Supplier's report used one panel already separated from the stack. I've asked them to state that.

Can you send representative panels as a stack, including the ordinary surface and flatness variation? The sample should resemble what stores will supply, not the one panel everyone chose because it looked easiest.

And include the stack preparation in the workload. If someone separates every sheet first, that is a different feed option with its own handling time.

Maya, that might still be worth comparing. We shouldn't make automatic stack separation a requirement if manually separated presentation would leave a useful job.

Agreed. I meant price it honestly, not rule it out. Ella's original proposal starts with a stack, so changing the start needs to be visible.

Both feed options are in the comparison. Representative stack sent with the receiving-fixture details.

Ask them to check the released panel condition too, not only whether it appears to sit in the nest. A flexible panel can look settled while still being held out of shape.

What identifies a double panel in their proposed check?

Supplier proposes a thickness check, but has not demonstrated its single, double or uncertain indications yet.

Have them show those known cases and the response to uncertainty, with the actual panel surfaces. I wouldn't choose the detector from its name alone.

Stack trial produced a double pickup. The proposed check identified the known single and double samples, but one damaged sample gave an uncertain indication.

Uncertain may be the correct indication for that sample. What does the proposed process do with it, and was it inside your accepted incoming range?

Outside the agreed incoming range. It held for operator review, as specified. Support at release still leaves a corner lifted on an accepted single panel.

Keep that support problem open in both feed options. Manual separation might remove one task, but it will not repair the way an accepted single panel lands in the fixture.