallon (about 1/10 a day's wage at the time), no one bought whale oil any longer. Anyway, that's a non-point. I added early oil, yay, it worked, whatever.

The thing that has turned shiny again is the Pen-y-Darren. The 25-year time-limit on most scenarios seemed a hurdle, now it seems a godsend. Start with the Pen-y-Darren, arguably the worst locomotive to ever appear in RT3, then spend 25 years and millions of dollars to improve it in such areas as events will allow until it the player arrives at the Planet. I think RoR suggested as much in a post a couple of weeks ago.

So, early oil? Yes.
Pen-y-Darren? Yes.
Everything else? NO!
Why? I don't see the benefit of spending a year making a whole game mod for just one scenario, maybe 2 or 3 at most.

Now I have locomotives to model. God help me.

[Just Crazy Jim]

You may be a mad RT3 modder if you make a caboose into a locomotive and reverse the model of a locomotive and make it a caboose just to see the trains all running backwards....

[Just Crazy Jim]

Finally settled on a beastly locomotive for 1804 that isn't the Pen-y-Darren: John Fitch's 1798 locomotive model as preserved at the Ohio State Museum. Just clunky looking enough to get across the point that it needs a good deal more development, which will be carried forward by events in the scenario. A somewhat romaticised version of the tale at this link and another at this link and a another at this link. The fans say he demonstrated the model to George Washington and his cabinet in an effort to secure funding. The romance version differs a bit from the facts, but no funding was gained as the US government was skint at the time. Another niggling detail is that George Washington wasn't president in 1798, that job-holder would have been John Adams. Any way, the model is real and dates from roughly 1800 +/- a bit, but definitely predates Fitch's death in 1798, so maybe he did show it to Washington and his cabinet, just not in 1798.

At the moment it's missing a few details, chiefly huge, billowing black smoke and a trickle of steam. And offset drivers, I am guessing they're drivers or maybe a driver and a flywheel, guessing because the 219 year old model is missing some of its original parts and isn't in the best of shape. So guessing is the best one can do in its current shape.

[Gumboots]

Interesting find. I hadn't heard of this one.



I think you have it back to front. Reaching around the grasshopper linkages* to throw coal in the firebox would be a dangerous exercise in gymnastics. My bet is it was intended to have the firebox and boiler at the rear, with the crew riding on the tender.

I think the "offset drivers" are flywheels that are geared to the actual drivers. My guess is they would have their crankpins just outboard of the inner framing.

*It's pretty obviously a sort of walking beam engine.

[Just Crazy Jim]

You're probably right, but my positioning was based on the idea that the bent stack (shown in the following images, not present in the image you posted) was aimed at the rear.

resizedimage110151-engine.jpg (4.68 KiB) Viewed 8052 times

But then, it may have been turned to the position shown in the images for reasons of crating or display within the confines of a glass case. Also, there is the fact that is it not a full scale locomotive and the design follows the fire being fed from the side on a small circular track. If it had ever made it to the full-scale model, the design would very likely be firebox at the rear.

Either way, re-positioning the smoke point is proving less easy that I like. Not even getting a Bong of Doom, just bang! crash to the desktop. And, yes, I read the manual and watched the video and scoured the Blender forums. I am guessing (blind stab in the dark) that the smoke point is one of those bits you have to manually edit with a text editor. I looked at the wiki for the 3dp file format (link), which was somewhat enlightening, but having no clear guide to which is X or Y or Z and it not being clear from the wiki if the values are floats or 32-bit integers, I didn't dabble.