Delrin to Aluminum Tolerance

Is the Delrin glass filled? In any case, the Delrin has at least an order of magnited greater expansion ratio.

Since it will be restrained (in compression) when heated, tight is better.

Look up the CTE forboth and be sure the joints are tight at the lowest temperature and youl be fine.

From my phone.

Zytel is another option to Delrin for the corners, going to finalize that tomorrow glass reinforced for sure.

So cte should be relatively mild in the room temperatures production facilities this will operate in.

So would you estimate “tight” would be for example;

[1.500”+/- 0.003” rail inside]

to

[1.495” +0.002/- 0.0” plastic insert]?

Going to confirm machined and extruded tolerances tomorrow, but still appreciate your input.

Probably doesn't matter much if it's acetal (Delrin^®) or Nylon^®.

You want an interference fit. [1.500”+/- 0.003” rail inside] to [1.495” +0.002/- 0.0”] is too loose in my opinion. 1.497-1.503 is OK. 1,495-1.497 may be a problem. If you have a 1.503 and even a 1.497 there is no interference.

The plastic will deform when pressed in to the extrusion so you want it tight when coldest.

Problems I see are, if the extrusions are saw cut and not chamfered inside and the machined plug is not tapered on the end, insertion may be a problem.

Both of those operations will increase the cost.

An adhesive may be a good back-up.

Thanks lyn! The plastic insert will have slightly (.032”) rounded edges. Adhesive is out cause it needs to be disassembled, but once assembled there will be a binding part added exhibiting strong tension from outside the rails inward (toward frame center). The main concern is a firm uniform fit so that the plastic inserts don’t have much play or shift (we’ll as little as possible to maintain precise geometry/ not warp the frame)..

So if [1.497 +.003/-0.0 plastic to 1.5 +/-.003 rail] is OK... what would your good fit (tolerance) be? (Given imperfect machining and extruding exists lol!) thanks!

OK, adhesive is out.

Now, I'm assuming that you really mean "exhibiting strong tension compression from outside the rails inward."

The fit becomes lees of an issue and the contact length of the insert becomes more important. Longer is better, but uses more material.

Can you tell us what the panels are? Is this a window? Wall? Pass through?

It will help if you can provide a narrative of the purpose and use.

Yes I meant inward compression.

The length of inserts are about 1.25” insertion. Can’t increase the dimensions so the firmness of the (up down side to side) fit needs to be relatively unyielding and nothing a few hammer taps can’t accomplish. I don’t expect perfect, just hashing out a good plastic to aluminum fit and tolerance to incorporate and send out for prototype.

Thanks again hoping to hear your numbers to try for tolerance!

Maybe even just a straight 1.500” aluminum and 1.500” plastic..?

Assuming you will be buying an off the shelf Al extrusion the variation will be whatever the supplier says it is. I doubt that you can specify this. Depending on what you specify ±.003 is probably typical.

The tighter the tolerance, the more expensive the machined plastic part will be also.

Linear Temperature ExpansionCoefficient
- α -
10^-6
(m/(m K))

A line to line fit will probably be OK. If it's a tight fit, a larger hammer will be needed.

Maybe +.0 -.003 for the plastic.

It will come down to trial and error, depending if you are really hand fitting and hammering them in place, or if a press jig is used.

Still too much unknown for more detailed response.

One other thing to consider is warming the Al and cooling the plastic insert to allow the plastic to "grow" after insertion. This is typically done when inserting a metal shaft into a hole in a metal part to achieve an interference fit.

All this adds cost, I know.

...any thought to bonding?

If by bonding you mean glass reinforced, yes that’s being heavily considered!

Bonding to Delrin needs very good surface prep. as the material is naturally quite slippery.

This will be a machined, glass filled part which will help adhesion a lot.

Did you miss the OP's statement that it is a machined part? $$$$$$$

I have no idea as to how many parts he is talking about.

With just a few dozen parts the machine setup will dominate. With the correct tool selection you may not even have to change tools.

With an intermediate number and CNC control it will probably not matter.

With many parts machining makes no sense.

Neither do I. Prototyping usually indicates a small quantity..........................

GW is on the money with the suggestion.

Ribbed parts allow closeness and flexibility of fit to overcome tolerances while still enabling a snug outcome.

Many hardware suppliers would already have corner parts for extruded square tube to enable construction of articles like window frames or outdoor gate and fence panels.

Ribs will also enable easier dismantle when the time comes.

Thanks for all your input! Ribs are an excellent suggestion we’re I actually using square tube, it’s very precise and only square on three sides, the fourth side has intrusions inwardly and the corner matches it for a solid fit.

At at this point I think I’m looking at going to prototype with a straight 1.5 to 1.5 fit and incorporate a rubber mallet to fit..

GW has a valid point, if quantities justify automation. Machining centers can do amazing things.

I got the impression this is not a huge product run.

Good luck.

So I guess that’s true they could machine ribs into the inserts at least on three sides, but I’m not familiar with the dynamics of ribbed inserts, it seems like it would give way to much flex between the base and rib. Once the frame is under compression it [can not] lose any tension in any direction, the geometry of the frame must remain nearly perfect and if the ribs absorb any of the compression, then they may be prone to give more and a more until settled/deformed and compromise initial frame tension. (In my mind anyways)

Ribs can take various forms - from being the component that flexes, to being extremely rigid and allowing the tube wall to act as the spring.

No matter what - tapered plug entry and taper the tube end so it does not sheer the plastic as it is inserted.

Ribs can also be axial rather than circumferential as in the earlier image. This would localise your interference fit & make insertion easier.

A friend has started 3D prints parts with ABS.

He is getting impressive results. 3D printing may be an option for trial fits, especially if you end up going with glass filled parts that don't print (??).

A caution at this point: If the joint has residual compressive stress, the plastic part will eventually develop stress cracks. The higher the residual stress, the quicker the stress cracks grow.

The polymer composition of the plastic has a large effect on how quickly the cracks develop. In previous designs using acetals, I inserted a rubber grommet around fasteners and clamps to reduce point stress and distribute loading. You might want to include a flexible insert. You also should seriously consider designing for clearance and glue it together with RTV silicone or flexible 2-part adhesive to get a high strength connection with zero residual loading.

Have you done any fatigue life evaluation of the joint?

www.chemicalbook.com is an online chemical knowledge dictionary, in addition to a platform of chemical industries’ purchasers and suppliers with a directory of thousands of Chinese chemical suppliers. (Location: China)