Rather than add another digression to the layout thread, I thought I'd start something new...
While I'm fond of scratchbuilding, I do most of my models as kits as I:
- often want more than one
- can take advantage of low volume manufacturing techniques
- can fund development and production via sales
The driver for me is better models. In this installment I'll talk about a hopper wagon kit I've been developing. I've spent entirely too long on this, but it has been an interesting exercise integrating established and new technologies to produce something that is extremely detailed and accurate, performs well and is easy to build.
There are three technologies that I currently use extensively: Investment casting, 3D printing and etching.
Castings are heavy, robust, reasonably accurate and quite well detailed, but are expensive and can require some clean up.
3D prints are accurate, quick to produce, surprisingly robust and extremely detailed. They are light and exposed detail can be vulnerable. There are some technical limitations that need to be worked around too.
Etches are accurate, moderate in weight, and robust. They can accurately represent sheet metalwork and provide excellent frameworks.
These techniques, coupled with wire, tube and other base materials, provide a flexible toolkit for kitset development.
I was after something with internal hopper detail, moderate weight, relatively quick/easy to build, with suspension and free running. I need at least a dozen of these and I know from experience that building the same model repeatedly quickly pales if the process isn't quick and relaxing.
In this first post I can foreshadow the end result. In some follow ups I'll show how I used the various techniques to capitalise on their strengths and illustrate how it goes together. This one is complete other than fitting side chains.
That is exquisite!
Relatively quick and easy to build? Hmmm. Please don't tell us you can whip out a dozen of those beauties in just a couple of hours. -- Russ
May be a stupid question: As I can make out, there are printing layers visible on the inside of the hopper. To my knowledge, they are difficult to remove, the more there are also a lot of rivets all around...
Do you have a useful hint?
Quote from: Peter_T1958 on February 26, 2023, 10:14:33 AMMay be a stupid question: As I can make out, there are printing layers visible on the inside of the hopper. To my knowledge, they are difficult to remove, the more there are also a lot of rivets all around...
Do you have a useful hint?
That is a very good question.
With the angles involved with the hopper, printing in 1 piece inevitably means that you cannot optimise print angles. In this case what you see is not so much layers per se as a pattern. Because of the angle you get 1 pixel shift every 2 layers. There's supposed to be some anti-aliasing going on that should smooth that. It probably does a bit, but this is an area of 3D printing that needs work (For the technically inclined the anti-aliasing is typically done on a 256 step grey scale but resin response does NOT map well to that).
I typically print these on my 50u printer at 50u layers, but the one illustrated in the top post is done at 35u/35u layers. Higher res printers will improve matters, the cost is longer print times and typically fewer units per print due to a smaller bed.
The conventional wisdom is to angle prints to minimise such artifacts, but with regular objects you a) tend to optimise one face at the cost of making others worse and b) any artifacts tend to be in odd directions. In this case horizontal lines do not attract the eye as they might be expected in a wooden surface.
For what it's worth, it would be possible to avoid (or disguise) these artifacts by adding some noise to the layer images, or perhaps texture (woodgrain) to the planks. Not straightforward though. You can also manipulate through resin choice and exposure. Some resins are inherently softer and some over-exposure will tend to soften layer lines without impacting detail much. I've chosen a pretty hard detail resin here, optimally exposed.
With all that said, the key question is whether the layers of the raw print unacceptably affect the finished model. Firstly, the images presented originally are a large magnification. The wagons will be presented on a 5m layout viewed from a distance of maybe 600mm. Secondly, I've taken the images under lighting and post-processing to highlight details - which tends to make any print artifacts look worse than they are in practice. Thirdly, to some extent these patterns are due to subtle differences in reflectivity of the unpainted surfaces, rather than large steps.
The attached pic shows a hopper part way through painting. I'm fairly sure this is a 50u resolution hopper so lower res than the unpainted one in the initial post. At this magnification you can see layering, but you don't see that more broadly spaced pattern that you do on the unpainted print. You probably can see it if you look for it in the right light, but that's not really the issue.
In terms of your question on how to fix the problem, a surfacing primer or product like Mr surfacer will probably take care of most of it if you need things smooth. I've used pretty thin coats on the illustrated example without surfacer.
At the end of the day, 3D prints have their upsides and downsides as do all methods. In this case I ask myself 'could I make a better model that meets my needs some other way?' and the answer is no. I also feel that, once painted, the print artifacts are minor enough that they don't justify the effort to try to reduce them further. Remember that my model is a layout, so the wagon is one small part of a whole. It's not viable to do a Chuck Doan on every one.
