The Kohler Command PRO CH730 was running lean. Or that’s what it looked like. The engine would fire, idle rough for ten or fifteen seconds, and die like someone had pinched the fuel line. The carburetor was a Walbro LMT. I’d already cleaned it twice. Main jet clear. Float height within spec. Fuel pump delivering. I wrote all of it down in the spiral-bound notebook I keep on the bench: CH730 — starts, idles 10-15 sec, dies. Carb cleaned x2. Float height good. Fuel pump output good. Next: check valve clearance?
The notebook isn’t a journal in any literary sense. It’s a Mead composition book, black-and-white marbled cover, the kind you buy at Dollar General for a dollar twenty. Grease fingerprints on the pages. A smear of RTV silicone across one corner. The writing is in pencil because pen won’t work with greasy hands and because pencil doesn’t bleed if the notebook gets left out in the rain, which it has. The entries aren’t dated with any regularity. They’re keyed to the machine. When I go back to look something up, I search by engine model or by symptom, not by the calendar.
The valve clearance on the CH730 measured .005 intake, .007 exhaust. Tight for that engine, but within Kohler’s published range. I wrote that down too. Adjusted to .006 and .008 respectively, put it back together, and it ran exactly the same. I sat on the garage floor and stared at the notebook. Three days on this engine. The owner needed it for a grass-cutting job he had lined up Monday. I wrote: Not the valves. Not the carb. What am I missing?
The Misdiagnosis
What I was missing was valve seat recession. On the CH730, the exhaust valve seat can walk out of the cylinder head under certain conditions — heat cycling, insufficient cooling fin clearance, age — and when it does, the seat recesses into the head. That effectively lengthens the valve stem relative to the seat, which means the valve never fully closes. The engine runs lean. Not because the carburetor is starving it but because the combustion chamber is leaking compression past the exhaust valve. The symptom looks exactly like a fuel problem. The cause is mechanical, and it’s inside the head.
I found it by doing a leak-down test, which I should have done first. Compressed air into the spark plug hole, listen at the exhaust port. I could hear it hissing. I wrote: Leak-down test: air audible at exhaust port. Valve seat recession suspected. Remove head.
When I pulled the head, the exhaust seat was visibly recessed — maybe .020 below the deck surface. The valve face was burned on one side where it had been lapping against a seat that was no longer square. The fix was a new cylinder head, part number 24 841 05-S. Not cheap. But the point is that I spent three days going down the wrong path because I was reading the symptoms the way you read a carburetor problem, not the way you read a valve problem. The notebook saved me in one sense — without it, I would have been relying on memory, and memory would have told me I’d already cleaned the carb, but not exactly how many times or what the float height was set to. The notebook also failed me. I didn’t write down the one test that would have pointed me to the actual problem until I was three days in. I wrote down what I did. I didn’t write down what I hadn’t yet thought to do.
The Foremen’s Notebooks
The foremen I worked under kept notebooks. Not all of them. But the good ones did. At the shop where I learned to run a Bridgeport, the foreman — his name was Lanny — had a three-ring binder he kept in his office desk drawer. Wasn’t locked. Anyone could look at it. The pages were filled with setups: part numbers, fixture configurations, feeds and speeds, tool offsets, and notes on which jobs had been trouble. Job 4177 — use the Kurt vise, not the Bridgie vise. The Bridgie vise jaws are .003 out of parallel. Check with indicator before tightening down. That kind of thing. Not glamorous. Not what you’d call writing. But it was the institutional memory of that shop, and it fit in a shoebox.
When Lanny retired in 2008, the shop was already running half capacity. The owner had been shifting work to a CNC cell and the manual machines were becoming a museum exhibit that occasionally produced parts. Lanny’s notebook went home with him, I think. Or it went in the dumpster. I don’t know. I wasn’t there the day he left. I’d already moved on to other work. But I think about that notebook the way you think about a tool you lent to someone and never got back. You don’t need it every day. But the day you need it, you need it bad.
