Staples doing 3D Printing In Store

[Zox]

Just trying to explain how shapes are represented and why the striations may take some time to disappear.

I suspect that, eventually, there will be a combination of additive (3-D printing) and subtractive (CNC machining) technologies into a single unit, that will be able to create just about any shape with great precision.

CNC machining compensates for 3-D printing's inability to do slopes; being able to work on one printed layer at a time compensates for CNC machining's inability to get "inside" objects.

[rswinnerton]

Yeah really. 20 years ago, wireless phones were the play toys of the rich. 10 years ago, they could be had for a penny but could only make calls. Texting was expensive and usually didn't work across different carriers. Today, look what you can do with a free iPhone.  The same will come with 3D printing at home. 10 years from now, they'll be almost as common as inkjet printers are today. Instead of going to Chulvis's shop, we'll go to his website, pick out a car and paint scheme, download a file and 30 minutes later, voila!

[Scottl]

And following the iTunes or similar pay model, people who invest in the 3D modelling exercise will be able to sell their designs.  It's going to be great! 

[DKS]

I think "never" is a rather rash assumption.

I agree. While it is true that certain kinds of technology have failed to pan out as anticipated 20-odd years ago--such as free-floating full-color 3D television images, as portrayed in countless sci-fi books and films--other technologies have literally exploded and surpassed assumptions (many film photographers were once claiming that digital imagery would never catch up to film). My gut says this is one form of tech that will enjoy continued progress until a point is reached that would satisfy even pickier modelers. That's not to say it will necessarily be affordable any time soon, but that's a different kettle of fish.

I was a newbie at any 3D modeling software, but well versed with a couple of 2D CAD programs.  I drew up plan drawings in Cadrail for my little 3D project in N-scale, then...with measurements and 2D drawings in hand, drew it in Sketchup in about  7 hours on a lazy Saturday.  That included not knowing anything about, or ever having used Sketchup before, so I had to learn how to use the program along with my first drawing.  The online tutorials really made a big difference in my learning curve.

I did pretty much the same thing, in about the same amount of time. But you and I had the advantage of a strong background in computer image generation (I've been using vector drawing apps since they first appeared on the pre-Windows PC market). This makes a huge difference in how quickly one can adapt to a new app. And I agree that SketchUp's online tutorials are very well done; without them, my first drawing--the factory superstructure below--would have taken ten times as long.

[Lemosteam]

Not to take this down a path of being too argumentative, but I can't think of anything 20 years old in technological terms that compares to what is possible now.  I totally accept that the current printers have limitations, but this is a rapidly transforming industry that is on the verge of becoming mainstream.  Once this hapens, it will bring in money, R+D, and new ideas that will improve it and we will indirectly benefit from it.

All I have to think of is that $3000 colour laser printer we bought a decade ago:  it is now superceded for a tenth of the cost with better tech specs.  Or thin screen TVs, cell phones, satellite phones, etc....

I understand it appears that way in many other arenas.   Costs have come down, no question, but an I phone isn't any clearer and doesn't do anything more than phones of old (referring to voice communication only).  I'm just telling folks what I know and have experienced.  ALL of the tech I have seen recently from all of the "affordable printing" category isn't any better than what I experienced then.  Take it or leave it.

Mark4 and others have machines that have much more accurate capability than the shapeways of the world and that's why product manufacturers use them for prototyping their designs- to see and feel them before cutting half-million dollar injection tooling.

Since I was trained, I have been following this tech to see how it could serve me as Robert has done.  I'm simply saying the tech today cannot meet my personal standards for N Scale, I'd rather cut & shut a bash then try to fill or hide the marks the machines leave behind.

Robert I agree 3D cad really is very intuitive, once one learns the commands and proper steps.

Zox, even CNC cutters have radii that will leave behind high points or edges.  Where I work automotive clay is CNC machined to the general design, but the clay is not perfect afterwards and CANNOT be scanned into CAD until PEOPLE smooth and scrape the clay to it's desired form, pleasing to a chief engineer's eye. Then it can be scanned into CAD.

