← Back to Field Notes

Knowledge

February 24, 2026 · The Bloomfield Team

How One Machinist's Notebook Became a Searchable Knowledge System

The notebook was a standard 8.5" x 11" Mead spiral, college-ruled, with a green cover that had faded to olive after years in a machinist's tool chest. It belonged to a CNC lathe operator at a 65-person job shop in central Indiana. He had been filling notebooks like this one for 23 years, one per year, sometimes two. The current one sat in the top drawer of his Kennedy box, next to his calipers and a mechanical pencil.

Inside, every page was dense with notes. Part numbers. Tooling selections. Feed rates that deviated from the program because the material ran differently than expected. Notes on which collet pads to use for specific bore sizes. Sketches of custom soft jaws. Warnings about specific customer part numbers: "4418-R: bore finish sensitive, use new insert every 40 pcs." Supplier notes: "Ryerson 303 SS runs gummy, switch to 303 from Metalwest."

When the shop owner learned that this machinist was planning to retire in 14 months, the first reaction was to ask him to type up his notes. The machinist agreed to try. After two weeks, he had completed six pages in a Word document and told his supervisor that typing up 23 years of notes was going to take longer than the 14 months he had left.

He was right. The notebook approach needed a different solution.

What the Notebook Contained

The shop owner and the machinist sat down over a series of lunch breaks to inventory what was actually in those notebooks. The contents fell into five categories.

Tooling selections and modifications. For hundreds of part numbers, the machinist had recorded which insert grade worked best, what the optimal speed and feed settings were, and where he had deviated from the standard tooling list. On many parts, his selections differed from the CNC program defaults because he had found, through years of running these jobs, that a different approach produced better results. For a family of 17-4 PH stainless valve stems, he had developed a complete alternate tooling strategy that reduced cycle time by 22% compared to the programmed approach.

Setup notes and fixturing solutions. The notebook contained sketches and dimensions for custom soft jaws, collet configurations, and workholding modifications for difficult parts. For a complex medical device housing that the shop had run for eight years, he had documented three different workholding approaches: the original, the one that failed, and the current version that solved a deflection problem on the thin wall section.

Material behavior notes. Over 23 years of running production on dozens of alloys, the machinist had built a detailed understanding of how specific materials from specific suppliers behaved under cutting. He noted that 6061-T6 aluminum from one supplier consistently ran softer than the same alloy from another, requiring a feed rate adjustment to prevent built-up edge on the tool. He recorded that 304 stainless from a particular heat lot had caused tool breakage on a deep bore operation, and that reducing speed by 15% and switching to a coated carbide grade solved the problem.

Quality notes and customer preferences. For repeat customers, the machinist had accumulated knowledge about what would pass inspection and what would be rejected, knowledge that went beyond the drawing tolerances. One customer's inspector consistently measured surface finish with the stylus traveling perpendicular to the lay, which produced readings 8 to 12 Ra units higher than measurements taken parallel to the lay. The machinist compensated by targeting a tighter finish than the print required, ensuring that the parts passed regardless of measurement direction.

Problem history and root causes. When something went wrong on a job, the machinist wrote down what happened, what he thought caused it, and how it was fixed. These notes formed a diagnostic reference that he consulted whenever a similar problem appeared. For a bearing sleeve that occasionally exhibited chatter on the OD finish, he had tracked the problem across four separate occurrences over seven years and identified that it correlated with collet pad wear beyond 0.002", a threshold that was well within the normal replacement interval but apparently enough to introduce the vibration on that specific part geometry.

From Notebook to System

Typing the notebooks into a Word document would have produced a long, unsearchable text file. The knowledge would have been digitized but not usable. The format would have been the same as the notebook, linear and organized by date, which is the least useful way to retrieve operational knowledge.

The approach that worked started from a different direction. Instead of digitizing the notebooks page by page, the process started with the shop's existing digital records: job history from the ERP, inspection data from the quality system, tooling records, and customer order history. This data was structured, organized by part number and job number, and covered the same 23-year period as the notebooks.

The machinist's notebook entries were then connected to the corresponding job records. A note about tooling on part number 4418-R was linked to the 47 jobs in the ERP history for that part. A material behavior note about a specific supplier's 304 stainless was connected to every job that used material from that supplier. A quality note about a customer's inspection preferences was linked to that customer's entire order history.

This connection transformed the notebook from a personal reference into a contextual knowledge layer on top of the shop's production data. The knowledge was no longer organized by date. It was organized by part, by material, by customer, and by problem type.

What the System Looks Like in Practice

When an operator opens a new job for part number 4418-R, the system presents the complete knowledge package for that part: the tooling selections from the machinist's notes, the setup configuration from the last successful run, the quality notes about bore finish sensitivity, and the historical job data showing cycle times, scrap rates, and any quality issues across all 47 previous runs.

When a programmer is writing a program for a new part in 17-4 PH stainless, the system surfaces the machinist's alternate tooling strategy for that alloy, along with the job data showing the 22% cycle time improvement it produced. The programmer can decide whether to adopt it, and the rationale for the approach is documented and available.

When a new operator encounters chatter on a bearing sleeve and types "chatter on OD finish" into the system, it returns the machinist's diagnostic history: four occurrences over seven years, all correlated with collet pad wear beyond 0.002", with the solution documented. The new operator checks the collet pads, finds they are at 0.0025" wear, replaces them, and the chatter disappears. A problem that might have taken two hours of troubleshooting was resolved in ten minutes.

The machinist's 23 years of accumulated knowledge became available to every operator, programmer, and estimator in the shop, searchable by part number, material, problem symptom, or customer.

What Happened After the Retirement

The machinist retired on schedule. In the six months following his departure, the shop tracked the impact on the CNC lathe department where he had worked.

Setup times on the parts he regularly ran increased by an average of 8%, compared to the 25% to 35% increase that shops typically see after losing a senior operator. The new operators referenced the knowledge system during setup and caught most of the deviations from standard that the veteran machinist had identified over the years.

Scrap on his regular parts increased by 3% in the first quarter, then returned to the baseline by the second quarter as operators learned to check the system for material and tooling notes before starting a run. Without the system, comparable retirements at other shops in the region produced scrap increases of 15% to 25% that persisted for 12 months or more.

The quality rejection rate from the customer with the surface finish inspection sensitivity stayed at zero. Every operator who ran that part saw the note about measurement direction and targeted the tighter finish. The kind of knowledge loss that typically costs hundreds of thousands of dollars was prevented by a system that made one person's experience available to everyone.

The Broader Lesson

Every manufacturing shop has notebooks. They might be spiral-bound, or they might be sticky notes on a machine, or they might be a folder of personal files on a shared drive. In every case, they represent decades of accumulated knowledge that belongs to an individual rather than to the organization.

The knowledge in those notebooks is the most valuable data a manufacturing operation possesses. It is specific to your parts, your machines, your materials, and your customers. It was learned through production experience, which means it reflects reality rather than theory. And in most shops, it is completely inaccessible to anyone other than the person who wrote it down.

Converting that knowledge from personal reference to organizational asset does not require a massive documentation project. It requires connecting what the experienced people know to the production data the shop already generates, and building a system that delivers that knowledge contextually, at the moment someone needs it, organized around the work they are doing.

The notebook was always the right instinct. The machinist understood that the knowledge mattered and was worth recording. The format was the limitation. A spiral notebook cannot be searched, cannot be shared, and cannot be connected to the digital systems that run a modern shop. When that limitation is removed, 23 years of expertise become a permanent competitive advantage.