# NIST TN 2006: published study figure notes

## Sources and credit

- Primary report: J. A. Slotwinski, W. E. Luecke, E. A. Lass, and A. Possolo (2018), *Interlaboratory mechanical-property study for Cobalt-Chromium alloy made by laser powder-bed-fusion additive manufacturing*, NIST Technical Note 2006, https://doi.org/10.6028/NIST.TN.2006.
- Companion public dataset: William Luecke (2018), *Data from NIST TN 2006*, DOI https://doi.org/10.18434/T4/1502566; landing page https://data.nist.gov/od/id/7394D482FB305D2BE053245706814ED41954.
- The NIST dataset record says its files contain data from report Figures 2, 3, 4 and 6 and Table 4. Its README identifies `TN2006-ILS-results.csv` as mechanical-property results appearing in Table 4 and Figure 3.
- Dataset record license: https://www.nist.gov/open/license. NIST requests appropriate source/photo/image credit and does not endorse this site: https://www.nist.gov/copyrights-disclaimers.

## What is plotted

`nist-tn2006-yield-by-lab.svg` plots every reported `Sy` value for Lab 1 through Lab 5 from NIST's `TN2006-ILS-results.csv`. The column-attributes file defines `Sy` as 0.2% offset yield strength under ASTM E8/E8M and gives MPa as its unit. The five-laboratory subset contains 34 rows: 8, 6, 6, 7 and 7 test-level results, respectively. Dots are the values in the published companion data file, not points digitized from a raster figure. Horizontal/vertical placement and jitter only arrange those values for readability; jitter does not change strength values.

`nist-tn2006-between-within.svg` is a direct transcription of the last ratio column in Table 5 of NIST TN 2006: between-laboratory variability divided by the median within-laboratory variability (`τb / median({σbj})`). Values are 2.7 for tensile strength, 4.9 for yield strength, 3.5 for elongation and 0.5 for elastic modulus. A ratio greater than one means that the report's between-laboratory estimate is greater than its median within-laboratory estimate. These are model-based results for the report's five-laboratory comparison, not universal constants.

## Interpretation limits

The NIST report covers eight participating laboratories overall, but the Table 5 model and ratio plot concern five laboratories using nominally similar laser powder-bed-fusion systems and powder from the same batch. The results describe this alloy, study, process subset and measurement context; they do not rank laboratories, identify one cause for every difference, or qualify a manufactured component for a particular use. The independently authored synthetic demonstration on `strength.html` remains separate and uses no NIST values.

## Reproduction and integrity

Run `python generate-nist-tn2006-study.py` in this directory with Python 3. The script checks the downloaded NIST CSV against its SHA-256 value in the dataset record, checks the attributes/README files against their recorded hashes, confirms the exact Table 5 input values, verifies all 34 plotted rows and the per-laboratory counts, then writes the four accessible SVG variants and a machine-readable validation summary. Downloaded source files are retained unmodified.