Port Cutouts and Fit Tolerance on PDA and Scanner Cases

Port cutout tolerance for PDA and barcode scanner protective cases is the specification line that quietly decides whether your programme ships on time. Everything else about a rugged case — durometer, colour, texture, drop rating — can be argued in a meeting. Whether a charging cable seats in the port with the case on cannot. It either does or it does not, and you find out when the samples land on your desk with two weeks left in the schedule.

The reason cutouts go wrong is not carelessness. It is that four independent sources of variation add up, and only one of them is under the moulder's control. This piece sets out the stack-up, shows the arithmetic that tells you how much clearance to design in, explains the mould-cutting practice that lets you fix an undersized opening cheaply, and gives a validation sequence to run before steel gets cut.

Why the cable doesn't seat

Picture a USB-C charging port on a handheld computer. Your drawing says the cutout is 12.0 mm wide. The moulded part measures 11.6 mm. The cable's overmoulded boot is 11.9 mm across. Nothing seats.

Trace it back and you find no single error large enough to blame:

• The mould cavity was cut 0.03 mm off nominal, which is normal machining variation.
• The silicone shrank 3.2% instead of the 2.9% the tool was compensated for. On a 12 mm feature that is only 0.04 mm, but on the 145 mm overall length it is 0.44 mm, and the port sits off-centre by a fraction of that.
• The device housing itself varies. The device maker holds its own tolerance on where that port sits relative to the outside surfaces, and nobody told you what it is.
• The cable you tested with in the office is not the cable your customers use. Boot dimensions on USB-C cables vary substantially between suppliers.

Four small numbers, one non-functional part. The fix is not to demand tighter tolerances — moulded rubber does not do tight tolerances economically — but to design the clearance so that the sum of the variation still leaves a working opening.

The five contributors, and how to add them up

Here is the stack-up for a typical moulded silicone or TPU case cutout. The bands below are representative for this class of part; get your own supplier's numbers rather than adopting these.

Two ways to combine them:

Worst case — add the bands: roughly ±0.5 to ±1.3 mm depending on feature size and how much you know about the device. This is the number to use when a failure means the part is unusable, which is exactly the case for a charging port.

Root-sum-square — take the square root of the sum of the squares: roughly ±0.3 to ±0.6 mm. This is the number to use for cosmetic gaps and for features where a small misalignment is tolerable, because the contributors are independent and rarely all go the same way at once.

The practical rule that falls out of this: design a functional opening at least 0.5 mm larger per side than the biggest object that has to pass through it, and reserve worst-case stacking for the openings where failure is binary. A 1 mm gap around a USB-C port looks sloppy in a render and saves a tooling revision in reality.

Note which row in that table is largest. It is usually the device, not the case. If you can get the device maker's mechanical drawing with tolerances — not just an STEP model of nominal geometry — you have removed most of the uncertainty from the programme. Ask for it. Many will provide it under NDA, and a nominal-only STEP file is a known trap: it tells you where the port is supposed to be, not where it might actually be on the unit in a driver's hand.

Cut the steel safe

This is the single most valuable habit in case tooling, and it costs nothing to specify.

Adding steel to a mould means welding and re-cutting: slow, and it leaves a witness line on the part. Removing steel means machining away material: fast, cheap and clean. So you cut the tool such that every dimension you are unsure about errs toward more steel in the cavity, which produces smaller openings and thicker walls on the part. Then, at first sample, you open the features that need opening. This is "steel-safe" design, and mould shops that build cases for a living do it by default — but only if you ask, because the alternative gives a better-looking first sample and a worse second one.

Applied to cutouts specifically:

• Cut port openings 0.3 to 0.5 mm undersize per side on the first tool. Expect to open them at T1.
• Cut button pads slightly proud, not slightly recessed. Material can be removed; feel can be tuned down but not up.
• Cut the internal cavity of the case slightly tight rather than loose. A case that is marginally hard to fit can be relieved; one that is loose needs the whole core reworked.
• Leave the scan window and camera apertures undersize and open them after you have seen the optical performance, not before.

Say all of this on the drawing, in a note, in plain language: "All apertures to be cut steel-safe; first article to be sampled undersize for opening at T1." Then budget for the revision. A tooling quote that carries no allowance for a first modification round is a quote that will produce a change order later — a pattern that shows up across tooling and mould costs in China generally, not just in case work.

The four openings that cause the most trouble

Charging port and I/O. The connector opening is not the constraint; the cable boot is. Buy three or four of the cables your users actually use — including the rugged ones with strain-relief boots — measure the boots, and dimension against the widest of them, plus clearance. If the device charges in a cradle through pogo contacts rather than a cable, that is a different problem entirely, described below. Also check the depth: a case wall 2 mm thick moves the connector face 2 mm deeper into the opening, and a short-nosed plug may no longer reach.

Scan window on a barcode reader. The exit aperture has a cone, and case material inside that cone causes decode failures at the angles a warehouse operator actually holds the device — which is to say, not straight on. Get the horizontal and vertical field angles from the device maker and dimension the aperture so that no case material intrudes at the extremes. A soft silicone lip is worse than a rigid one here, because it deflects into the path when the device is gripped hard. If the case has any lip at all across the window, chamfer it away from the optical path rather than leaving a square edge.

