A custom machine that runs in the customer's factory is easy to argue about when the test records are thin. If the equipment was built to a drawing rather than a catalogue, the acceptance criteria are written by the project team, and both the buyer and the builder benefit when those criteria are measurable and captured on video. Before a machine is designed, it helps to know what data a buyer will actually ask for during testing of non-standard automation equipment: cycle numbers, tolerances, defect rates, downtime, and the operator's own words.
That is the perspective we apply as a non-standard automation equipment builder. Below is the checklist we ask our own engineers to fill in on every build, so you can use it as an acceptance plan of your own.
Start With The Process, Not The Machine
A test plan that only describes the equipment tends to miss the reason a project fails. Record the process the machine has to reproduce first, because everything else is measured against it.
- Sample parts that represent the real population. Collect the parts your line actually produces, including the out-of-tolerance pieces at the edges of your normal variation. A build tested only on perfect samples tells you nothing about the 3% that matter.
- The full dimensional and visual envelope. Photograph and measure the parts going in, then record the dimensions, surface condition, coating thickness, burr height, or flatness that the machine must handle.
- Input supply condition. Air pressure at the machine, line voltage during shift change, ambient temperature, humidity, and the way parts arrive in a tray, bulk, or on a conveyor all change results.
- The takt you must hit. Write the required cycle time, the shift pattern, and the annual volume the equipment is expected to support.
Data To Capture At Every Stage
A build normally passes through several checkpoints, and each one generates a different kind of evidence. Splitting them up stops a late problem from being argued about in hindsight.
1. Factory acceptance testing (FAT)
This is the run made at the builder's workshop, before disassembly and shipping. Capture:
- A continuous video of a full run, long enough to show at least one changeover and one stoppage recovery.
- Measured cycle time per station, plus the changeover time between product variants.
- The pass rate across a meaningful batch of your real samples, with every rejected part kept, tagged, and photographed.
- The alignment or placement error of finished units, measured with the same gauge the production line will use afterwards.
- Utility readings at the machine inlet, not at the wall socket.
Keep the rejected parts. They are the cheapest training set for the debug round and the clearest evidence in a warranty discussion.
2. Installation and site re-verification
Conditions in your plant are never identical to the builder's. Repeat the cycle time and yield checks with your own utilities, your own operators, and your own incoming parts. Record what was changed during shipping and reassembly, including any recalibration and any software revision number.
3. Trial production run
Run the machine inside the real schedule for a shift or two, and log:
- Stops, with the cause written down at the time rather than reconstructed afterwards.
- The time an operator needed to clear each stop and restart the cycle.
- Scrap, rework, and the rate of first-pass yield across the whole run.
- Whether the operator needed a tool, a manual, or a phone call to complete normal tasks.
4. Handover and after-sales baseline
Close the project with the documents that make future troubleshooting independent of one engineer: a bill of materials with part numbers for wear items, the pneumatic and electrical drawings, the recipe or parameter set that was signed off, the spare-parts list, and the contact route for support. A dated parameter snapshot is what allows a fault six months later to be compared against the accepted state.
What This Changes In The Quotation Stage
The same list read backwards is a design brief. A buyer who hands over sample parts, dimensional limits, target takt, utility data, and a written acceptance method gets a machine that matches that description; a builder who receives those inputs can price the sensors, gauging, and guarding the job really needs instead of the cheapest workable option. Where a project also involves a cleaning step, it is worth noting that our PowerDe cleaning equipment line is developed as purpose-built machinery in the same way, since the process condition drives the specification in both cases.
How We Test Our Own Builds
As Shenzhen PowerDe Automation Precision Equipment, we take projects from R&D through integration and commissioning, and the stages above are the reason our hardware testing service starts with your samples rather than with our catalogue. If you are preparing a request for a custom machine, gather the four data groups — process inputs, acceptance metrics, site conditions, and handover documents — and send them with your enquiry. That is usually the difference between a build that is argued about and a build that is signed off.
Tell us what your line produces and what data your quality team needs from it, and our engineers will tell you which of it can be measured on the machine you are specifying.
