How can teams define acceptance criteria for 3D curved-glass hot-bending machines?

2026-11-03 11:42:00

Leave a message

A sample of curved cover glass arrives, the process engineer runs it for a week, and then comes the question that decides whether the machine goes on the purchase order: did it pass or fail? If nobody on the project team can answer that in writing, the evaluation drifts, trial after trial produces opinions rather than results, and the equipment supplier is blamed for a decision that was never defined. For anyone specifying a 3D curved glass hot bending forming machine, agreeing on acceptance criteria before the machine reaches the factory floor is cheaper than negotiating afterwards.

Start from the specification sheet, not from opinions

The most reliable acceptance criteria are the numbers the supplier already commits to on the product page. On PowerDe's machine, the published parameters include a cycle time of under 180 seconds per piece, an applicable size of 14 inches and within, a forming temperature range of 600–850 °C, dual high-frequency heating, and oxygen content in the cavity of 50 ppm or less. The same page lists a product pass rate of 80% or higher and an equipment failure rate of 2% or lower, and notes 30 molds in the normal cycle.

Every one of these is a candidate acceptance line. The job of the buying team is to decide which ones will actually be measured, by whom, and on what sample. A criterion such as "forms correctly" is not testable; "produces parts that meet our drawing's curvature tolerance over a continuous run at the stated cycle time" is.

Match the machine envelope to the parts you actually make

Before any performance number matters, the machine has to fit the product family. The applicable-size limit published for this machine is 14 inches and within, so a team evaluating it for automotive center displays or large tablet covers should verify against their largest part dimensions first, and should raise anything outside that envelope with the supplier at specification stage rather than at FAT. The same applies to glass thickness, coating, and the mold set: the machine monitors mold thickness before discharging, which is a feature worth writing into the acceptance plan as a check, not just a brochure line.

Turn process control into measurable criteria

Hot bending is a thermal process, so the acceptance plan should say how the thermal and atmosphere conditions are proven, not only that they exist. Two functions described on the machine are directly useful here:

  • Heating temperature and heating time can be set and displayed on the touch screen. Acceptance should require that recipe values be recorded for each trial batch, so a good run can be repeated and a bad one traced.
  • Circulating oxygen detection runs on four nitrogen-filled chambers, with the oxygen value controlled by a controller. Acceptance should require the recorded cavity oxygen reading to stay at or below the 50 ppm figure during the qualifying run.

Written this way, process control becomes evidence. The team leaves the trial with logs, not impressions.

Define the trial itself: sample, duration, and who measures

Ambiguity about the test is the most common source of disputes. Before the machine ships, the two sides should agree in writing on: the glass grade and part number to be run; the number of consecutive pieces or hours that constitute a qualifying run; the pass-rate statistic being judged (the published line is a product pass rate of at least 80% and a failure rate of no more than 2%); which dimensional and visual checks apply, and with what gauges; and who from each company signs the result. If your organization uses hardware testing services or an internal quality team, name that owner in the acceptance document now.

Put the documentation requirements in the contract

A machine that passes once is not the same as a machine whose acceptance is proven. The acceptance package should collect the trial recipe and screen records, oxygen readings, pass-rate and failure-rate calculation sheets, the dimensional report for the sample parts, and a signed list of any deviations. Where a specification in the contract differs from what is printed on a public product page — model wording, parameter units, or the size envelope — the contract governs, and the discrepancy should be closed in writing before the trial. PowerDe's engineering services, from product development and engineering support to its semiconductor equipment line under semiconductor process equipment, cover customized configurations, so teams with non-standard parts can confirm the exact scope, parameters, and applicable size for their application before ordering.

A short checklist for the acceptance document

  • Size and part coverage confirmed against the machine's applicable-size range.
  • Cycle time, pass rate, and failure rate stated as numbers, with the measurement method named.
  • Temperature window and oxygen content captured from machine records, not observation.
  • Trial sample, run length, gauges, and signatories fixed before shipment.
  • Documentation package listed in the purchase contract, including how deviations are handled.

Teams that write these five things down usually find acceptance goes quietly: the numbers were agreed, the run produced the records, and both sides signed the same page.

send inquiry
Contact Us If there are any questions

You can reach us by phone, email, or by filling out the online form below. Our experts will contact you promptly.

Contact Us