Camera and illumination. Two separate failures: vignetting at the corners of the frame, and internal reflection from the flash or aimer bouncing off the inside face of the case aperture back into the lens. The second one produces a washed-out image that looks like a lens fault and is actually a case fault. Matte, dark material on the inside face of the aperture and a generous chamfer fix it. If the device has an infrared sensor, the constraints are tighter still and overlap with the thermal and optical considerations for AI edge and thermal-camera hardware.

Cradle and dock contacts. The one that most often gets discovered late. A docking cradle is designed around the bare device, and many docks have a maximum case thickness they will accept — sometimes published, often not. Three things have to work at once: the case has to clear the dock's retention features, the charging contacts have to meet with enough pressure, and the device has to seat deeply enough for the dock to detect it. There is no way to validate this from CAD. Someone has to put the cased device in the actual cradle. Decide early whether that someone is you or the factory, and if it is the factory, plan to ship a dock.

Button feel deserves a line too: a moulded silicone button pad over a device's own dome switch changes actuation force and travel. Users notice immediately. Specify the pad thickness and the air gap behind it, and accept that this is a feature you will tune at T1 rather than get right on paper.

A validation sequence before you cut steel

Run this in order. Each step is cheap; skipping one is not.

1. Get the device drawing with tolerances, not just nominal CAD. If unavailable, measure a sample of at least five real units at the critical dimensions and record the spread you find.
2. 3D-scan a real device and compare the scan to the CAD you were given. Discrepancies here explain most of the surprises later.
3. Build the stack-up table for each functional opening using your supplier's actual shrink and process bands, and decide worst-case or RSS per feature.
4. Print a fit check. A 3D-printed shell in a rigid resin will not tell you how a silicone case stretches on, but it will tell you within a day whether the port positions are right. Do this before mould design is frozen.
5. Send a golden device to the factory, with a written record of what was sent, its serial number, its condition and how it comes back. Establish this in writing — a device that goes to a factory and does not return is a recurring and entirely avoidable annoyance.
6. Send the accessories too: the cradle, the bulkiest charging cable, the hand strap, the belt clip. The case has to live with all of them.
7. Define the first-article check. A dimensional report against your drawing, plus a functional checklist: cable seats, dock seats and charges, scan decodes at extreme angles, buttons actuate, case fits and removes without tearing.
8. Approve on a golden sample, not a photo. Keep the approved sample sealed and dated as the reference for later production inspection — the same discipline that governs pre-shipment inspection generally.

Fold steps 4 through 7 into the standard sample order and approval process so they have a slot in the schedule rather than competing with it.

Common questions

What tolerance should I put on a moulded silicone cutout?

Rather than writing a single ± figure, name a rubber dimensional tolerance class — the international standard for rubber product tolerances defines classes that scale the allowance with the nominal dimension, which is how moulded rubber actually behaves. Then add a separate note for the handful of functional openings where you need something tighter than the class, and expect the supplier to price those specifically. Asking for a blanket ±0.1 mm across a silicone part is a request that will either be quoted very high or quietly ignored.

Can a factory guarantee the fit before tooling?

Nobody can guarantee it from CAD alone, because the device's own variation is outside everyone's control. What a competent supplier can do is model the stack-up with you, cut the tool steel-safe, and build a first-article validation into the schedule. Treat any supplier who promises a perfect first sample on a new device with more caution, not less.

Is a ready-made mould safer than a new one for fit?

For a device model a factory has already tooled, yes — the fit problem has been solved once already and the sampling is fast. A factory that maintains a library of public moulds for standard terminal models can put a physical sample in your hands in days rather than weeks, which is a different risk profile from a custom programme. That trade-off is set out in more detail in the piece on what a public POS case mould library actually gets you. For any device model that has not been tooled before, you are back to the stack-up.

Who pays for a cutout revision at T1?

Settle this before the purchase order. A revision caused by an error in the drawing you supplied is yours; one caused by a mould cut off-drawing is the supplier's; one caused by device variation neither party knew about is a negotiation. Writing the steel-safe intent into the drawing helps here, because it establishes that a first modification round was expected and priced.

What to ask the supplier next

1. What dimensional tolerance do you commit to on moulded cutouts — as a millimetre band, or as a named rubber tolerance class — and does it differ between silicone and TPU?
2. What shrink rate are you compensating the tool for, on which material grade, and how much lot-to-lot variation do you see?
3. Will you cut all apertures steel-safe, and is a first modification round included in the tooling price?
4. Do you accept a physical device sample for fit validation? How is it logged, and how and when does it come back?
5. Do you check cradle and dock clearance before cutting steel, and do you need me to ship a dock?
6. Will you issue a first-article dimensional report against my drawing as standard, or is that an extra?
7. What is the sampling window for a custom tool versus an existing mould, so I can schedule the validation loop rather than discover it?
8. What quality-management certification do you hold — certificate number, issuing body and validity date — and can you send scans of the originals so I can confirm them with the issuer?

That last question applies to every supplier, and it is worth asking plainly: a logo on a slide is not a certificate, and the scope on the document has to cover the process you are buying. On the capability side, the WJM Silicone factory profile describes in-house mould design and CNC fabrication with a 12-engineer R&D team and over 10,000 completed custom mould projects, alongside a library of ready-made moulds for standard POS terminal models — a useful reference point for what to expect an answer to look like when you put the eight questions above in front of a shortlist.