Re: 3D printing (Was: New early P&R steel.....)


Dennis Storzek
 

On Mon, Dec 23, 2019 at 01:37 PM, <dalemuir2@...> wrote:

That means accuracy should be within approximately 0.1% to 0.2%

For reference, Shapeways specification for Fine Detail Plastic is way looser: ±0.3- 0.7 mm for every 100 mm

I don't do 3D printer work, but I do know something about designing plastic parts. I'm also familiar with "jobbing" work out to other shops.

It doesn't make any difference what tolerances the Polyjet SHOULD be able to hold... Shapeways has told you what THEY guarantee to hold, and that's all you can expect to get. They publish that their maximum shrinkage is .7%. If your parts are shorter than this, they owe you replacement parts. If not, you got what you paid for.

It may come as a surprise to some people, but injection molded parts also have problems with differential shrinkage. The resin manufacturers publish guidelines for shrinkage, always as a range, typically .004 to .006 inch per inch for high impact styrene, almost four times that for acetal. The part designer has to take this into account for any parts that will be used in assembly. Slotted holes are a must for long parts. These should be easy to do in the solid model, and the printer will faithfully replicate them. Likewise, unlike the old saw about not forcing a square peg into a round hole, that's exactly what you want to do (more typically round pegs into square or hex shaped holes). This allows some space for an overly tight peg to displace material into. There is no model building material that can guarantee "line-on-line" fits, not even photoetchings. The part doesn't shrink, but there is always the possibility of over or under etch.

If the ends of long parts need to visually align and you don't have control of the process that makes them, you have to design in someplace where you can take a little tuck if needed. If the end plates of your girders have exposed rivet detail, maybe they should be separate parts so the main girders can all be sanded to the same length, then the ends applied. I ran into this when I was designing brass replacement sides to fit under Rivarossi passenger car roofs. I measured three random roofs, and came up with an .018" difference, so, on the theory that I hadn't found the longest possible, I made the letterboards .006" longer than the longest, with the instruction that the modeler must file them to fit the roof HE has, then center the doors in the resulting opening. .012" over the width of a .280" wide door isn't going to be noticeable, but a .012" gap at the corner of the letterboard certainly would be.

The good designer anticipates the problems and chooses where he wants to deal with them.

Dennis Storzek

Join main@RealSTMFC.groups.io to automatically receive all group messages.