Hi Matthew, Thanks for your analysis. With your feedback we can whittle away the differences between the ideal and the feasible. The dimension in my drawing are measured from one of the 3 cross arms I have above my garage. For reference I took some measurements from an Atlas pole, that represents the acme of poles. The cross arm is 0.038 by 0.0315 by 0.778 versus 0.025 by 0.01875 by 0.75. Their oversizeness is not offensive to my eyes even though that comes out to 6"x5"x10'4.5". The sway brace is 0.02" wide, which corresponds to 3.2"; quite oversize but not too shabby looking either. Their main disadvantage is having only 4 insulators. They are also quite fragile. The Atlas insulator is essentially right on top of the cross arm and there is almost no discernible pin. I imagine eliminating the pin would strengthen the insulator attachment quite a bit. The insulator has minimal shape; a taller than wide flat top cone. The insulator is 0.052 tall by .0310 diameter at the bottom and tapering to 0.024 at the top. That scales out to an 8" insulator, which is too high. I can see the insulator being a small amount wider in diameter than the cross arm and just a bit taller. How do those dimensions look to you?
Brian, Thanks for your update. See my response to Matthew regarding supersizing the parts to make them producible. Do those dimension look more workable? Nancy
On the topic of weight. Perhaps an easy approach to increasing the weight of cars would be to blend tungsten powder into the polymer feedstock before selective last sintering. Tungsten powder is readily available through golf shops since it is added to golf clubs in localized areas of the club to increase mass to achieve different performance. I t does not take very much. For instance to double the weight of a component would only require 5 volume percent and to triple 3 volume percent. Basically the polymer has a density of 1gram/cc (essentially the same as water). Tungsten has a density of 19.6grams/cc. Thus to increase the polymer weight from 1 gram to 2 grams one would add 1/19.6 of tungsten to double the weight. I would not expect this to have much effect on the Selective Laser Sintering process. One would need to be careful handling the mixed powders since powders of this extreme density variation can segregate or "unmix". Of course, convincing some one to do this might not be so easy. This could probably be used in casting processes as well. Possibly another approach.
Assuming that 3-D printing is still not up to the task of producing cross arms, does anyone have any pointers for how to get a traditional, injection molded part produced? I have many examples of incredibly detailed parts that show me that the technology exists to produce at least semi-detailed cross arms. I imagine the process would involve creating a master part, probably out of metal, that would be used to create the mold. I would like to understand the process steps and the costs.
That's the plan! I haven't set up a store yet though - I'm hoping to do so this weekend. I probably won't put the boxcar up for sale for a few weeks though (I would like to double check my final modifications before I start selling them). Sounds like a cool idea to me! Shapeways offers metal parts which are made by (what I understand to be) a process similar to what you are describing. I have shied away from these metals as a model railroading material however because they are rather expensive and not anywhere near as exact. That said, I've read or heard about people in the "real world" applying 3D printing technology to make everything from aluminum airplane components to artificial bone implants. As such, I'd guess that denser plastics or metal model railroad components might only be a few more years away! That said, for the time being at least, I at least am stuck using whatever Shapeways offers
When you do, send me a PM and I will be your first customer!! I have a gift certificate that I have been saving for one of my designs, but those gondolas are awfully tempting. :tb-wink:
Bump! Seeing all the amazing work going on in 3D, I started messing around in Sketch-up 8, and yes, there is a learning curve, particularly where curving surfaces are involved... I'm slowly coming up to speed, and the tutorials are helpful though obviously not designed with model railroaders in mind. So a question for you more advanced 3D modelers, Matt or others: what's the best way to do appropriately sized rows of rivets and how does one make them go around round corners? I'm running into a block wall here and hope I don't need to revert to using rivet decals... that would kind of defeat the purpose. Any thoughts would be appreciated. Kind regards, Otto K.
Otto, Sketch complete spheres the diameter of the rivets desired. Scale rivets you won't see in N, so you have to make them slightly oversized. A dimension of .008" is a good diameter to start experimenting with. Center the sphere on whatever surface it is intersecting. This will give the appearance of a hemisphere rivet head no matter what type of surface it is embedded into.
I've found this process does a decent job at allowing semi-mass production of otherwise unavailable rolling stock. Though the quality still seems dependent on how the contractor for Shapeways decides to orientate the model. Here is one I did that turned out very well. It's a four-pack of CNW ore cars. You just cut them off the sprue like you would an injection molded car. The warped car on the upper right was easy to fix. I made some bracing out of styrene to spread it back, placed the car in a cup of water and put it in the microwave for one minute. Once it cooled, the warp was gone.
