• Welcome to Westlake Publishing Forums.
 

News:

    REGARDING MEMBERSHIP ON THIS FORUM: Due to spam, our server has disabled the forum software to gain membership. The only way to become a new member is for you to send me a private e-mail with your preferred screen name (we prefer you use your real name, or some variant there-of), and email adress you would like to have associated with the account.  -- Send the information to:  Russ at finescalerr@msn.com

Main Menu

Rewanui - a 1940 West Coast NZ layout in 1:64

Started by Lawrence@NZFinescale, February 08, 2021, 08:47:25 PM

Previous topic - Next topic

Lawrence@NZFinescale

Hi Paul

I'm no Blender expert and my PC is into it's second decade and not particularly whizz to start with.  I would say the coal stage is about the limit before I'd need to start looking for workarounds on processor etc.  The print software starts to choke on stls over 0.5GB too.  Up to that I'm fairly cavalier on poly count, but near that level I need to start being a bit more efficient.  On the other hand such limits are on parts not final models, so not all that limiting in practice.  The coal stage is in 2 printed parts.

In terms of the texturing I was relatively crude. There's not a lot of point in generating a digital model that is better than the printer's output resolution. A member 50mm long is 1000 printed pixels. 50mm is over 10' in 1:64 so it wasn't hard to find images of planks at sufficient resolution online. I used two pics, one of which is shown (I did convert to gray scale and tweak levels in photoshop) [update: not shown as it is 370kb :-(].  The mid grey area was added as a place to park bits of the uv map that I didn't want textured. For larger scales, you might want a bit more resolution, although you start running out of print bed.  I'd be trying to avoid repeats for wood grain.

CAD to Blender is problematic.  Choices seem to be DXF (giving vertices and edges only) or stl.

So yes, there is a bit of retopology.  However timber structures are often quite simple (this one sure is) and since most elements are simply a stretched cube I mainly just redrew and copied (the uv mapping needs redoing for each unique element though).  I also cheated a lot with repeats, untextured backsides etc.  This was more than laziness as it is a bit slow and there is a need to keep the file size down.

Blender has very different controls and workflows to a CAD package and is a steep learning curve (for me anyway).  The UV unwrap/map process is very straightforward once you know how though and pretty quick with the right workflow.

Megascans looks nice.  As a means to do brick and stone, could be quite exciting.
Cheers,

Lawrence in NZ
nzfinescale.com

WP Rayner

Lawrence:

Yes, it's amazing how quickly polys can accumulate, especially when retopologizing. Since you're basically only working with elongated cubes and flat panels, creating UVs is a simple process. As soon as you get into more organic or complex mechanical forms, it becomes something else indeed. As you say, it's definitely an advantage that you are working in small output sizes, governed both by scale and your printer. File size can quickly become an issue. A couple of my more complex rendering and animation projects had over one million polys, file size over 1gb, which even with the liquid-cooled 32 core machine, caused it to chug and complain on occasion. Megascans is a great resource. I used them quite a bit, especially when creating weathered surfaces in renders. They have an excellent range of brick, stonework, and concrete textures, most available in a range of LOD.  Would be interesting to see if any of those worked in a 3D printing workflow.
Paul

Stay low, keep quiet, keep it simple, don't expect too much, enjoy what you have.

Lawrence@NZFinescale

Paul

It's tempting to try some stone or brick for the hell of it, but I'm trying to focus.  My learning curve on this was a Napoleonic battle standard.  These were heavily embroidered, which I copied from a museum image and used to texture a draped flag, then printed.  VERY effective, but how my customer painted it I have no idea  ;).  Of course a plane is the simplest possible model.

As it is generally useful to break down models for both printing and painting the size issue almost takes care of itself.

The next step is to finish the vat and coal stage models.  Thereafter I may try this on a wooden wagon as I have the base model already as well as the need to do them in 1:64, 1:48 and 1:33.

Cheers,

Lawrence in NZ
nzfinescale.com

Ray Dunakin

Visit my website to see pics of the rugged and rocky In-ko-pah Railroad!

Ray Dunakin's World

Lawrence@NZFinescale

Inspired by Volker's valves and the generally high standard shown here, I've tried to up my game.

Here's a progress shot of the 6000 gal vat for Rewanui.  This was government railway (albeit a long way from head office) and in the 1940 period maintenance would have been quite good.  So I've aimed for work-stained, but well maintained.

This is basically a number of 3D prints with some real timber/ply and brass/styrene.

The vat (with hoop tensioners) is 1 piece (a hollow tube).  Other separate parts are the tank base and the transverse/longitudinal tank support timbers.  The entire support structure is 1 piece (verticals, bracing, fitting, concrete and the short ladder).  All this was done in CAD.  The hose was modelled in Blender using a fabric simulation draped over the CAD model. There's a clamp holding it onto the spigot.

