Manufacturing · Process and quality engineers
Your limits are claims, and the records on file can check them.
A tolerance, a control limit and a change interval each state that the process stays inside a line someone drew. Most plants already hold the measurements that would confirm or refute that.
EvidenceIn a public forging record of 81 parts and 18 measured dimensions, one stated limit was missed by all 81 of 81 parts, and 9 of 81 exceeded the length limit on at least one measurement. A 90% range around a prediction of length, built the usual way, covered 73% of parts [62%, 81%] on days it had not seen.
No figure in money comes from this data. The specimen has a cost calculator, and its starting rates are placeholders for your own.
Evidence
- Specimen report Process Calibration Audit: a radial forge The finished deliverable, including the day-2 data-structure check.
- One-page brief Process limits, in one page For quality and process engineering.
- Paper Tool Change Intervals, Audited Against Measured Wear KW-2026-03, for the method.
Offer
Process Calibration Audit
- Input
- The records where a limit was stated and outcomes were logged, exported as they already exist. 6 months or more.
- You get
- A data-structure check on day 2, coverage of each stated limit by regime, and the threshold curve at your own limit with your own rates.
- Timing
- 2 weeks from data receipt
- Reference price
- Fixed, set by the size of the record set. About EUR 6,500 for one line with 2 years of records.
Try it
The Band Checker counts measurements outside a stated tolerance in your browser, on your own CSV or on the bundled process sample (480 synthetic parts against a two-sided tolerance), and your file never leaves your computer.
First step
Send one paragraph describing the series or the records to matthew@mtanti.com, and you will have a quote within one working day. A standard mutual NDA comes before any data moves, and your data is deleted on delivery and never used elsewhere.
What this does not support
The forge is a research testbed: 81 parts over 5 production days, with tolerances that may not be a customer's. It says nothing about your scrap rate, your yield or your costs, and it does not show that in-process monitoring does not work. The intervals treat parts as independent, which parts made on one day are not. No cost in the data set is real.