What-If Limits

What-if limits are scenario edits in an analysis. They do not rewrite source measurements, apply production limits, or create a saved proposal in Decide. You can save the scenario as part of an analysis for later review.

Edit and reset limits

1. Open editable work in Analyze and set the Cpk table to Full lot. Both limit editing and Review impact are disabled in Selected mode.
2. Choose What-if. Edit the test’s LSL or USL and press Enter or leave the input to accept it into the working draft. Scatter and histogram also allow limit editing in their expanded views.
3. Inspect the preview and choose Review impact when you need the full-lot before/after calculation.
4. Choose Done to finish the current round of edits. This keeps the scenario in the current analysis; it is not a server save.
5. Use Create Analysis or Save to preserve the investigation as a server version.

While closing a chart with draft changes, Keep Changes retains them in the analysis, Discard Changes restores the scenario from before those draft edits, and Continue Editing returns to the chart.

Use a test’s revert control (Revert both what-if limits to baseline) to remove its override, or Reset All to remove all scenario overrides. Entering both baseline values manually also removes the test’s override after the edits are accepted. The baseline is the analysis’s effective non-scenario limit data, not a promise that no account spec override exists.

Chart preview is not lot impact

Scatter and histogram respond locally while you drag or edit a limit. Read the footer’s scope labels: Sample, Estimated full-population, or Exact describe different kinds of totals. Checking full-population totals… means authoritative totals have not arrived yet.

Chart transitions and Test yield describe that test’s measurement population. Even exact per-test counts do not say that a die passes the whole test program. An in-range measurement on one test does not rescue a die that fails another test. Do not sum per-test changes to predict lot yield.

Review the lot, not only the visible chart

Review impact evaluates the full lot at the selected snapshot. It is not restricted to the brushed dies or the chart’s wafer filter. Use its server result for the lot-level before/after comparison.

What Review impact compares

The review uses actual die fail sets so that failures shared by multiple tests are not double-counted. It shows before/after yield, rescued and newly failing dies, per-test changes, and affected wafers.

The starting point depends on when you open it:

Current state	Comparison
Editing a changed draft	The scenario kept before this editing round versus the current draft. With no prior scenario, that starting point is the non-scenario baseline.
No new draft changes, including after Done	The non-scenario baseline versus the scenario kept in the analysis.

Check the displayed baseline and proposed limits before interpreting a delta, especially when revising a scenario that already had changes. Keep chart previews separate from this server result while requests are pending or fail.

Save or export the investigation

An analysis save includes its active scenario limits, including edits still in what-if mode. Done is a useful editing checkpoint, not a boundary that excludes unfinished edits from a save. See Saved Work And Sharing.

There are two distinct CSV exports:

• The Cpk table’s scenario CSV exports limit rows for reuse or review. It is not a lot-impact report.
• The impact review’s CSV exports the before/after comparison for external review. It does not apply those limits or approve a release decision.

For a first walkthrough, return to Quickstart. To attach the saved investigation to a lot decision, follow Proposal Review.