When we commissioned our first 5-axis cell in 2023, the business case on paper was simple: fewer setups, fewer fixtures, shorter cycles. Two years and roughly 1,900 jobs later, the machine delivered — but almost none of the savings arrived where the spreadsheet said they would.
Average lead time on complex fixture work fell from 18 working days to 11. Only about a third of that came from spindle time. The rest came from decisions we had to make around the machine, and those are the ones worth writing down.
Setups were never the real bottleneck
On a 3-axis machine, a typical housing needed four setups. Each one meant a fixture, a re-datum and an operator waiting on the CMM before the next operation could start. We assumed collapsing four setups into one would collapse the schedule. It did not — because the queue in front of the machine did not change.
What actually moved the needle was that a single-setup part no longer has to come back. It leaves the cell finished, so it stops competing for capacity with itself. Once we stopped scheduling parts and started scheduling the whole job to completion, the queue drained.
Three changes that carried the savings
- Fixture library over one-off fixtures. We standardised on a zero-point system and now reuse roughly 70% of setups instead of fabricating new plates for each order.
- Probing moved inspection into the cycle. In-process probing replaced two of every three CMM trips, which removed the single biggest source of waiting in the shop.
- Quoting changed. We now quote from a simulated cycle rather than a shop estimate, so the promised date and the achievable date are the same number.
“The machine gave us capability. The scheduling change gave us the lead time. If we had bought one without doing the other, we would have written off the investment.”
What we would do differently
We under-invested in CAM training for the first six months and it cost us. Programming a 5-axis toolpath badly is slower than programming a 3-axis toolpath well, and for a while our new machine was the least productive asset in the building. If you are planning a similar step, budget the training before the tooling — the machine will wait, the learning curve will not.
If you are weighing up a similar investment and want to compare notes on the numbers, our engineering team is happy to talk through what we measured.