Snap, glue, weld or friction?

Hi Guys

I have a question for you. I am trying to join two polycarbonate components together without having to glue or ultrasonically weld. Can any of you guys maybe suggest an alternative such as Friction or press fit? Would it be at all possible to get some sort of snap feature in there? All ideas are very welcome. The component has a diameter of approx 16mm and a combined height of 30mm (just to give you and idea of scale) There is about 2mm of the male collar going into the female.

Thanks in advance

It’s hard to suggest without understanding what this will do and what kind of load it needs to withstand.

If it’s just a cosmetic piece and it popping out isn’t a big deal,glue would probably be fine, though 2mm isn’t really a lot of material to engage with.

How are these parts made? Are they molded? CNCed?

The parts are to be injection molded and we are are trying to avoid glue. We have been through gluing and while it is strong enough it is time consuming and therefore expensive. The cap must not pop off…and although there will not be too much force acting on it is possible that, as times it could get a knock. It is not cosmetic, the parts house another component.

Thanks for your response!

Well snaps are an option, but you need to take into account your tooling design since it will most likely involve adding at least some side action in the tools to create the necessary undercuts.

Is there a reason you can’t increase the engagement between the two parts? If you did that you could probably add some small crush ribs on the outside of the cap part to get a better friction fit - but 2mm of engagement isn’t going to be enough to be very secure.

We would like to keep the cost and complexity down as much as possible. I can maybe increase the distance by 2/3 mm. have you any info on designing a friction fit? How much bigger should the male part be?

I can’t say I do, I’m not an ME - but just looking at that distance it doesn’t really look like enough to give you a good tight fit with that limited engagement. Obviously the more engagement you can get the better you would be (you also need to factor in the draft on the surfaces of your part…you also seem to have a very very thin wall section, not sure if thats just an example and not the final part.

Can you do a screw in part? Put threads on each component?

A press fit might work for that. You’ll really want to work with your injection moulding provider to figure out the best dimensional specs for that. The mating surfaces on each part are going to vary with whatever diameter and roundness tolerances their process generates. PC is not a brittle material, though, so there’s room for error.

I’m a little concerned by the walls on the female part in your drawing. It looks like there’s a spot where a sharp corner is combined with a really thin wall, and that might rupture.

As for assembly, there’s two ways to do it. Usually you have a press of some sort and you just cram one part into another. But you can also cool down the male parts (or heat up the female parts), assemble by hand, and then the fit will tighten as they converge to the same temperature. This is usually done with metal parts, but PC expands and contracts about 5 times more per unit temperature change than mild steel, so it seems like something that you could try.

This looks like a great candidate for a press fit.

One of the pieces should have standard bullet-nose pins molded in. The mating piece would have a hex shaped hole.

Usually I’ve seen the pins have a height:diameter ratio of about 1.5-2:1.

the tough part about this approach is dialing in the right interference. The pin diameter should be a few thousandths larger than the flat to flat distance on the hex (both of which you will probably want to tolerance to plus/minus .001).

Dialing it in can be expensive because it’s tough to prototype. You can’t machine a hex hole (if you can, I’ll pay for the name of the machine shop), and while you can urethane cast the parts, you won’t be able to use the production material. You might want to look at prototype (aluminum) tooling if your volumes are high enough.

As far as specific numbers go, it depends on the application. Less interference means you can snap them by hand, but less pull out resistance. More means stronger, but will likely need a press and a fixture.

Also consider if the pins will be in shear. It might be nice to have some other features to take up the shear load.

that really should be an ultrasonic weld. It’s surprising you’d consider inferior joining methods but not a u/s weld which is exactly what is called for here.

other than that… solvent bond and you’re done.

I agree: ultrasonically weld that sucker… You’ll pay for the tooling (the horn), but you’ll be glad you did.

Here’s a good design guide from Branson:


I’ll agree, ultrasonic weld would be the traditional and perhaps most cost effective way to do this. It’s also easy to design, there are dozens of resources out there that will give you exact geometry for the energy director.

That being said, look into press fits. There aren’t any hard guidelines and tolerances are more critical, but assembly is a snap (couldn’t help it), and you can get away with no tooling. If your load requirements aren’t extreme, it might be the most effective way to go.

Trying to think of something good besides press fits or USW… I’ve got nothing

My only concern with pres fits may be stress cracking… Because it’s cylindrical, I’m thinking of hoop stress here… What’s the application? Service environment? Service temperature? Chemicals? What’s the reason for going with polycarbonate?

I would go with annular snap fit. Think aspirin bottle cap with a sharp inner edge and nothing to grab outside to open it. I recommend Jordan Rotheiser’s book “Joining Plastics” page 398 and material calculations on page 401.