I think the layering looks like distressed wood so it has no negative effect at all. If it bothered you, though, couldn't you print and built up the car from five individual flat pieces? -- Russ
Quote from: finescalerr on February 26, 2023, 01:07:35 PMI think the layering looks like distressed wood so it has no negative effect at all. If it bothered you, though, couldn't you print and built up the car from five individual flat pieces? -- Russ
Yes Russ, you could. The distressed wood bit is fine - it's the fact that the 'distressing' carries on through the ironwork that's perhaps less acceptable. But, truly, at normal viewing the wood looks smooth not distressed. In fact I've been playing with adding further distressing as it looks too smooth really. Remember that those planks are a scale 9" or 3.6mm wide. Even the thumbnail image is a significant magnification on my desktop.
As this thread progresses I plan on showing the parts and something of the how and why I chose to do them as I have. Every choice has benefits and downsides. I generally give the easiest way a go first. If it can be made to work I stick with it, otherwise I try something else. Obviously experience helps, but this hopper kit has gone through a few iterations.
The 5 'flat' pieces would be anything but as they need to be detailed on both sides. It's still possible, but now you have narrow unbraced parts that need printing vertically - you need to stiffen them somehow for them to print well. Possible - but not without disadvantages. And the doors are a nightmare to do this way - I've tried.
The parts need to fit and you need to assemble them. Possible, but in my experience poor corner joints are a lot more obtrusive than print artifacts. So even if my skills would be up to it, many of my customers would struggle.
It Takes your Breath away - Beautiful stuff
Barney
Quote from: Lawrence@NZFinescale on February 26, 2023, 11:58:18 AMIn terms of your question on how to fix the problem, a surfacing primer or product like Mr surfacer will probably take care of most of it if you need things smooth. I've used pretty thin coats on the illustrated example without surfacer.
Further to Peter's point, I grabbed a dud print and quickly brush applied some light grey Vallejo surface primer.
The print shows two patterns. The most obvious at this magnification are the print layers at 35 micron. There is a more subtle pattern at around 0.6mm pitch. Without magnification (and as shown in the pics of the wagon in the initial post) this subtle pattern is by far the most obvious. As can be seen, prime and paint largely take care of the larger pitch pattern. The layering remains, but this is invisible until you start blowing up images to high magnification. Don't forget those planks are 3.6mm wide, the square nuts are 0.3mm.
My point is that while these print artifacts can be a bit distracting in unpainted and high magnification images, they really are not evident in the finished model.
Lawrence, This question no doubt will further reveal my lack of knowledge of 3D resin. But besides pretty well filling in the shallow print lines with a primer, I read somewhere that, after curing, some resins can be lightly brushed with a solvent which will tend to blend/smooth the layer lines together.
That doesn't seem necessary with your N scale wagons, but if they were a larger scale and the layer lines through the ironwork were still noticeable, would it be possible to brush whatever solvent onto the ironwork alone and eliminate those layer lines but not dissolve the rivet heads? Is the process that controlable?
Quote from: Bill Gill on February 26, 2023, 06:37:16 PMLawrence, This question no doubt will further reveal my lack of knowledge of 3D resin. But besides pretty well filling in the shallow print lines with a primer, I read somewhere that, after curing, some resins can be lightly brushed with a solvent which will tend to blend/smooth the layer lines together.
That doesn't seem necessary with your N scale wagons, but if they were a larger scale and the layer lines through the ironwork were still noticeable, would it be possible to brush whatever solvent onto the ironwork alone and eliminate those layer lines but not dissolve the rivet heads? Is the process that controlable?
I've not heard that. It's possible you are thinking about acetone vapour polishing, which as far as I know is only applicable to filament prints.
1:64 is a lot larger than N scale too :-).
In any case the lines are small no matter the scale and paint/primer is readily controllable. If it goes wrong you can strip it off. Chemically smoothing things sounds a bit tricky and irreversible to me.
To put my question into perspective, it was thought to be a general one, but sometimes the language barrier doesn't allowe me to be precise enough ! Now the focus has shifted a bit too much to the planks of your beautiful wagons, although they came out very good indeed. I am sure, the fine lines won't be visible after some primär and color and I would be very pleased I could archieve your results !!!
I was interested in that matter as I am confronted with that problem to a much greater extent (Hmmm, it's only logical as I am working in 1/25 scale ::) ). I do not want to hijack your thread when I am posting here a picture just to point out what I am dealing with.
