Fitting one part inside another (tolerances)?

First post to Shapeways forums! Merry Christmas all.

In my design I've got a tube (grey) with 27mm ID and I want to slide the red part into the tube. I don't want to print as one as there are some other things i want to do with the red part before assembly. I want the red part to slide easily into the tube with an easy friction fit (removable) and no play as concentric alignment is important for my design to work properly. Should I under size the red discs so they'll slide into the grey tube or could i print them with 27mm OD? I plan on printing both with the strong flexible plastic.

I'm also concerned the grey tube may not be round after cooling from the printing process. The walls are 1.5mm thick. Should it likely print ok?

thanks much,
Fred

I definitely wouldn't go with zero clearance. I'd go with something like 0.08 to 0.1.

Another way would be to put a number of radial spring tabs that would be set to something like 27.5 OD so as to provide backlash compensation. The cool thing about 3D printing is there is no manufacturing steps so you can make things very complex without having to worry about what resources would be consumed to manufacture it.

Orientation affects roundness.

I always recommend oversizing things and then sanding or otherwise machining things to get a "good" fit. Having no play is going to be difficult with straight geometries like that. And there's no guarantee the two parts would be oriented the same way in the printer either, assuming they are the same material. Having them be different materials further complicates the problems.

I just got an ornament back with a hole and a 7mm rod that is supposed to fit into it. Despite having a slightly oversized hole it was not round enough to let the rod enter but everything is fine after using a small round file on the hole. Always assume there will be problems with tight tolerances even if you use the Shapeways spec numbers. The best designs for parts that are mated together are to be tolerant of dimensional warping or size variations.

Ignoring the inner plates you'll have enough issues with the end unless you provide some sacrificial internal rim you can sand or file down to get an acceptable friction fit. I'm not sure if your use requires enough of a seal to minimize air or liquid leakage but that would make things even harder. You might consider using a screw thread to hold the pieces together but that creates additional uncertainty in terms of getting things right the first time. Or you could use an external mechanism like a hose clamp to hold the pieces together. It' difficult to make many recommendations without knowing how this will be used.

Could you incorporate an 'O ring' into your design. It may take up any irregularities in the curve of the tube.

Thanks for the o-ring idea. I tried printing grooves for o-rings in another cap/tube assembly but either undersized the groove width & depth in my drawing or it printed narrower than the design. In either case I had a hard time getting the o-ring to stay in the groove in the cap so it would fit into the tube properly.

In some earlier designs similar to what I'm doing here I've included features to use grub screws to fasten the core into the tube. For this design I'm going to try a threaded cap. This pdf has some design guidelines and at the bottom are suggestions for threads in SLS using hemispheres as the positive thread to reduce friction.
http://www.3dsystems.com/company/datafiles/SLS_Guide.pdf?use rActiveBreakpoint=1

At the base of the threads I added a beveled face with a corresponding camfer in the tube so when the cap threads into the tube the ramped faces should help concentrically align the assembly. There is 0.2mm of clearance everwhere. When tightened up the bevel and camfer will come into contact and the clearance between the positive hemispheres and the negative threads will be taken up... that's the theory at least. Hope I drew it all correctly. Might eliminate the flange on the cap so the cap can tighten into the tube without bottoming out on the flange, just the camfer.

Hey thanks for the link on SLS design. Some great tips in there! -S

Keep in mind that the SLS printed part you end up with will not have a high degree of fidelity on those positive hemispheres. They will be there, but they will look nothing like your drawing. They will be muted. That is a good idea regarding how to reduce friction, I just hope there is enough there to make it function properly.

You never said what the application was, but, another thing you can do is use 3D printing to make your own custom taps and dies for making perfect threads. To do the cutting think machine tool inserts like these. You can 3D print the threads and then chase the threads to perfection using a custom tap and die.

The printed result is what I'm anxious to see. The hemisphere positive threads are 2mm dia. So I'm having just those two little parts printed to see how they turn out.

I've been wondering about printing threads, have watched some youtube tutorials and subsequently drawn some parts. The question i have is if i want to print a part that will thread onto a threaded steel rod where I know the thread specs how do I properly draw the thread so the part will spin on easily and not be a tight friction fit? Its a larger part the threads are I believe 1/2x20.

And I have already tried printing threads for the grub screws then chased them with standard taps and it worked very nicely.

At two mm they should be fine. I was thinking they were smaller.

To make a 1/2 20 nut is doable I feel, but I've never tried it. There is the out of round issue that was mentioned and the threads will not have very good fidelity at that size, but I still think it would work to a degree.

What I would do is make a batch of variations and find the answer via trial and error. Actually, Shapeways should have already determined this for people, but they never have.

Another way would be to use the example of your reduced friction thread, but provide a means so that the positive hemispheres can move in and out while under spring tension. As an example of what I'm saying look at this 1984 Popular Mechanics "Squirm Drive" article. Click on page 76.

Here are a couple pics of the threaded parts. They printed nicely. They thread together with little resistance. But when tightened down they misalign a little bit and also loosen quite easily. It looks like I could've gone with 0.1mm clearance and been fine.

So for use in the design I originally posted (grey tube with red insert) I'll likely go with continuous positive threads instead of the hemispheres since there won't be as much depth available for threading. Likely make a more fine pitch as well. I'd like the red core to thread in snugly and not easily loosen.

Another experiment with a different fastener design and tighter clearances. The threads didn't stay snugly together very well probably in part because the 0.2mm clearance everywhere was to loose.

Instead of threads this assembly uses a twist lock action. There are three 3mm dia cylindrical tabs that rotate into slots over 0.1mm bumps at the entrance. Going over the bumps is -0.1mm clearance. Then at rest, locked into the slot there is 0mm clearance. It snapped into place but if i pushed the two parts together i could rotate the tabs without feeling any bump.

So to lock the red core part into the grey tube in my OP there will be 4 larger 3x6mm tabs that will similarly twist lock into the grey tube. Kind of like what is drawn onto the parts in the 2nd pic.

Here are pics of the drawn assembly and the printed result. Can't use 0 thickness edges like were printed on the end of the outer cylinder ;-)
The twist lock action at the other end of the outer cylinder are drawn with a 0.2mm space between the tab and the cylinder and i can flex them with a pencil point, but they're likely to fragile.