TID Lot Acceptance Report Generator

Preliminary analysis aid. Check the result against your own analysis and, where needed, a test on the intended part. Full disclaimer.

1. What it calculates

From the parameter readings of a total dose lot acceptance test, one table per parameter with units down the side and dose steps across the top, it computes at every step the sample mean, standard deviation, the exact one sided tolerance factor K_TL for the sample size at that step, the bounds x̄ ± K_TL s, and a pass or fail against the limits; plots every unit, the mean and the bounds against dose with the limit line; states the first dose step at which the bound crosses the limit and the verdict at your specification dose; and produces a report you can print to PDF or download as Markdown, with the statistics as CSV and each plot as SVG. The default criterion is P = 0.99 of the lot with γ = 0.90 confidence.

2. Inputs

• Test identification. Part, manufacturer, lot, test date and facility, dose rate and source, bias and anneal conditions, notes. Printed at the top of the report; the tool does not use them in the calculation.
• P and γ. The pairing from the list, or custom values. Also the specification dose (optional) and whether the dose axis is spaced as listed or proportional to dose.
• Parameter card. Name and unit; upper and lower limits (leave blank if none); whether to evaluate measured values or the shift from pre irradiation against a delta limit; the log option for lognormal parameters; and the data table.
• Data table. First row: an optional unit label header (SN, unit, ...) followed by the dose steps in krad(Si). A header that is not a number, such as "RT anneal" or "100C anneal", is an anneal step and is plotted after the dose steps. Each following row is one unit. Tab, comma, semicolon or space separated; paste from a spreadsheet works. Blank cell: missing measurement. F, fail or x: functional failure, which removes the unit from that step and all later steps.

3. Calculation

• Per step. N, x̄, s over the units with a measurement; K_TL(n, P, γ) = t'_γ(n − 1, z_P√n)/√n from the noncentral t by quadrature and bisection; upper bound x̄ + K s, lower bound x̄ − K s.
• Verdict. PASS when the upper bound is at or below the upper limit and the lower bound at or above the lower limit (whichever are given) and no unit has failed functionally; otherwise FAIL. A step with one unit has no factor and is marked n = 1.
• Log option. Statistics on log₁₀ of the values; bounds transformed back. Values at or below zero are dropped.
• Shift mode. Each unit's pre irradiation value (column headed 0 or pre) is subtracted before the statistics; the delta limit is the upper limit on the shift.
• Report. First dose step where the bound crosses the limit and the last passing step before it; verdict at the specification dose if a step exists there or if the crossing brackets it; a lot verdict across parameters.

4. Assumptions

• At each step the parameter (or its log) is normally distributed across the lot and the units are a random sample of one population.
• Bias during irradiation was worst case for each parameter; the tool has no way to know.
• Measurement repeatability is small compared with the part to part scatter; if not, the bounds are conservative.
• Steps are independent evaluations; the tool does not fit a trend through dose.

5. Implementation notes

• K_TL was checked against the Odeh and Owen tables for n = 3 to 100 and by Monte Carlo coverage; agreement to the third decimal. The n = 2 value for 99/90 is 18.50.
• Everything runs in the browser; the data are not sent anywhere. The example data set is synthetic and marked as such.
• Print uses the browser's print dialog with the page chrome hidden; choose "Save as PDF" there. The SVG download is a standalone file with fixed colors on a white background for reports.
• Column headers with letters mixed into a number (for example "50A") are treated as anneal steps, not doses; write "RT anneal".

6. Limitations

• No overtest factor, design margin or enhanced low dose rate logic; the tool evaluates the criterion at the doses you measured.
• No interpolation between steps: the crossing dose is the first step at which the bound is outside the limit.
• A bimodal lot defeats the normal assumption. Look at the unit lines on the plot before trusting the bound.
• The tool does not know your contract's revision of Method 1019 or ESCC 22900 or which parameters carry post irradiation limits.

7. References

Back to the tool · Total Dose Hardness Priors · Why 1.5 times the specification dose