UK NZ CNC Partnership: How a Split Supply Chain Actually Works

• 1
  What a UK NZ CNC partnership really splits
• 2
  Which part families suit the split
• 3
  Holding ±0.005 mm across a long supply chain
• 4
  Material choice and what it does to the plan
• 5
  Freight, customs and the lead-time math
• 6
  Where the arrangement breaks down
• 7
  When the split works and when it does not

Most UK and New Zealand manufacturing collaborations are described in marketing language. Strip that away and there are only three things that can move across the border: design authority, metal removal, and metrology. A UK NZ CNC partnership works when each of those three sits in the country that does it cheapest or fastest for that specific part family.

In practice the split is usually this. Design intent, tolerance stack-up and drawing release stay with the UK engineering team, because they own the assembly and the certification paperwork. Machining moves to a supplier with the machine hours available, which for many part families is a partner plant outside the UK. Final dimensional reports either travel with the parts or are produced at the receiving site.

The mechanism matters because it decides who owns risk. If a UK buyer keeps design authority and signs off on the inspection report, the partnership is a capacity arrangement. If the buyer also hands over tolerance definition, the partnership becomes a development arrangement and needs far tighter communication.

One more thing. A UK NZ CNC partnership is not the same as a UK supplier with a New Zealand sales office. The parts have to physically cross a border, clear customs, and survive a long air or sea leg. Everything below follows from that single fact.

Split-country machining suits parts that are hard to make but easy to ship. A 400 g aluminium bracket with tight bores and a complex profile is close to ideal. It is light, it does not corrode in transit with the right packaging, and its value per kilogram is high enough that air freight does not destroy the business case.

It also suits low-volume runs where the UK shop is fully booked. Sending a 20-piece batch of 316L stainless manifolds to a partner with 16 simultaneous 5-axis machining centers can pull weeks out of a schedule. The machine hours exist; the buyer just cannot access them locally in time.

Parts that do not suit it: large weldments, anything over roughly 4,000 mm in the longest dimension, single-source castings that need rework, and assemblies that must be fitted together on site. Freight cost and handling damage scale with size, and one bent frame wipes out the saving on the whole batch.

Surface finish also sets a boundary. As-machined finishes at Ra 1.6–3.2 μm travel and rework fine. Fine finishes at Ra 0.2–0.8 μm can be damaged by handling, so they should be packed in individual cavities or finished at the destination.

Tolerance is where most split-supply arguments happen. A machine can hold ±0.005 mm on a part it machines in one setup. The same nominal tolerance becomes a negotiation when the part is measured in the UK on a CMM that has never been correlated with the machine tool's own probe system.

The fix is straightforward. Agree on the datum scheme before the first chip is cut, and write it on the drawing. Datum A should be a surface that is machined in the first operation and never used as a clamp face afterwards. If the datum moves between operations, the tolerance stack moves with it.

Temperature is the second variable. An aluminium part measured at 30 °C in a warm shop will read differently than the same part at 20 °C. For work at ±0.005 mm, both sides should measure at 20 °C after the part has soaked for at least 4 hours. This is not bureaucracy; it is the difference between a rejected batch and a passed one.

Third, keep one inspection authority. Either the machining partner issues the dimensional report, or the UK site does. Two reports that disagree are worse than one report nobody reads. Reports on request, and the receiving site should be able to re-measure the same features without reworking the datums.

Material decides shipping weight and, sometimes, customs paperwork. Aluminium grades such as 6061-T6, 7075 and 6082 are the default for split supply: light, stable, and available in most machine shops. Titanium TC4 (Ti-6Al-4V) and Inconel are also machinable but the part cost is high enough that freight becomes a rounding error.

Stainless 303, 304, 316 and 17-4PH sit in the middle. They machine well, resist transit corrosion, and are common in medical and food-equipment work. If the part needs passivation after machining, decide early whether that happens before or after shipping, because passivation changes the surface and can move a tight bore.

Magnesium AZ31B and AZ91D need extra care. They are light, which helps freight, but they are more sensitive to moisture and to clamping marks. If a magnesium part travels, it should be packed with desiccant and inspected on arrival before it enters the assembly line.

Plastics change the calculation in the other direction. POM, PEEK and PA are light and machine quickly, but they move with humidity and temperature. A plastic part held at ±0.005 mm should be measured in the same climate as the assembly, not in the shop where it was cut.

Air freight from a partner plant to a UK site typically adds days, not weeks, but it does add paperwork. Commercial invoices, HS codes and country-of-origin statements have to match the drawing revision. A mismatch at customs holds the whole batch, and the schedule slips by more than the machining time ever did.

Sea freight is cheaper and slower, and it suits heavier parts where the schedule allows. For a 10,000-part run of a stable part, sea freight can be the right call. For a 20-piece prototype batch, it rarely is.

Lead time should be built backwards from the assembly date, not forward from the order date. Add machining days, inspection days, packing, freight, customs and a buffer for a re-measure. A quoted 3–5 days to ship from the machine shop is not a door-to-door number.

Packing is part of the engineering, not an afterthought. Machined faces should never touch each other. Individual cavities, VCI paper for steel, and desiccant for magnesium and aluminium are standard. If a part arrives with a nick on a sealing face, the saving is gone.

The first failure mode is a drawing revision that only one side has. If the UK team changes a hole position on Friday and the machine shop cuts on Monday from the old file, the parts are scrap. Version control has to be explicit, with one file name convention and one release channel.

The second is inspection disagreement. This almost always traces back to datums or temperature, not to machine capability. When two sites disagree on a feature, the fastest resolution is to ship the part back for a joint measurement rather than to argue over numbers.

The third is capacity drift. A partnership built on spare machine hours breaks when the partner's own orders pick up. Build a second approved source before you need it, or accept that the arrangement is temporary and plan around it.

None of these are reasons to avoid the model. They are reasons to write down who owns design, who owns machining and who owns inspection before the first purchase order goes out.