Engineering Answers / Design & DFM
A slide is the standard answer to an undercut and often the wrong one in silicone. The material will stretch over a shallow undercut and recover — if the depth, the hardness and the lead-in are all inside the window.
Silicone can strip over an undercut up to roughly 10–15% of the diameter, at hardness below about 60 Shore A, provided the lead-in is chamfered 15–30° and the undercut has a rounded edge rather than a sharp lip. Beyond that depth, above 70 Shore A, on a filled compound, or where the undercut has to hold a seal or a tight dimension, a slide costs less than the scrap.
An undercut is any feature that blocks the part from being pulled straight off the core: an internal groove, a snap bead, a side hole, a lip that overhangs.
There are only two ways to mold it. Either the tool has a slide that withdraws before ejection, or the part strips over the feature because the material is flexible enough to stretch and recover.
Plastic usually needs the slide. Silicone often does not, and the difference in tool cost between the two answers is large enough that it is worth understanding exactly where the window closes.
| Undercut | Strips? | Condition | Main risk |
|---|---|---|---|
| Internal groove, 5–10% of bore | Yes | Under 60 Shore A, chamfered lead-in | Slight dimensional loss at the groove |
| Internal groove, 10–15% of bore | Usually | Rounded edges, slow ejection | Lip distortion on a thin wall |
| Internal groove over 20% of bore | No | — | Tear at the lip |
| External snap bead, shallow | Yes | Bead radius equal to bead height | Bead flattening on ejection |
| Side hole through a wall | No | — | Needs a slide or a pin |
| Undercut above 70 Shore A | No | — | Insufficient elongation |
| Undercut that must hold pressure | No | — | Recovery is not complete enough to seal |
Two of those rows are the ones that get argued about at the quotation stage, so it is worth being blunt about them.
Side holes never strip. A hole through a wall has material on both sides of it. There is no direction in which the part can be pulled that does not tear the hole. If the design has a cross-hole, it needs a slide, a pin, or a redesign that puts the hole in a boss that can be drilled afterwards.
An undercut that seals is not a stripping undercut. This is the mistake that reaches us most often from medical and consumer designs: an internal lip whose whole purpose is to press against a mating part. Stripping stretches that lip past its elastic range at the very surface that has to seal, and it does not come back to nominal. The part passes a dimensional check straight off the tool and leaks a week later.
Four details separate a design that strips cleanly for a million cycles from one that tears on the first shift.
On a part where the undercut is close to the limit, we cut the stripping insert as a separate piece so it can be replaced without remaking the cavity. That turns a yield problem into a maintenance item: if the undercut tears, change the insert and soften the lead-in rather than scrap a tool. It costs a little at build and saves a rebuild later.
A slide is not just machining time. On a silicone tool it adds:
Against that, a slide gives you a repeatable undercut dimension, no stretch marks, and no dependence on ejection speed. On a part where the undercut has a tolerance or a sealing function, that is usually the right trade regardless of tool cost.
Before committing to either answer, mold a small trial insert with the undercut at the depth you want and strip it twenty times. Measure the undercut after the first strip and after the twentieth. If the dimension holds and there is no visible whitening, strip demolding is safe. If it drifts, no amount of lead-in chamfer will fix it — the material is being asked to do more than it can recover from.
That test costs a fraction of a slide and answers the question with data instead of opinion.
A working rule is 10–15% of the diameter the undercut sits on, at hardness below 60 Shore A. A 20 mm bore with a 2 mm groove is 10% and strips reliably; a 4 mm groove on the same bore is 20% and will either tear the lip or leave the part deformed. The percentage matters more than the absolute depth, because what the material has to do is stretch radially by that fraction.
It leaves evidence. A stripped undercut usually shows a faint stress mark or a slight whitening at the lip, and the undercut dimension runs a little under nominal because the material has been stretched and does not fully recover instantly. On a cosmetic part that is acceptable; on a sealing lip it is not, because the same stretch that lets the part off the core is the stretch that has to be recovered before the lip can seal.
When the scrap rate from tearing exceeds the cost of the slide, which happens faster than most buyers expect. A slide adds machining, a wear surface, a maintenance item and a parting line that can flash — but it makes the undercut dimension repeatable. On a part running hundreds of thousands of units where the undercut has a tolerance on it, the slide usually wins on yield alone.