(https://images12.fotki.com/v1686/photos/4/3824994/16426959/Ladeflche-vi.jpg)
In the broadest sense, some sort of hopper too, also with rivets, bolts and washers. And at the first sight, the print didn't look all that bad, so I gave it some primer. Only now I detected all those lines. Of course I tried to smoothen them by sanding. The result: All flat surfaces could be solved; alas a some rivets lost their lives too, so : unusable!
So I secretly hoped, that a solution could be found as Bill Gill mentioned ... :'(
Thanks; Peter
Quote from: Peter_T1958 on February 27, 2023, 12:22:47 AMTo put my question into perspective, it was thought to be a general one, but sometimes the language barrier doesn't allowe me to be precise enough ! Now the focus has shifted a bit too much to the planks of your beautiful wagons, although they came out very good indeed. I am sure, the fine lines won't be visible after some primär and color and I would be very pleased I could achieve your results !!!
Hi Peter. I understand. I would like to move on from the discussion of print resolution as it applies to my wagon as it really is not problematic.
Quote from: Peter_T1958 on February 27, 2023, 12:22:47 AMI was interested in that matter as I am confronted with that problem to a much greater extent (Hmmm, it's only logical as I am working in 1/25 scale ::) ). I do not want to hijack your thread when I am posting here a picture just to point out what I am dealing with.
Don't worry about a hijack. I started this thread to give and receive ideas. Print artifacts should be independent of scale (unless you are obliged to use lower resolution), and generally I find large scale much easier.
Quote from: Peter_T1958 on February 27, 2023, 12:22:47 AMIn the broadest sense, some sort of hopper too, also with rivets, bolts and washers. And at the first sight, the print didn't look all that bad, so I gave it some primer. Only now I detected all those lines. Of course I tried to smoothen them by sanding. The result: All flat surfaces could be solved; alas a some rivets lost their lives too, so : unusable!
It is hard to give specific advice on one photo, but the general problem is well known. The print looks pretty good - What resolution is it printed at? How large is it? I think it is worth quickly painting the piece. In my experience such lines are far less obvious in a finished model, and it's the final appearance that is important. At least you'll get a feel for the magnitude of the problem, which always looks worse at this stage. The following are things you might try - not all of them may be appropriate to your situation.
- As I noted earlier, printing on an angle has advantages but, on objects such as yours, diagonal lines are a lot more intrusive than vertical/horizontal lines would be. Consider printing flat (or vertical), though this will cause another set of problems with support.
- Printer resolution. If I recall, this is your own model rather than the 1/35 kit run? If so better resolution and/or thinner layers may help.
- Resin and exposure. If you are able, you might try other resins and over exposure. On a large object over exposure won't impact detail much but should smooth your lines a bit. Some resins seem to produce a smoother result than others. Likewise print settings (such as equilibration pauses/temperature) can impact layering. A caution though, you could spend a lot of time on this approach for not much benefit, so only worth doing if it's easy.
- Leave out bolts and rivets (but print holes or location marks) and add them as separate parts later. That way you can sand the flat areas. In 1:25 it should be possible to print bolts and rivets on a stem and fit them to holes. This may be a benefit if your print resolution is on the low side as you can fit hi res bolts to low res sheet metal.
- Support. If the model is able to move/flex during printing then layer lines will be more problematic than if it is rigid. This can be quite resin dependent too. This does not look like your issue though. Make thin sections thicker if you can do so unobtrusively. My hopper has scale planks at the top edge, but these taper to thicker section towards the bottom.
- Split the model into subcomponents to make print set up easier to optimise and simplify part clean up.
- I'm no expert, but I would look at the Mr Surfacer products if you have not done so. These look to be the closest thing to the Bill Gill solution that I'm aware of but I have not tried them.
Outstanding information. -- Russ
Moving on to the how and why...
This model was always going to have etched parts, as details like the handbrake lever cannot really be done any other way.
So once the decision is made to have an etch, it makes sense to make the most of it as shown in the pic of the fret. There's cosmetic parts that are usefully in metal (shunter's step, handbrake lever), parts that are opportunistically in metal as I'm doing it anyway (inside faces of the hopper doors) and functional parts (chassis subframe and suspension sliders).
The chassis subframe is a fold up structure which keeps everything square. The angled spreaders are soldered in and the folded spreaders then removed (as they get in the way of the deck castings). The pivot points for the suspension are folded in. There's a location point for the handbrake pivot to reinforce the 3D printed underframe and the axleguards are etched to replicate the iron bar of the prototype with good strength. The rubbing points for the bearings are also etched.
With the exception of the axleguards virtually all of this will be hidden in the finished model, but these etched parts are vital to functionality and a fine appearance that could not be robustly achieved with 3d prints alone.
Making up the subframe is a pretty simple task taking just minutes.