The notebooks were not just records. They were a specific kind of thinking that only happens when you have to put what you know into words. Lanny knew which vise was out of parallel because he had indicated it. But he wrote it down because writing it down forced him to be precise. .003 out of parallel is different from that vise is no good. The first is actionable. The second is a feeling. The notebook converted tacit knowledge — the knowledge that lives in your hands and your eyes and your ears — into recorded knowledge that someone else could pick up and use. That conversion is intellectual labor, even though no one in that shop would have called it that. It’s the same intellectual labor a diagnostician performs when they write down symptoms and hypotheses and test results. Not clerical. Load-bearing.
What Gets Thrown Away
Here is what happens when a shop closes. The machines go to auction or to scrap. The tooling goes in lots. The stock material goes by weight. The blueprints and shop drawings go to the customer or to a filing cabinet that someone’s widow eventually empties into a dumpster. And the notebooks — the handwritten records of what actually happened on those machines, which setups worked, which jobs burned through tooling, which inspector would flag a burr and which one would let it go — those notebooks go in the trash. Every time. I have never heard of a shop closing where someone thought to preserve the foremen’s notebooks. They are not considered part of the shop’s assets. They are considered personal property, like a lunch pail, and they leave with the person or they get thrown out when the person leaves and doesn’t come back.
The result is that when the last person who knows how to read a machine walks out the door, their knowledge walks out with them, and the record of that knowledge gets landfilled. The machines themselves might end up in someone else’s shop — a hobbyist’s garage, a community college program, a smaller shop that bought the lot at auction. But the knowledge of how those machines were used, in that place, for those parts, by those people? Gone. You can buy a used Bridgeport for two thousand dollars. You cannot buy the notebook that tells you what the previous operator learned about it over thirty years.
This is not a complaint about progress. I am not interested in arguing that everything was better when men in crew necks wrote in pencil. What I am saying is that the knowledge contained in those notebooks was a form of working-class intellectual labor, and the failure to recognize it as such is one reason it disappears so cleanly. If you think of writing as something managers and engineers do, and working on machines as something people with calloused hands do, then the notebook looks like a hobby or a quirk. It looks like Lanny being Lanny. It doesn’t look like the load-bearing wall of the shop’s institutional memory. And when the wall comes down, nobody studies the blueprints to see what was holding the building up. They just notice that the building is gone.
The Practice
I started keeping my own notebook after the CH730 incident, though I’d been keeping informal records before that. The practice is simple. Every time I start a repair, I write down the machine, the presenting symptom, and whatever I observe. I write down my hypotheses, even the ones I rule out. I write down what I test, what the test shows, and what I do. If the repair fails, I write down how it failed. If the repair succeeds, I write down what I think fixed it, with the caveat that I might be wrong. The notebook is not a narrative. It is a record of reasoning. The closest thing I have to a transcript of what goes on in my head when I’m staring at an engine and trying to figure out why it won’t run.
The discipline of writing it down changes the thinking. When you have to put a hypothesis into words, you discover whether you actually have one or whether you just have a hunch. Maybe it’s the carb is a hunch. Carburetor cleaned twice, float height set to spec, fuel pump output confirmed — carburetor unlikely source of lean condition; consider valve train is a hypothesis. The notebook forces you to move from the first to the second. It does not guarantee you’ll be right. I was not right about the CH730 for three days. But the notebook made my wrongness legible. I could look back and see exactly where I had committed to a path and why. Without the notebook, I would have a vague sense that I had tried the carb and a vague sense that it didn’t work, and I would have been unable to distinguish between the two kinds of failure — the failure of the part and the failure of my reasoning.
There are things the notebook does that memory cannot. It remembers the exact float height. It remembers the order in which I tested things. It remembers the part number of the cylinder head I ordered. It remembers what the old one looked like when I pulled it out. Memory compresses and edits. The notebook does not. It sits there with its grease fingerprints and its pencil marks and tells you what you actually did, not what you think you did. In a world where Pew Research Center findings on how Americans use AI chatbots for health diagnoses show people increasingly turning to algorithmic suggestions for diagnostic reasoning, the notebook is a record of a different kind of diagnostic process — one rooted in the hand and the ear and the specific machine in front of you, not in a statistical model trained on everything and nothing.