The vat hoops are 0.3mm monofilament fishing line. They sit into shallow recesses printed around the vat.  Very easy to get tight, square hoops this way, and they don't look 'moulded on'.

Pipework is brass tube over brass wire.  The 3d printed valves and fittings are also threaded onto the wire giving a surprisingly robust construction. (Handles on the valves are 1.6mm diameter).

The gauge is printed - the graduations are done as 0.1mm recesses and filled with a wash during painting.  The wire to the indicator runs over a little wheel at the top and is 0.15mm nylon (invisible thread) blackened with a marker pen.

Still to go are the ladder, operating gear, 'water', a cover for the drain and the float valve assembly.  The ground will eventually cover most of the concrete foundations, hence the ragged look.  I used talc in the paint to provide texture on the concrete, but fear I overdid it.  As this will be largely hidden I'll probably just resolve to do better next time.

The vat is based is based on the standard drawing that would have been supplied as part of the contract, supplemented with photos as the thing was 30+ years old as modelled.  Of course some of the smaller details (particularly at the top and inside) are interpolated as information is incomplete. 

Enjoy
Cheers,

Lawrence in NZ
nzfinescale.com

Hauk

Quote from: Lawrence@NZFinescale on April 13, 2021, 01:56:03 PM
Inspired by Volker's valves and the generally high standard shown here, I've tried to up my game.

This sums up the value of this forum quite nicely.
Your water tank inspires me to up my game!
Regards, Hauk
--
"Yet for better or for worse we do love things that bear the marks of grime, soot, and weather, and we love the colors and the sheen that call to mind the past that made them"  -Junichiro Tanizaki

Remembrance Of Trains Past

finescalerr

What I'll never understand is how many people find the modeling on this forum intimidating. Should I conclude, then, that such guys also never look at a beautiful girl? Very confusing. -- Russ

1-32

Hi Lawrence.
That is a lovely model very well painted but the water spout is what really got my attention.
cheers Kim

Lawrence@NZFinescale

Quote from: 1-32 on April 14, 2021, 12:24:36 PM
Hi Lawrence.
That is a lovely model very well painted but the water spout is what really got my attention.
cheers Kim

Thanks Kim

The hose is very easy to do with software.  Learning the software not so easy.  It is done in Blender using animation tools. I'm no expert at this but I can get it to work eventually.

1. Import (from the CAD model) the elements of the vat that the hose needs to interact with (ie the spigot, horizontal beam and platform edge).
2. Create the hose as a horizontal tube attached to the spigot.
3. Assign the properties of a stiff material to the 'hose' and set the interactions with the other parts (ie pin the hose to the spigot and make it collide with the platform edge and beam)
4. Simulate - The hose behaves more or less as a real world hose would, staying attached to the spigot and falling under 'gravity' to drape and fold over the obstacles.
5. Tweak - adjusting all the parameters above to get the effect you want takes time and the simulation is not that fast either.
6.  Once happy 'bake' the result, add a clamp and join hose, clamp and spigot into a part for printing.

It is amazing what can be achieved with software like this and really neat to be able to print the results.  I recently played with a 7mm tent for a fellow lurker here. However, it is a different hobby (or aspect anyway).  I enjoy it, and particularly the way I can connect the virtual to the real to create something special in the space between them.  If you don't enjoy time at the keyboard I wouldn't recommend it.  If you do it opens some doors to things that would be very difficult to do any other way.
Cheers,

Lawrence in NZ
nzfinescale.com

1-32


Lawrence@NZFinescale

Quote from: 1-32 on April 14, 2021, 02:14:22 PM
yes, the seam on the pipe is special.


Thanks.  Things like the seam are little tweaks that are super easy to add into the digital model.

Appreciating models is a little like theatre - you have to suspend your disbelief a bit.  It follows from that that the more consistent the model is the easier that is to do and that how convincing the model is overall will be determined by the least convincing element.  I've done vats before and folded hoses from lead foil.  Not bad, but definitely the least convincing bit - so this is an improvement.

Lawrence
Cheers,

Lawrence in NZ
nzfinescale.com

Ray Dunakin

Visit my website to see pics of the rugged and rocky In-ko-pah Railroad!

Ray Dunakin's World

HelgeAndreas

Hello!
This thread was unfortunately new to me..
Outstanding work. Unfamiliar techniques to me too, but very convincing.

H

finescalerr

Lawrence, you really must become our resident 3-D CAD and printing guru. -- Russ

Lawrence@NZFinescale

Quote from: finescalerr on April 15, 2021, 12:15:51 PM
Lawrence, you really must become our resident 3-D CAD and printing guru. -- Russ

I'm not sure about that, but happy to help.

Half the battle is knowing a thing is possible.  Finding out how is usually just a Youtube tutorial away...
Cheers,

Lawrence in NZ
nzfinescale.com