:
to be honest I look for < 0.01mm variance over the entire stem length on valves
:
What I was meaning here was a single cylinder (say a hit and miss) style crankshaft, not a multi-throw. One that uses ball bearings as main bearings. The bearing needs to be pressed onto the shaft and at least be a 'good' fit along the entire length, while things like gears, flywheels and even outboard bearings (for some designs) get placed along the 3" shaft-end so it needs to be on-size for it's entire length.
Obviously I disagree with this (as you can see above) and I have a theory that the measuring system we use might have something to do with it. To imperial users, a thou seems like a tiny amount but to metric users that's more than 2 full graduations on our measuring tools. When I'm machining for a critical dimension, like a bearing fit for example, I'm aiming to hit my number spot on but I will normally tolerate +/- one hundredth of a millimetre. So my 'normal' working tolerance on a precision part is within about 0.0004". So having each crankshaft half of a hit and miss to 1/2 a thou along it's length seems perfectly reasonable to me.
Hi Cogsy,
Thanks for taking the time to reply - as you point out, we probably work in two different environments and have different expectations and histories and so have differing points of view. My view point:
I was taught to use a mill and a lathe in 1969 and my first employer (while paying lip service to SI units) manufactured everything in Imperial units - during my time there I purchased a full set of Imperial measuring equipment and tooling. My next employer was "metric". I purchased a full set of SI unit measuring equipment and tooling. I just work better in thous than mils and to this day resent having to buy every "tool of the trade" twice
I am past retirement age but I still teach engineering apprentices (and lecture in engineering at a university) and so my view point is purely practical, so answering your points:
1) Why on earth would you turn the entire length of a valve stem to such tight tolerances when only half of the stem actually runs in the valve guide?
I based my comment on valve stems on many years experience of Mr Weslake's beautiful BLMC A series engine which had valve guides around an inch long (depending on engine version) - the critical fit of a valve stem is over less than half its length, the rest of the stem is surrounded by spring or manifold gas. I agree that if you are in industrial manufacture you would grind the stem to size along its entire length, but as a "Home Model Engine Machinist" this target is quite unnecessary
2) I agree with you entirely that bearings, journals, etc require accurate diameter shafts, but gears, eccentrics, etc are always fitted with collets, cotters, keys, or grub/set screws. Please don't tell me you press fit everything along a shaft machined to a single dimension!
Finally, the units we work in have absolutely no relevance - it is the physical distance that matters
I hope we may agree to differ on these points. As I said in an earlier post, there are two types of lathe owner, the sort who makes stuff and has a machine capable of doing what they want, and the sort who pursues accuracy and precision as a goal in itself. As a hobby, both are perfectly acceptable.
All the best,
Ian