Non-contact is one of those requirements that sounds like a small detail and reorganises the whole design. The moment parts must not touch each other, the tray becomes a positioning device rather than a container, and every tolerance in the chain starts to matter.
Start from the part, not the tray
Three questions before anything else:
- What surface is critical? A machined face, a painted surface and a glass lens have completely different tolerance for contact.
- Where is the part’s own tolerance? A casting varies more than a machined billet, and your pocket has to accept the largest part while still holding the smallest.
- How is it put in and taken out? By hand, by robot, by a vacuum gripper. A design that works for a careful operator may be unusable for a robot moving at speed.
Locate on something stable
Locators should pick up on the part’s most repeatable features – machined faces, bores, established datums. Locating on a cast surface or a trim edge introduces variation you did not ask for, and the pocket then has to be oversized to compensate, which defeats the purpose.
This is the step most often skipped in a rush, and it is why trays end up with generous clearance everywhere and parts still manage to touch.
Clearance is a range, not a number
A pocket dimension has to accept the part at maximum material condition with manufacturing variation in the tray, while still restraining the part at minimum. In practice that means the useful clearance is usually smaller than designers expect and larger than the customer wants – typically a few millimetres either side, adjusted for how many parts sit in a tray and how they move together.
What holds the part down
If the tray is stacked, the layer above is holding the parts below. That works until it does not, and the failure is the part lifting during transit and landing back on its neighbour. Where that is unacceptable, the options are:
- A lid or intermediate layer that closes the pocket
- A positive retainer – a clip, a strap, a formed lip over the part
- A captive pocket depth where the part cannot clear the surrounding wall
Check the tray, not the drawing
The only reliable validation is to build samples, load them with production parts at both ends of the tolerance range, and run them. We would rather send a sample that gets argued about than a drawing that gets approved.
Cost reality
Non-contact trays cost more. More pockets per part, tighter tooling, more engineering time. If the requirement came from a customer, that cost is usually recoverable. If it came from internal caution, it is worth checking whether contact on a non-critical face is genuinely unacceptable – sometimes it is, and sometimes the requirement was written once and never revisited.
Prototyping is not optional here
Non-contact designs do not survive contact with a drawing alone. The first sample almost always reveals something – a part that rocks in its pocket, a locator that fights the operator’s natural grip angle, a surface that the part rests on in a way nobody expected.
The cheap way through this is a soft prototype: cut the pocket geometry from a spare panel, fix it to a base, and try it by hand before any tooling is committed. It will not tell you about durability, but it will settle the geometry, which is the part that is expensive to change later.
Budget for two rounds of samples. Designs that go straight from drawing to production tooling usually need a third round anyway, by which point the tooling has been modified twice.
What we build
Our automotive dunnage work is mostly non-contact trays with welded locators and edge frames. Send us the part model and the critical surfaces and we can usually say quickly whether the geometry is workable or whether the part itself is going to make this expensive.