Small self-tapping screws driven into plastic bosses are the joint that punishes manual assembly hardest. This is the case where automated fastening changes the most — and the case our machines were originally designed around.
At a glanceTypical fastening: M1.7 – M2.5 · single- and dual-platform screw driving and dispensing cells configured per part · built in Dongguan, China since 2011
Most engineers arriving here go straight to one of these:
Consumer electronics enclosures are assembled with M1.7 to M2.5 screws, almost always self-tapping into a moulded plastic boss. The boss is the weakest part of the joint: a screw driven a few degrees off-axis cuts a new thread instead of following the existing one, and the boss cracks rather than the screw stripping. The part is scrap, and because it happens at final assembly the whole sub-assembly is lost, not just the housing.
The second pressure is volume. A single unit can carry eight to twelve screws, and a line running several thousand units a day is asking an operator to repeat the same precise motion tens of thousands of times per shift. Torque drift over a shift is not an operator failure — it is what happens to any human hand after four hours, and it is why end-of-line inspection keeps finding the same defect with different operators on different days.
The third is contractual rather than technical. Brand owners increasingly require that every screw on a unit is recorded — which screw, what torque, pass or fail — so a field failure can be traced back to a specific station. On a manual bench there is no such record, and retrofitting one costs more than automating the station in the first place.
These are the parameters we need from you to configure a station rather than to guess at one.
Confirm the exact thread (M1.7, M2.0, M2.3, M2.6) and head form. Vacuum pick-up handles these sizes; a screw feeder is only worth it once the same screw runs in real volume.
Plastic bosses usually want low torque with tight tolerance. The smart driver logs torque and angle per screw so the window can be proved rather than assumed.
The fixture must support the boss from underneath, not just locate the housing — an unsupported boss will flex no matter how accurate the screwdriver is.
Count the screws per unit and multiply by the required output. Our arm platform runs 0.37–0.41 s per screw, plus travel between screws — travel usually dominates, not driving.
If the customer wants a tightening record per unit, settle the data format and the MES interface before the machine is configured. It is a software setting, not a rebuild.
Send the part, the screw size and the torque window and we will come back with a configuration and a cycle-time estimate.