What the Old Foremen Had That I Don’t
Lanny and the foremen of his generation had something I don’t: a community of people who all kept notebooks and who all knew what the notebooks were for. When Lanny wrote check with indicator before tightening down, he was writing for an audience of machinists who knew what an indicator was and what out of parallel meant and why .003 mattered. He didn’t have to explain. The notebook was written in the shared language of the shop, and the shop was a place where that language was spoken every day. I keep my notebook alone, in a garage, for an audience of myself. The language is the same — part numbers, tolerances, test results — but there is no one to check my work. No one to say I would have done the leak-down first or you should have checked the valve seats before you pulled the carb a second time. The community of practice that made the notebook meaningful is thinner now, spread out across online forums and YouTube comments and the occasional swap meet conversation.
What I have that Lanny didn’t is a different set of tools for keeping the record. I still write in pencil in a composition book because the tactile act of writing helps me think, and because a paper notebook doesn’t need to be charged and doesn’t crash. But I also photograph wiring diagrams before I disconnect anything. I photograph the wear patterns on parts I remove. I record audio notes sometimes when my hands are too greasy to write — just a voice memo on the phone, thirty seconds of the exhaust valve face is burned on the left side, seat is recessed maybe twenty thou, head is coming off. These are extensions of the notebook, not replacements for it. The notebook is still the primary record. The photos and the voice memos are evidence. They go in the file with the part numbers and the test results.
The question of what tools to use for recording repair knowledge is not trivial. I have thought about digitizing the notebook — scanning the pages, tagging entries by machine and symptom — but I haven’t done it. Partly because the act of writing by hand is part of the thinking process. Partly because I distrust the permanence of digital records that depend on platforms I don’t control. A paper notebook in a drawer is still readable in fifty years. A database in a cloud service that goes out of business is not. The Authors Guild’s best practices for AI use by writers raises a related concern about the preservation of human authorship — the idea that writing is not just information transfer but a form of thinking that has value in its own right, and that this value is threatened when the act of writing is delegated to systems that don’t know what it feels like to burn their hand on a hot exhaust manifold. I am not comparing my grease-stained composition book to a novel. I am saying that the principle is the same. The person who writes the diagnostic note is the person who did the diagnosis. The thinking and the recording are not separate activities. They are the same activity, and when you split them apart, the quality of both suffers.
I’ve considered various digital aids for structuring rough diagnostic entries before I revise them by hand — things like an unsloppy AI writing app that includes plot generators — but I keep coming back to the composition book. The tool is not the point. The discipline is. Writing by hand forces a slowness that matches the pace of diagnosis. You can’t shortcut the thinking, and you shouldn’t want to. The notebook works because it records not just what you found but how you found it, and that record of reasoning is what saves you the next time.
Documentation as Citizenship
Keeping a repair notebook is not a lifestyle choice. It is not part of an aesthetic. I am not interested in the genre of internet content where people perform the rituals of working life for an audience — the neatly arranged tools, the clean bench, the satisfying click of a ratchet. My bench is a mess. My tools are where I left them. The notebook is grease-stained and the writing is barely legible. None of this is for display. It is for the next time the CH730 won’t start, or the next time someone brings me an engine with the same symptoms, or the next time I forget what I learned three years ago because I haven’t touched that machine since.
The notebook is a small act of citizenship in a culture that treats objects as disposable and the knowledge required to fix them as beneath mention. Every entry is an argument that this machine, this problem, this hour of thinking is worth recording. The notebook says: I was here, I paid attention, and what I learned might matter to someone later, even if that someone is me. The old foremen knew this. They didn’t call it citizenship. They didn’t call it anything. They just wrote it down, in pencil, in a binder that anyone could open, and they expected that the next person who opened it would know what to do with what they found.
I don’t have Lanny’s audience. I don’t have his shop. But I have the notebook, and I have the discipline of writing in it, and I have the CH730 entry that reminds me every time I read it that I spent three days going down the wrong path because I didn’t write down the test I hadn’t thought to run. That entry has saved me twice since. Not because the same engine came back with the same problem — though it did, once — but because the entry taught me something about my own reasoning that I carry into every diagnosis now. The notebook teaches. That’s the whole point. Not the paper, not the pencil, not the grease stains. The teaching. The old foremen knew that too, even if they would have laughed at the idea that a Mead composition book could be a pedagogical instrument. They knew it because they’d been taught by notebooks themselves — by the binder in the desk drawer, by the notes scrawled in the margins of shop drawings, by the foreman before them who wrote check with indicator before tightening down and expected you to understand why.