Very interesting, Lawrence. I will follow closely the construction and explanations.
Bernhard
(https://images14.fotki.com/v1683/photos/4/3824994/14463210/Encyclopaedia-vi.jpg) ;)
Thanks again for your thoughts! And yes, as I have to send my own model to a printing service I have no influence with regard to resin, printer resolution and so on. But I already had to split a lot of parts into subcomponents (as you proposed), this also might be the solution here, I will try to omit the finer details to add them later.
Cheers,Peter
In today's installment - Suspension.
Why bother? Rigid 4 wheeled wagons can be a trick to get perfectly flat and even when that is achieved the track on which they run may not be. Suspension greatly improves track holding, particularly when running scale flanges. More than this though, the moving wagon looks better as it appears to have mass and flows around the track rather than bouncing over any imperfections. Suspended wagons also run quite a bit quieter, so it's all benefit for little cost.
The original prototype for the suspension was the first etch I drew myself back in the late '90s. It wasn't an original idea then, although this implementation is all mine and has evolved a bit over the years.
In principle, the wheel bearing is mounted in a hidden slider that is free to move behind the axleguard. The slider has an attached leaf spring made from music wire (guitar string). The spring rate can be adjusted by altering the gauge of the string. I typically use 0.008" string in 1:64.
The first pic shows the etched slider. On the left is the etching cut from the fret. The long tag is looped around and fed through the slots to form the spring retainer (centre). This looks tricky, but is really easy and very reliable. I feed the tag through with tweezers and then pull tight with flat jaw pliers. On the right the wire, and pinpoint bearing have been soldered in and the excess tag trimmed back. I deliberately make the spacing a little loose so that the fit of the axles can be adjusted using etched spacer washers behind the bearing.
The second image shows the slider assembly in place in the chassis subframe. The turned bearing sits between the horns which are a relaxed sliding fit.
This is possibly the simplest implementation of the suspension method that I've done. The bearing sits in a void within the cosmetic printed axlebox which is rigid (more on that later). Different brake/axlebox/spring arrangements have demanded slightly different solutions. Typically I use a cast axlebox that is soldered to the outside of the slider and moves with the bearing. I also attach brakes to the sliders so that the brakes move with the wheel and can be close fitted without the need for a large gap. This wagon has a spring arrangement that doesn't suit my usual approach and isn't fitted with air brakes, hence the simple solution.
Nice, a simple yet clever solution!
You have too much spare time on your hands.
I came across a smilar system in an English forum, when some fear came up, my Narrow gauge tramway loco (somewhere pinned down in this forum ::) ) could derail at every slight bump due to its scale flanges. Naturally, I do not have complete peace of mind now, when I read your lines... :D
Seriously: A very clever way of sprung mounted axles indeed, but I should have included that before... Now it's too late to improve and all what remains, is a prayer !!!
Anyway, I will safe this article on my harddisk under «modelling hints»!
Clever design and masterful execution.
Maybe I should limit myself to simple static models ....
Russ
Thanks Russ. But if it is clever and masterful it's by iteration over time. MkI was not as elegant of as effective.
Nobody cares how long Michelangelo spent creating the ceiling of the Sistine Chapel; they only care about result. -- Russ
Quote from: Peter_T1958 on March 01, 2023, 11:29:45 AMAnyway, I will safe this article on my harddisk under «modelling hints»!
I hope it helps!
The underframe of this wagon was of timber and a 3D print was a good way to represent this. As the pic shows, the functional etched parts are largely hidden away within the 1 piece printed part except those areas where the etch conveniently represents the metal of the prototype.
The brass and printed parts can be glued together with CA glue to give a structure with no visible joins to be tidied up.
The handbrake pivot would have been quite vulnerable if only 3D printed, but it is backed up by an etched pivot point. This will allow the operating shaft to be securely soldered in later on and provide a robust assembly.
Very clever!
Moving on...
I chose to cast the decks primarily to add weight, but it also allowed the brake hanger and shoe to be included as well as providing a robust mount for the couplings. I did toy with the idea of pewter, but at the time I didn't have a viable supplier. The brass is more expensive for production but is a little more convenient to assemble. Later in life these wagons had plate decks so I could easily do that version if the inclination took me.
I carefully drill and tap M1 to mount the couplers - more on that in a future installment.
Handgrab tails are part of the casting, which is relieved from behind so drilling out is painless. The handgrabs can then be soldered into place from below using a little spacer jig to ensure consistent height.
Once cleaned up the casings just clip into the etch/print assembly (as shown). Once I'm happy with the fit, I attach with CA glue and then reinforce with some epoxy.