Expected price increase (Shapeways)

[RAIL N SCALE]

Expected price increase
Recently, Shapeways, the company that prints our 3D products, announced a price change. This means that in the coming weeks our products will become (considerably) more expensive. Unfortunately, at the moment it is not clear to us when the price change will take effect. We advise you not to wait too long with your order!

RAILNSCALE-Team
https://railnscale.com/

[nkalanaga]

I always thought that, like most technology, 3D printing would become cheaper as it caught on...

[peteski]

I always thought that, like most technology, 3D printing would become cheaper as it caught on...

Maybe they have a high number of rejects/reprints and the cost is passed onto customers?

[C855B]

3D output is at that awful stage of the tech at the moment. I've mentioned before that I ran 2D high-res service bureaus in the '70s and '80s. The life cycle with 3D is deja vu. We had to deal with silver-bearing film and photographic paper*, the associated chemistries and processing equipment, and ever higher demands put on us by customers with more sophisticated work that frequently required re-running. Multiple times.

* - Have you ever heard of the Hunt brothers? Film/paper prices went up nearly 4X, and... funny thing... never went down.

Sound familiar? The pioneering 3D printers have the same expensive problem in supplies and processing, and increasingly complex files requiring more equipment time, post-processing, and higher reject rates.

What hurt the 2D service bureaus was a different technology - affordable laser printers hitting the 600x600dpi threshold. We stopped getting the easy jobs that paid the light bill. Once "office grade" printers could do 1200x1200, service bureaus were done. I think that's the direction 3D is headed. Advances in surprisingly cheap filament-based "makers" are already eroding most of the casual business where absolute resolution isn't needed. So the simple jobs are gone or going away while folks like Shapeways get the hard stuff that takes increasingly more equipment time and has higher re-do rates. The increasing costs per job have to be recouped.

Filament-based machines have a way to go, but that tech or something like it will get to acceptable resolutions and replace the expensive noxious chemistries of the vat printers for most hobby work.

[Lemosteam]

Maybe they have a high number of rejects/reprints and the cost is passed onto customers?

This.  And they refuse to address quality issues that would prevent this.

@C855B Technology will NEVER overcome the surface quality unless someone develops materials that can morph and blend into curvatures on their own, like fluid colors on a 2D print.

Remember the old adage, no matter how many half-steps you take, you will NEVER arrive at your destination.

From a 3D printing perspective the image below is what I mean.  The design intent looks nice and smooth, but under 10x size you can see the layer thickness of the black print (this is what creates the undesired striations on a contoured surface.  The 50x image shows the green represents a 100% decrease in layer thickness, and will still result in striations.  The orange represents another 100% decrease, still resulting in striations. 

Unfortunately, layer thickness is not the only culprit.  I have shown only a cylindrical design.  Now imagine curvature coming out of the image toward you.

Each cross section now must also interpret the shape of each 3D model cross section for EVERY layer.  Laser/jet/print table movement (SLA) and flash accuracy (DLP) error states (offset, shape) now come into play with the curvature of the surface as designed.

The question is, how small does the layer thickness and cross section accuracy have to get before an N Scale surface is completely acceptable out of the printer?

Remember as well that the machine cannot double the speed at which it prints so the thinner the layer, the longer it takes to print and time=money.  Time may already be at its threshold, and prices will only rise, unless the printing, cleaning and curing is perfect the first time round.

[Philip H]

would those stepping/curvature issues be solved by laser scintering inside gelatinous resin?  I can see the ever finer striations remaining an issue in the lay down to print machines, but I'd think the ones that harden resin with lasers would not suffer the same fate.

But I'm an oceanographer, so what do I know?   

[Lemosteam]

I hope I was not sounding condescending, I was not trying to be.

All printers that use a layering technology suffer the issue (SLA, DLP and Filament) just in slightly differing ways.  With SLA and DLP the fluid hardens instantly in a cylinder or cross section, respectively, the diameter of the laser (it does not expand to form shape) that is from the surface of the fluid to partially into the previous layer so they fuse.

If you did layers of multiple single laser beam (SLA) location it would probably look like this (exaggerated of course), laser diameter reducing slightly through the material:
___
\_/
\_/
\_/

Layers of a photo hardening (DLP) would be similar to the laser only across an entire cross section.
____________
\__________/
\__________/
\__________/

Similar for a filament as the diameter of the filament basically represents the laser diameter, except i assume there might be some material oozing out at the base so likely upside down from the laser example above.
___
/__\
/__\
/__\

Now the sections can be relatively smooth around their edges, but the laser (SLA) still moves in stepped increments, I assume the photo hardening is based on the section resolution (still digital like my image, maybe smoother than stepped movements , and the filament printer moves in stepped increments too.

One day I will create an exaggerated 3D model of these scenarios.

Also recall that an STL file is a reduced resolution file of the CAD model as well using triangles to simulate compound curvature.

As an oceanographer, you know that analog (water motion, etc.) is the only way to produce smooth curvature at least until you get down to the size of the sand grain and even then the grains are not stacked in layers to form shape, they fall into each other and compact to the smallest gaps between grains.

I don't mean to be doom and gloom but digital printing has it's limitations.  It's not like God forming clay into natural organic shapes, it is after all a digital reproduction.

I liken it to digital versus analog sound, one is obviously better than the other, which is why records and phonographs are booming (pun intended).