I've said enough and I am not a basher- I'll be a user when that time comes.  As I said I was just trying to explain the process and until the process improves to a much finer level, accuracy in N Scale will have to be sacrificed.  Like I said, maybe someone will invent a machine or process that fills or removes the missing surface...  Just telling you that in my line of work people have been complaining that STL cannot produce a class A surface, just like those who complain about the striations on shapeways parts.

[robert3985]

Chris333, I'm afraid that will NEVER happen.  Until someone invents a 3D printer that can FLOW material to fill in the surface striations, they will always be there.

Striations will always be a large as the smallest pixellation that the printer can create- in this case it is the thickness of the paper and glue, and the accuracy of the cutter.  In other cases it is the diameter of the material coming out of the head and the thickenss it can be appled, and yet in others it is the diameter of the laser and the thickness it hardens the fluid to.

Sorry for the dissertation and the quandry...

3D printers are already available that produce parts that look like injection molded parts.  Actually, they look better than injection molded parts because they have no relief to them, which makes it possible for injection molded parts to release from the molds.

The printers are made by EnvisionTEC and the process is called "Perfactory".  These machines do not "layer", but cure pixel by pixel, the smallest voxel (volumetric pixel) dimension being .0006", or .096 of a 1/160th inch.

I am sure you agree that your eye is unable to perceive an indentation or protrusion that is an N-scale 1/10 of an inch.

As to your illustration and contention that a smooth surface is something that only polishing can render, please be aware that there is no surface that is perfectly smooth.  The infinite value of pi demands that any arc be made up of increasingly tiny straight lines, which means that every curved surface in our reality is comprised of tiny flat planes...the smaller they are, the smoother the surface...but, they will always be there.

However, practically...the day is here with machines that are fully capable of producing parts that ARE BETTER than injection molding as far as detail is concerned, and very nearly as good as far as surface quality is concerned (I hedge my statement as to quality since I have not seen clear canopies created using Perfactory machines yet.  Maybe they look equally as good as injection molded ones.  I don't know yet.)

The price is around 106 grand for a Perfactory machine that would sit in your home workshop, with its accompanying de-waxer, but EnvisionTEC offers a desktop model (less resolution, less capability) for about 1/10th that price.  Won't be long until resolution, versatility and price are all where some of us want them to be, and we'll have one or more in our workshops.

PS.  I thought this was interesting.  Ford, when mocking up their vehicle interiors, goes to great trouble to make them look as mass-produced as possible, since interiors are not just function, but also design and very close to the end-users eye.  Perfactory 3D machines were the only ones that could render several objects to Ford's designers' specifications...in particular including the knobs and buttons of interior switches and instrumentation, many of which have glossy surfaces.  Maybe the surface quality IS there...

[SkipGear]

I've had perfactory parts in my hand, they still look pixelated. I have a customer at our shop that is working on something unrelated to the hobby industry and was having parts for his invention done on a perfactory machine.  I was supposed to sand and paint the parts to get an injection molded look so his parts looked like a finished product. As is, they still looked like a 3d printed model, better than most anything I have seen, but still were grainy and pixeleated. This part had a very complex and organic 3d curve in it so it was a perfect piece to see what a perfactory machine could do.

The thing that has improved over 20 years is the price, not the quality. I believe that 2 million dollar machine that John is talking about used intersecting lasers in a resin bath to build the parts. How much finer resolution can you get than a beam of light?

[Chris333]

Whatever the lime green stuff is from Finelines. Give me that at an affordable price.

[GaryHinshaw]

The infinite value of pi demands that any arc be made up of increasingly tiny straight lines, which means that every curved surface in our reality is comprised of tiny flat planes...the smaller they are, the smoother the surface...but, they will always be there.

I'm sorry, Robert, but the above comment triggers my bullsh^t detector.  "The infinite vale of pi demands that any arc be made up of increasingly tiny straight lines"?  Umm, first, pi is not infinite, it is 3.14159....  It is irrational, which may be what you meant, but it's definitely not infinite.  Second, in differential geometry, the concept of an infinitely differentiable manifold exactly describes what you would think of as a perfectly smooth surface.  There is no serious mathematician in the world that would claim that "any arc must be made up of increasingly tiny straight lines".  Certainly in the physical world, things get tricky at the atomic scale, but that has little or nothing to do with the value of pi.

The rest of your post is fine though.    Sorry to go OT, carry on.

