Australian steam locos

[Hawk]

Yep! That's a bit better.

[Gumboots]

I'm getting carried away with this stuff, and should get back to doing something more sensible.

So there was a serious NSWGR proposal to build a 4-6-4+4-6-4 express Garratt. This was proposed back before they decided to build the C38 Pacifc's. The idea of course being that it would be able to run on any track in the state, unlike the C38's which were limited to a few main lines due to their axle loading. Also, Garratts wouldn't need to be turned around in small towns or whatever. Their symmetry means they could run backwards at full speed. The whole idea made sense if you were looking for something more flexible than the C38's they ended up building. You could not only get a lower axle loading, but you could get more grunt too, so potentially the Garratt would be faster and more versatile.

Anyway they never built them, and there are no drawings of the proposal anywhere. There is one model that was made back around that time, but that's all. I'm not all that keen on the model, so thought I'd play around with some ideas of my own. So this is based on standard NSWGR running gear (they used 69" drivers for express, etc) and is smack on the 100 RT3 unit limit for wheelbase length. It's worked out to be a feasible unit that would have around 45,000 lbs tractive effort at a 16 ton axle load. That should make it good for an easy 90 mph with full express loads.

Makes sense to give it some nifty styling, since those Victorians and South Australians had fancy trains and there's no way they could be allowed to score more street cred than NSW. The way NSWGR worked, the streamlining would have to have been fairly simple to build and cheap to maintain. Lotsa chrome trim would have been frowned on. Fairly easy access for maintenance would have been required too. Would also need good visibility in both directions, along with good coal and water capacity for long distance.

So anyway I got enthused about this whole idea and spent some time modelling it up in Blender. Verts + polys count is still decent (currently about 2600). IMO this is something they could have built with the technology at the time, and within the practical usage limits for an NSWGR locomotive.I think it looks pretty good.

#AC39_front.jpg

#AC39_side.jpg

#AC39_rear.jpg

[Hawk]

That's a cool looking loco.

[RulerofRails]

Looks great! If you're really into this, how much would it take to skin it up for the game?

I'm not a modeler, but I can imagine spending a lot of time getting heaps of small details right. That would become not-fun. I'm just trying to say is there a way that some of the small stuff can be sacrificed in order to get a decent-looking in-game model before the whole thing turns boring and nothing gets finished.

[Gumboots]

Umm, yeah. That's the perennial problem innit. Guess what happened with that !!%@#$# Schools class. Tricksy complicated little thing, that one. Oh it only has eight wheels, he said. How complicated can it be, he said. I'm going to have to get ruthless with that thing if I ever want to finish it. I sort of detailed myself into a corner and then went bonkers.

So this is why streamliners are good. Streamlining hides all the doohickeys that would otherwise chew up polys and time, and with the Blender script available the geometry for them isn't a big deal. That's also why I'm keen on locos like the LNER P2. Very clean and easy to model.

This Garratty thing can be as simple or as complex as I want it to be, because nobody can tell me bits are missing or something is wrong. I'm designing the critter, so I gets to say what's right. It's largely sorted already for the actual modelling. Needs a few more bits, but nothing over the top. Skinning it shouldn't be too bad either.

---------------------------------------------

Edit: Oh and this is why I'm kinda developing a liking for the little old DX Goods too. Having roughed out the model to a pretty decent level I can see there's not going to be too much work to get it all sorted. It's quite cute of its type when it's modeled correctly. Nothing like the current 1.06 horror. Simple too. Inside cylinders and Stephenson valve gear means very little crud on the outside.

[Gumboots]

Threw a few more pieces on it. I had a brainwave about the headlights. I was using a 12-sided cylinder for the headlight casing and a basic 12-sided face for the lens. Then I thought that since the headlight casing is mostly buried in the streamlining, with only a thin rim showing, I could change the cylinder to a cone and halve the number of verts. Works really well. I even ditched the bottom two faces of the cone, since there's so little that I couldn't tell the difference at the bottom. This knocked off 11 verts and 14 polys from each headlight, without making it look any rougher. So for the two headlights there's a total reduction of 50 verts + polys just for this simple change.

I've been pruning it a bit in other places too. This has enabled me to keep the total count to 2800 so far, even though I've now added handrails and steps to the model.