Nice! Looking good! Otto - to add to what's already been said - when working with rivets (or any other repeating element for that matter), you'll want to use either a linear or radial array (depending on the shape that you are working with). A quick google search should get you up to speed if you aren't sure what I am talking about. As for corners, if they are tight (the corners of a tender for instance), I generally don't bother intersecting rivets. It's certainly not impossible to do so, but if you space the rivets out right it doesn't look like anything is missing. Steam boilers are clearly a different animal, but luckily those are easily done using a radial array
It was taught to me by Mark Gasson of Mark4Design to draw rivets as a cylinder, not a sphere. The poor resolution will take care of the rest. This was done on a Perfactory system which is much higher resolution than the Shapeways process. Jason
I second that on drawing rivets as cylinders, not spheres. Not only do they come out better, you really save on the poly count. If the rivets are small enough, I just use an extruded octagon.
Another trick I learned. If you have delicate details like stirrup steps, after you carefully clean the print take super thin CA and apply a drop or two to the fragile detail. Capillary action will wick it right in between the layers, greatly increasing strength AND rigidity.
Thanks guys for all the suggestions, I'll just have to keep experimenting.it just saves so much time if you have someone show you... Regards, Otto
Jason, Hmmmm...Verrrrrry interesting. Is it just me or is that door translucent? Will the Perfactory printer allow another translucent color such as green, blue or turquoise...or clear???? Obviously the material isn't FUD and will the RP company you went through give you a discount if what you want isn't needed "rapidly"? I've got several RP projects already drawn and three of them were going to Shapeways the first of next week for a trial run. I'll still probably do that, but if I could get my parts in a translucent green, blue, turquoise or clear, and with an increased resolution (!)....that would be PERFECT! Cheers! Bob Gilmore...another confirmed rivet counter
Bob, AFAIK, the material is what it is. There's no choosing a color. And no it isn't FUD, it's a completely different process than what Shapeways does. As for pricing and such, you really just need to contact Mark. It's been a while since I've had anything done by him, but I can give you an example of what you might run into. I had this caboose quoted for me by Mark. He told me I would need to buy four (to fill the build area) and they would be $80 each. I would have also needed to build the cupola as a separate piece. The Shapeways model cost me $15 but the quality is very poor. If you are sure about your drawing and want the best quality, the Perfactory is worth it. Jason
The EnvisionTEC (Perfactory) machines are capable of beautiful results no doubt. That said, if you decided to go that route, make sure that you get into contact with someone like Mark who has experience with one. By my understanding, in addition to being more expensive, EnvisionTEC machines are both more size limited and require more careful/deliberate part design and arrangement (such as requiring a separately rendered caboose cupola).
Jason, Thanks for the info. However, I decided to do my own research and I've been in contact with a rep at EnvisionTec, who makes the Perfactory system. You are correct in that the Perfactory machines use a somewhat different process than Shapeways and their subcontractors. I was quoted a price of between $4.00 and $6.00 per cubic inch of the materials used in the EnvisionTec machines. So, unlike Shapeways, you are being charged a whole lot more for your caboose bodies (at $80 EACH!!!) than just for materials and a small set-up fee. Also, there are no design limitations on the Perfactory system other than the "build envelope". I am assuming that including the cupola as an integral feature compromised Mark's minimum charge for printing, meaning he could not stuff four bodies into his "build envelope", but if the cupola was separate, he could fit it all in for the basic $320. This means he is using the Perfactory 4 Mini XL with ERM, which has a "build envelope" of 4.5" X 2.8" X 6.3". This particular machine has the best "resolution" which is a remarkable .0012" or .192 of an N-scale inch!! That's resolution on the X and Y axes. Resolution (depending on material used) on the Z axis can be as low as .0006" (.096 of an N-scale inch!!...LESS THAN A SCALE 1/10th OF AN INCH!) The materials available include an actual ABS material, a Polypropylene equivalent, a flexible material, medical plastic, a metalic "silver", wax for direct investment casting and a hard, high temperature material for making masters for vulcanized rubber molds. A couple of transparent materials are also available and are available in colors. The Perfactory process does not produce visible "layers" like the Shapeways processes. Perfactory machines interpolate a gray scale and deposit their pixels of material according to the gray scale density, so there is no discernible layering...just smooth surfaces and curves. The actual cost of materials for your caboose project could not have been more than for three cubic inches of material, which would have put it at between $12 and $18 (for all four caboose bodies), so Mark is making a good profit and paying for his machine, which cost him at least $102,000.00 delivered to his New Zealand door. You also have to have a "light polymerization chamber" to cure your prints, and which costs an additional $4,000.00. As we all know, ultimate quality is expensive, but boy, would I ever like to have one of these in my workshop!! I just have to win a lottery somewhere! Cheers! Bob Gilmore
Bob, there's no doubt you boned up on the machine and the process, but you're missing some of the practical parts of the equation. When Mark is running my four cabooses at $320, he's taking time away from $5000 (or whatever, but much more profitable) projects. Also, some companies put a premium on the 'Rapid' part of Rapid Prototyping. I can show you a quote for the same caboose that was $800 for a single caboose from an outfit in California and that machine was only slightly better than what Shapeways uses. Jason