[DKS]

^^^ This is why I love Railwire...

[Mark W]

I'll get excited when someone post a photo of something in N scale scale that looks like it was injection molded...

[daniel_leavitt2000]

I tried some 3D software, but it was damn near impossible to make anything with a laptop and touchpad. I think I need some serious new hardware to get the hang of it. I just don't want to drop the money with the new basement and the need for a mill for another project.

Seeing all the work Bryan has done with 3D, I really feel left out. I would do anything to have the skills nessisary to make buses and other vehicles.

[conrail98]

Guess, I shoulda come back here earlier, but I was too busy learning about Adobe's advances in web development. My point in posting the link was that 3D priting has kind-of been looked at as a niche or coming technology. Once brand names like Staples jump into it, you'll see it gaining greater usage across the board, not just in modeling. Also, printing presses and big printers used to be only found in industries or commercial settings, now we have small printers in every home. Let the big boys work out the kinks and in 5-10 years, we may seek Inkjet3Ds being sold at Best Buy and the like. Oh, and to see how far technology has come, all you have to do is look at this:

[Lemosteam]

The thing that has improved over 20 years is the price, not the quality. I believe that 2 million dollar machine that John is talking about used intersecting lasers in a resin bath to build the parts. How much finer resolution can you get than a beam of light?

SLA 500 by 3D Systems:  Still have the training manual!  LOL!  Check out that computing hardware!  No intersecting lazers- only one reflected off mirrors to perpendicular down to the fluid surface hardening a round dot that was continually moving.  Then the vat would lower one thickness and a wiper bar would make the fluid consistent thickness across the top of the last pas and repeat the process. The software made cross sections of the CAD model for every layer the machine had to "draw" on the fluid. The machine would spend most of it's time on the outline of the cross section to trap the fluid in the outline then would quickly go over those pools to semi harden them to save time.  The Time consuming bit was cleaning off the waste of the supports and risers that held the part up in the fluid and the UV hardening time in the "oven" to finish hardeing the pols that were trapped in the cross sections.  Old stuff there, but mot much different today, save for the perfactory style.



Someone please show me a macro image of a diesel radiator fan on an N Scale shell produced by any 3D printer that is as clean as an injection molded part and I'll agree.

Robert, Ford does that because one-off rapid proto is CHEAP for vehicle prototypes.  Try and find one surface in your car's interior that is smooth that is not glass or clear plastic for vision.  In production those RP parts will be replaced b y injection molded parts with surface stippling becuase it is the only way to hide surface imperfection from the consumer.  For high volume mass production, injection molding is the cheapest process.  For low volume product like you are working on, it may be good enough and cost effective.  I am aware of no one in a high volume industry that uses un-processed RP (straight out of the machine) parts for production- it just takes too long to make a part.

Also, in the PI terms, RP parts are manufactured on an X Y and Z axes so those tiny planes you mention are not manufactured.  Those tiny planes are cross-sectioned for use by the RP machine to reproduce and as far as I know there is no muli axis RP machine that can print in a non layered fashion.  Although I did look at Envision's video and it does appear to be a different process than slicing...

Mark W, I really want to see a macro image in the nooks and cranies of that model and an estimate of how much tim you spent cleaning it up before paint.

"Voxel" limitation is not something that software or advances in technology can eliminiate.  There are size and time limitations that prevent this from being as perfect and cost effective as it could be.

Gary, PI is infinite in terms of the number of decimal places that never end.  I think that is what Robert may have been referring to.

Another thing no one has mentioned is that the smaller the resolution gets, the longer it will take to print the parts because it is printing less of the part at a single instant.  Time=money=more cost to the consumer.  Shapeways has found a level of both that is "good enough" for the average paying consumer.

Here is an interesting article I came across:  http://www.explainingthefuture.com/3dprinting.html

[up1950s]

Just because a flat made arc can look like a curve it will never be a true curve . It will always be a faceted curve like shape which will be proven when thin enough lines are used and viewed close enough under magnification . It is a way to mathematically create parabolic curves to fit tangent lines .

If something is turned on a lathe , or sanded smooth , or bent by physical or atmospheric force , there is no flat .

The world is round , last I read , though in London there is loads of flats .