AC39_more_bits.jpg

Update: Added more features and trimmed more verts and polys. It now has steps all over the place, and poly+verts count is down to 2740. I'm thinking through this thing as if I'd have to use it IRL. How do I climb around it, and other things. The design details are developing in line with all of that. My aim is to keep the level of stupid low enough that I can't spot it myself. This should make the thing more satisfying to use in the game, if it ever gets that far. It'll feel like a real locomotive instead of another Free Willy.

AC39_steps_everywhere.jpg

[Gumboots]

Had some more fun with this, and learned something about problems with the AD60 Garratts. There's a thread with information here: warning - contains Australian railwaymen's language, some of it coming from blokes who worked on these things back in the day.

The short version is that the AD60's, as delivered to NSW, had three problems. These problems didn't prevent the class from doing a lot of useful work all over the state, but could have been easily fixed if the powers that be had consulted with the people who had to fire and drive the things. If they'd all thought it through before ordering the class, the result would have been much better. The problems were:

1/ Access to the cab. With the AD60 being built to the limits of the loading gauge, and with the access being from the side of the cab (the rear was closed in solid), if the train stopped for whatever reason in a single lane tunnel then you were stuck. Couldn't get out the door. The only way out was to try and crawl through the cab roof hatch, which of course was pretty much jammed against the roof of the tunnel. Naturally this was viewed as a major safety problem. The class ended up being banned (by the union) from working any routes that had single line tunnels, which reduced its range of operations, which is pretty counterproductive for a loco which was specifically ordered so that it could haul anywhere on any line.

2/ Heat in the cab. With it being mostly closed in, and with the steam pipe for the rear cylinders running beneath the cab floor, things got mighty hot in the cab in summer.

3/ This was often made worse by poorly maintained expansion joints (also under the cab) on the steam pipe, which gave the crew a lovely steam bath. Not bad in winter, but not good in summer.

So fixing the access to the cab and the ventilation would have been easy. Plenty of the African Garratts had cabs open at the back, and the Rhodesian Class 15 even had doors in the front of the cab so the crew could get straight onto the running boards if necessary. Naturally these were also great for ventilation if you wanted some extra.

The steam pipe problems should have been solvable with a bit more thought. More insulation for a start. Better design and/or maintenance on the expansion joint. Maybe relocating the expansion joint, if possible.

Anyway, after reading about all of this I figured I should redesign the cab on the mythical Australian express Garratt. So I did. It can now be fully closed in for running in either direction in nasty weather, or fully opened up so you don't fry in summer. If I'm going to invent an Australian express Garratt, I might as well do the job properly.

#AC39_Mk_4_cab_front.jpg

#AC39_Mk_4_cab_rear.jpg

[Gumboots]

Done a bit more. Added in the ashpan and main steam pipe. Also threw some niftier wheels on it, which look much like the final ones will. Wheels are currently modelled in 3D so are far too high-poly for the game skin, but good for the render. The final ones will just be done the normal RT3 way.

Anyway it's currently sitting bang on 3,000 for verts+polys, which is where I want it. Needs a little more detailing, but I have some temporary stuff on it at the moment so final count wont change much.

AC39_ashpan_and_steam_pipe.jpg

[Hawk]

That's coming along nicely.

[Gumboots]

Hey I played around with the idea of faking valve gear. Walmarts gear instead of Walschaerts.

So I can make it do this easily enough:

#AC39_valve_gear_test.gif

That's not bad for a bit of fun.

So the eccentric crank (see here for names of all the bits) on a real loco trails the crankpin by 90 degrees. I can get it doing that in RT3 coding, but it means the eccentric rod and expansion link have to be done as one sorta L-shaped piece. Don't ask why. It just does. RT3 coding is like that.

Anyway, this means that both move as one item, as you can see in the animated gif above. If you look carefully it's obviously fake, but it may be good enough for a bit of fun in the game, just to get a few more bits moving around.

There's another way I can fake it, which is by having the eccentric crank either 0 or 180 degrees from the crankpin, instead of the real 90. I haven't mocked up an animation of that yet but will get onto it. Doing it that way means the eccentric rod and expansion link could be done as two separate units that are hinged to each other, as they are supposed to be. It would also allow doing a very basic animation of the radius bar and valve stem, although for RT3 purposes these would have to be all one piece with the expansion link (in RT3 terms it would all be a piston, and the eccentric rod would be a connecting rod).

And with either of these systems the combination lever is a lost cause. The only way of animating that within RT3's limits is to make it part of the main piston.