[esaym]

>Do field technicians have dependable tools and methods of measuring with such high degrees of precision?

Yes, it is basic stuff in most mechanics. For instance, the valve clearance on most car engines is going to be .002-.005.

I worked aviation maintenance for years. Most larger planes are held together with 'hiloks' [0] They are an "interference" fit.

For a quarter inch hilok, you'd first pilot drill a hole around 3/32" (#40 drill bit), then up drill it to .242 (C drill bit), then swap out to a .242 to .247 "reamer" and ream the hole to .247. The hilok itself is around .248-.249" so you can't just "push" it in but tap it in with a hammer or rivet gun. There was a "go-no-go" gauge. The "go" end was .246 and the "no go" end was .248. So yes, tolerances of 0.001" are quite common.

[0] https://duckduckgo.com/?q=hilok&ia=web

[Animats]

Adjustments to tolerance are something aviation still has, but which has disappeared from consumer products. It's too labor-intensive.

Here's the "Adjustments" manual for a Model 15 Teletype, the 1930-1958 model. I've restored two of those. Making those adjustments isn't that difficult, just time-consuming. Adjustment to 0.002 in. is sometimes required, but it's not that hard. Most Teletype adjustments can be set with feeler gauges. You don't need a micrometer.

[1] http://aetherltd.com/public/model15manuals/138_Model15_Adj_O...

[BinaryRage]

I was staggered by how much adjustment was required when watching CuriousMarc's excellent Model 19 restoration series, seems incredibly finicky.

https://www.youtube.com/watch?v=_NuvwndwYSY&list=PL-_93BVApb...

[Animats]

I just put up a few stills of my own restorations, some short demo videos, and Github repositories for the interface board and the software. I had no idea anyone would want to watch four hours of Teletype repair videos.

Model 15 and 19 Teletypes are not hard to work on. Everything comes apart easily; it's all screws and lockwashers. They're human scale; it's not like building surface mount electronics or repairing an iPhone, where you work under a microscope. The adjustments aren't that finicky, except for a few near the selector magnet. It also helps that the whole thing is unidirectional - there's a straightforward path from input signal to typebar hitting the paper, and you can work through problems in order. As a nice feature, movement is powered in one direction and spring-loaded in the other, so if something gets stuck, it's just stuck in the operated position and doesn't get bent or broken.

Aircraft are built like that. Some parts require careful adjustment, but there's almost always an easy way to check that you got it right. Because, after all, you can't fix it in flight.

Mechanical design has a design philosophy embedded in it. If you work on complex mechanical systems, you can sometimes get a feel for how the original designer thought. Good machinery design is not a common skill. All the good Teletypes were designed by only two people - Howard Krum and Ed Kleinschmidt.

Very few people study this any more in the US, which hurts when you need to design production machinery.

[perl4ever]

Tolerances and path dependence are the fundamental reasons why I've always gravitated towards software and programming rather than building/repairing physical stuff.

It's interesting to me how getting the right physical configuration of something so it's reliable kind of encodes the complex procedure that gets it there.

It's so much harder than a program where you can identify all the parts and modify them in any order. And whenever you make a mistake, you can generally just reach in an fix it, rather than redoing a long disassembly or something.