The hard programmes live in one head: HR for aerospace CNC parts makers

Walk into an aerospace machine shop and ask which jobs make the owner nervous. The answer is rarely a machine. It is a part number: a thin-walled titanium housing on the five-axis trunnion, an Inconel seal ring on the mill-turn, a deep-pocket aluminium rib that chatters unless the toolpath is just right. Behind each one stands a single CAM programmer who proved it.

Proving such a programme takes weeks. There is the CAM model, the post-processor output, simulation and collision checks, a dry run in single-block, cautious first cuts, probing, and a first-article report that must agree with the drawing. When that programmer resigns or falls ill, the shop discovers the programme alone is not enough to repeat the job with confidence.

• Fixture tricks. Which clamp sequence stops the thin wall springing, and where a soft jaw needs a shim.
• Control-side edits. Feeds overridden on the second roughing pass, a dwell added before a finishing cut, a tool length nudged after a crash months ago.
• Post-processor quirks. A manual edit to the posted code because one machine's rotary axis unwinds differently.
• Tool choices. The insert grade and stick-out that survive Inconel, and the ones that look identical but burn.
• Inspection judgement. Which CMM feature tends to drift and which hand check catches it first.

None of this is hidden on purpose. It accumulates because the programmer is always proving the next job, and writing the last one down never reaches the top of the list.

When the person is away, repeat orders stall or are re-proved from scratch, with fresh scrap risk on costly forgings. Lead times quoted to the customer stretch. Setters and operators cannot grow, because every hard set-up waits for one man. And the programmer himself carries a weight that slowly pushes him toward a shop where he is not on call every evening. A larger plant nearby only needs to phone at the wrong moment.

• Nobody owns the job library; programmes live on one workstation or a pen drive.
• Proving is rewarded and documenting is not, in pay or in praise.
• There is no grade between senior setter and programmer, so no natural understudy.
• Repeats arrive months apart, long enough for everyone else to forget.
• Revisions are handled by the same person who proved the original, so knowledge concentrates further.

1. 1

   Check first: the single-person list

   List every part number that only one programmer can prove or re-prove after a revision. Note the machine, the material and how often it repeats.

2. 2

   Build a proving pack per part family

   Set-up sheet, fixture photos, clamp order, tool list with stick-out, control-side edits, post-processor notes, CMM watch points. Stored on the shared server beside the programme.

3. 3

   Name an understudy for each family

   Usually a strong five-axis setter. He sits in on the next proving and runs the next repeat with the programmer watching.

4. 4

   Create the missing grade

   A proving setter level between senior setter and programmer, with its own band. That gives the understudy a reason and a title.

5. 5

   Reward the teaching

   Count each completed proving pack and each understudy sign-off in the programmer's review and grade. Teaching becomes part of his job, not a favour.

The aerospace CNC parts makers page sets out the wider pressures in this kind of shop. Its learning and development page covers the operator-to-setter-to-programmer route, and the compensation benchmarking page covers banding the new grade against nearby shops.

Two moments expose the single head most sharply. The first is a drawing revision. A moved bore or a tightened profile tolerance means re-posting, re-simulating and often a fresh first article. If only the author can judge which toolpaths the change touches, every revision queues behind him. Let the understudy handle minor revisions with the programmer countersigning. The second moment is the night shift. Roughing Inconel overnight with nobody able to read spindle load or hear a chipped insert is how a forging becomes swarf. A proving pack with written abort signs, such as load spikes, squeal or blueing on chips, lets a night setter stop the cycle with confidence.

A programmer asked to document everything may hear: we are preparing to replace you. Say the opposite and show it. Put his name on every proving pack. Give him first say on new machines and CAM upgrades. Make the understudy part of his team. Owners who frame it as building his bench, not draining his head, usually find he cooperates.

This week, write the single-person list with your programmer. Next month, finish one proving pack for the family that repeats most often and name its understudy. The hub page for aerospace and defence components shows how this pattern recurs across the group, and the requirements form is where to describe your own shop if you want help.