[Maletrain]

Since the 3D printing striations are not going away any time soon, I would think that it would be best to design complicated things (like say steam engine shells) in multiple parts that are designed to simplify the smoothing process before assembly. 

I have been thinking about trying to get somebody to make parts for  a steam engine, and it seems to me that a boiler shell without domes or stacks or rivets would be the best start.  Domes and stacks could be printed separately.  Pilot decks and running boards could be fabricated from brass.  Rivets and stay bolts could be added like decals.  Cabs could be printed or fashioned from styrene sheet. Piping and grabs could  be fashioned from wires and tanks from tubes or bar stock.

The problem with current 3D printing really seems to be that there is a desire to do everything with that one manufacturing process, rather than to choose the best manufacturing process for each part and then do assembly.  The best RTR manufacturers are tending towards my suggested approach.  Some kit manufacturers seem to be going that way, too.

[C855B]

Aliasing will become moot at X resolution. That's what happened to 2D, and is what I expect to happen with 3D. IOW, get the voxel size small enough, you won't notice the artifacts.

[Lemosteam]

Since the 3D printing striations are not going away any time soon, I would think that it would be best to design complicated things (like say steam engine shells) in multiple parts that are designed to simplify the smoothing process before assembly. 

I have been thinking about trying to get somebody to make parts for  a steam engine, and it seems to me that a boiler shell without domes or stacks or rivets would be the best start.  Domes and stacks could be printed separately.  Pilot decks and running boards could be fabricated from brass.  Rivets and stay bolts could be added like decals.  Cabs could be printed or fashioned from styrene sheet. Piping and grabs could  be fashioned from wires and tanks from tubes or bar stock.

The problem with current 3D printing really seems to be that there is a desire to do everything with that one manufacturing process, rather than to choose the best manufacturing process for each part and then do assembly.  The best RTR manufacturers are tending towards my suggested approach.  Some kit manufacturers seem to be going that way, too.

^^^^ This is what I try to do with all my designs to make it easier for the kit builder to sand the parts with no details in the way.  I draw the limit at surfaces like that, but not for underbody stuff that will be black and hardly visible.  its a reasonable tradeoff IMHO.  I would even do this on a diesel, if I were so inclined to design one.  All of the doors would either be etched or separately installed parts and the main shell would have exacting holes to fit the pieces snugly.

[Lemosteam]

Aliasing will become moot at X resolution. That's what happened to 2D, and is what I expect to happen with 3D. IOW, get the voxel size small enough, you won't notice the artifacts.

This may be true, but at X resolution the manufacturing time and calibration maintenance may drive the cost beyond the affordable. I suppose the first copy can be used to make molds from if one is inclined to start casting to defer the printed cost, but then the designer must also be a manufacturer.

At the current resolution it appears that SW is already raising prices to make up for it, and I fear that this new cost structure may actually put the cost out of range for most.  Shop owners are leaving SW over this as i write.

Regardless, the 3D print will still have to deal with accuracy issues and machine movement, calibration, etc.  Easier to do on a 2D format because ink flow and only two dimensions to move in.    This is why most Auto manufacturers still use clay to develop analog surfaces (using digital as an assist only).

I am also not sure if 2D printing is dimensionally accurate either where 3D must be accurate to fit parts together.  Drawings still carry a note warning not to measure printed drawings- measure the CAD.

[Chris333]

Form what I've heard from others. Shapeways overseas is a much better service. Their machine must be calibrated better or more often so the prints come out better.



And their shipping is better 

[Lemosteam]

Since the 3D printing striations are not going away any time soon, I would think that it would be best to design complicated things (like say steam engine shells) in multiple parts that are designed to simplify the smoothing process before assembly. 

I have been thinking about trying to get somebody to make parts for  a steam engine, and it seems to me that a boiler shell without domes or stacks or rivets would be the best start.  Domes and stacks could be printed separately.  Pilot decks and running boards could be fabricated from brass.  Rivets and stay bolts could be added like decals.  Cabs could be printed or fashioned from styrene sheet. Piping and grabs could  be fashioned from wires and tanks from tubes or bar stock.

The problem with current 3D printing really seems to be that there is a desire to do everything with that one manufacturing process, rather than to choose the best manufacturing process for each part and then do assembly.  The best RTR manufacturers are tending towards my suggested approach.  Some kit manufacturers seem to be going that way, too.

@Maletrain , see this- it took me over 3 years and 1000 hours to design and release: 

https://goo.gl/photos/TU4dQ9gCizR6i7Za8

Model, venue and image by Chris Broughton

[Lemosteam]

Form what I've heard from others. Shapeways overseas is a much better service. Their machine must be calibrated better or more often so the prints come out better.



And their shipping is better 

@Chris333 Yes they were the ones who finally printed my Stillwells correctly, which I have since had to discontinue because of SW inaccuracies and they have not promised me that I can use that facility form them.

[peteski]

John,  the workaround for the striations is to coat the printout with a liquid primer/surfacer which will fill those steps int he printout, resulting in an acceptably smooth (analog  ) surface.  We already successfully use that technique (along with some sanding) on 3D printed parts.  If the printing steps get fine enough, will likely not even have to sand the part - just spray it with a primer/surfacer.  Mayny modelers prime their scratch built models before painting, so the number of steps needed to paint 3D printed parts will be no different than any scratch built parts.