Production Equipment Run-Off & Acceptance

New automated production equipment was being introduced into a high-volume manufacturing area. The equipment had been specified and purchased; my involvement began at the point where it had to be proven before it could be accepted and released to production.

The reason this matters more than it sounds: equipment acceptance is the last point at which a supplier still has both the obligation and the leverage to fix something. After buy-off, every remaining defect becomes the plant’s problem, paid for in downtime, scrap and engineering time for the life of the asset. The acceptance decision is cheap. Everything after it is not.

The specification described what the equipment was supposed to be. It did not, in a usable form, describe what would constitute proof — which is the more common failure in capital projects. “Capable of producing the part to drawing” and “target cycle time” are procurement language, not acceptance criteria. They cannot be passed or failed by two people looking at the same data.

So the first deliverable was not a test. It was a translation: every requirement restated as a characteristic, a method, a sample size and a pass condition, agreed with the vendor before anyone stood in front of the machine.

• The installation date was committed and there was production demand waiting behind it — schedule pressure ran against the acceptance decision the entire time.
• Some characteristics could only be verified at volume, which meant part of acceptance necessarily happened after installation and had to be structured as a staged buy-off rather than a single event.
• Acceptance criteria had to be negotiated with a vendor who had a commercial interest in a broader definition of “acceptable.”
• Repeatability had to be demonstrated across operators and shifts, not by the vendor’s best technician.

I defined the acceptance criteria, led the run-off, owned the punch list and its disposition, and made the buy-off recommendation. I also owned the PFMEA, control plan, work instructions and inspection criteria the equipment would run under once accepted — which is the reason acceptance and production readiness were treated as one decision rather than two.

Run-off produced the usual distribution of findings: a small number of genuine non-conformances, a larger number of workmanship and documentation items, and a handful of items that were really specification ambiguities rather than defects — places where the plant and the vendor had read the same sentence differently.

The decisive finding was not a broken part. It was that the equipment met cycle time in a demonstration run and did not sustain it over a production-length run, because a recovery step that was trivial at low volume accumulated. That distinction only appears if the acceptance criterion specifies a sustained rate over a defined quantity rather than a cycle time. It was the clearest argument I have encountered for writing the sample size into the criterion.

Capability was the other open question. At run-off there was not enough stable production to compute a meaningful capability index, and computing one anyway would have produced a number with no statistical standing. Rather than accept a weak study or delay the installation, capability was carried as an explicit open acceptance item with a defined closure requirement.

The equipment was released to production on the committed schedule, with the capability criterion closed on real volume data shortly afterward and the punch list closed against defined evidence. Nothing was accepted on the strength of a verbal assurance.

The durable output was the criteria structure itself. Writing characteristic, method, sample and pass condition for every requirement meant later equipment could be specified and accepted against something that could actually be failed — and it made the ambiguities visible during negotiation rather than during run-off.

A requirement without a sample size is not an acceptance criterion. The sustained-rate finding existed only because the criterion specified a quantity. A cycle-time number on its own would have passed.

Negotiate the criteria, not the findings. Every hour spent agreeing pass conditions before run-off saves several during it. Arguments about whether something is a defect are almost always arguments about a criterion nobody wrote down.

“Pass with an open item” is a real disposition, and it needs rules. It is legitimate when the missing evidence is missing rather than unfavourable, when containment covers the interval, and when closure has a defined sample and date. Without those three, it is just a failure with better manners.

Acceptance and production readiness are one decision. Accepting equipment without the control plan, work instructions and maintenance plan in place moves the risk rather than closing it.