Why This Checklist Exists
A sample PO should not only purchase hardware. It should freeze the minimum engineering boundary needed to judge whether the sample is acceptable, repeatable, and ready for the next program gate.
Review what to request before issuing a custom humanoid actuator sample PO, including drawing revision, CTQ limits, test reports, sample variants, NRE scope, packaging, and pilot assumptions.
A sample PO should not only purchase hardware. It should freeze the minimum engineering boundary needed to judge whether the sample is acceptable, repeatable, and ready for the next program gate.
Checklist
The goal is to make acceptance measurable before the quote, PO, sample, or pilot gate creates pressure to accept incomplete proof.
| Area | Buyer Question | Define Before Review | Evidence / Output | Red Flag |
|---|---|---|---|---|
| PO purpose | What decision should this sample make possible? | Feasibility sample, fit-check sample, EVT sample, DVT sample, pilot-reference sample, or production-intent sample. | Sample purpose, decision gate, required report depth, and what happens after pass or fail. | The PO says "sample" but no one agrees whether it is for fit, performance, validation, or pilot release. |
| Drawing and BOM revision | Which design baseline is the supplier building and reporting against? | CAD revision, 2D drawing revision, BOM revision, open items, buyer-owned components, approved alternates, and frozen interfaces. | Revision-controlled sample baseline and open-trade-off list. | Files are sent by email without a clear revision owner or freeze date. |
| CTQ and acceptance limits | Which dimensions and performance values decide sample acceptance? | CTQ features, GD&T notes, backlash limit, torque-speed or force-stroke target, thermal limit, noise threshold, and FCT requirements. | Acceptance matrix tied to drawing features and actuator performance targets. | Inspection is requested but no CTQ list or pass/fail value exists. |
| Report deliverables | Which records are included in the sample price and lead time? | FAI, CMM, material record, surface treatment record, torque-speed, thermal, backlash, FCT, calibration, burn-in, and traceability scope. | Report list with owner, format, timing, and whether each report is mandatory or optional. | The buyer assumes reports are included while the supplier quoted hardware only. |
| Sample quantity and variants | How many units and variants are needed to make the next decision credible? | Left/right variants, joint axis, quantity per variant, destructive test units, spare units, and pilot-reference units. | Sample allocation plan and serial-level purpose for each unit. | One best sample is expected to answer fit, performance, endurance, and pilot readiness. |
| Buyer-supplied boundaries | What must the buyer provide for the sample to be meaningful? | Control board, firmware, encoder requirement, harness, mating connector, robot link, test protocol, software, and NDA-controlled drawings. | Buyer-supplied item list, delivery date, compatibility note, and responsibility split. | The supplier cannot complete or test the sample because buyer-supplied parts arrive late or undocumented. |
| Change and NRE rules | What happens when the design changes after PO? | Included engineering revisions, chargeable changes, tooling or fixture NRE, approval owner, and quote reset conditions. | Change-control note and NRE assumption visible before PO release. | Scope drift is handled through informal messages instead of a controlled revision update. |
| Packaging and delivery | Will the sample arrive ready for incoming inspection and bench testing? | Destination, Incoterm, protective packaging, labels, serial marking, document packet, spares, and urgent shipment needs. | Packing rule, delivery documents, serial list, and incoming inspection handoff. | Precision surfaces, connectors, or encoder areas are shipped without a protection rule. |
A sample program can still have unknowns. What matters is separating controlled requirements from open engineering questions before money changes hands.
Put CTQ dimensions, performance limits, report deliverables, and pass/fail ownership into one table rather than scattering them across emails.
If CMM, FAI, FCT, backlash, thermal, or burn-in records are required, list them explicitly with timing and format.
State whether a passing sample triggers EVT, DVT, pilot planning, tooling review, or a revised sample round.
Gate Review
| Gate | Pass Signal | Fail Signal | Next Action |
|---|---|---|---|
| Before quote approval | The quote states hardware, NRE, report scope, lead time, and assumptions separately. | A unit price arrives without the engineering assumptions behind it. | Request a revised quote that separates sample build, engineering work, and evidence deliverables. |
| Before PO issue | Drawing revision, CTQ list, acceptance limits, sample quantity, and report scope are attached. | The PO relies on message history rather than a controlled scope. | Add a one-page PO scope appendix before release. |
| Before shipment | Serials, outgoing inspection, required reports, packaging, and open issues are ready for buyer review. | The sample ships before evidence is checked or linked to serial numbers. | Hold shipment or mark the sample as engineering-only until records are complete. |
| Before reorder | The team knows what passed, what changed, and what must be controlled for pilot. | A second order repeats the same undefined sample scope. | Convert sample lessons into pilot readiness actions before reordering. |
Proof Boundary
Yes, if open items are marked clearly and do not affect the acceptance decision. Unfrozen interfaces, CTQ values, or report deliverables should not be left implicit.
It depends on the sample purpose. Fit-critical housings, bearing seats, output flanges, shafts, and robot interfaces usually need at least a defined inspection scope. Pure feasibility samples may use a lighter report.
The most common mistake is treating hardware delivery as the whole deliverable. For custom actuators, the acceptance method, revision baseline, and evidence packet often matter as much as the physical sample.
Send the drawing baseline, joint role, performance target, and current gate. Mark unknowns clearly so the first review can separate feasible assumptions from missing acceptance data.
Inquiry Email
Include target torque/speed, quantity, and delivery location.
Application Engineer
+8618857971991
Talk directly about drawings, torque-speed targets, and RFQ data gaps.