Is that REALLY the size?

[Paul Alciatore]

Interesting discussion.

I believe that what can be done in a "home shop" is limited only by the knowledge, skills, and ingenuity of the user of that shop, the home shop machinist. The impressively accurate devices of our present day came from yesteryear's shops and labs that were far more challenged in the available equipment than today's "home shops". So I reject any argument that says that any particular level of accuracy is categorically not possible in a home shop.

Where were the first gauge blocks created? The first micrometer? The inch? The meter? The list goes on.

Can I or any of us routinely make parts with a tolerance of +/-0.0001" and REALLY KNOW that they are that accurate? For myself, I would have to say NO. Perhaps on occasion and with a lot of effort and some help from an outside calibration facility, I could. But it would be quite an undertaking and not without some expense. And the others here? I would think a small percentage could and perhaps they even have all the necessary items already in their garage or other home shop. Others who lack the knowledge and/or skills probably could not, even with all the help they could muster. I am not saying they couldn't learn: just that they are not there yet. And most of them will never be there because they have no reason to get there. Their present level is sufficient for what they want to do. And for most of what I want to do.

It is not the home shop. It is the home shop machinist.

[J Tiers]

Even for just comparing an other part it is difficult (unreliable) is the surface finish of that part differs from my part.

I use the half division mostly to compare parts that I have made the same way (small series) to make them more equal (not more accurate).

Yes, "more equal".

Also, the surface finish is a very good reason to discard ANY claims of close tolerance on most turned parts..... At least HSM ones from an old SB etc. A ground finish is best just so that the inherent ridges from turning do not exceed the tolerance.

[BCRider]

I do readings on my micrometer on half a division (0.005 mm (0.0002"). If you can do better and get that same value day after day, you do a better job than me. Beware that 10 mm aluminum will extend 0.23 um for every degree in temperature rise. So better don't hold that (or your micrometer) in your hand when measuring.

And that's a good realistic value. I'm doing the same thing I believe. Well, other when I saw a valid and repeatable difference between two readings on the same part in different spots. And even then it's fair to say that I'm scratching at the edges to see a tenth's worth. Two tenths would be more in the scope. As in checking for parallelism along a turned portion. It's the same surface texture, same temperature and within that time frame I'd like to think a consistent method with the mic.

[BCRider]

Yes, "more equal".

Also, the surface finish is a very good reason to discard ANY claims of close tolerance on most turned parts..... At least HSM ones from an old SB etc. A ground finish is best just so that the inherent ridges from turning do not exceed the tolerance.

Very true. The texture straight off a cutting tool sure isn't where to start looking for more than the starting point for a finely sized part. And that's why I would only be suggesting such tolerances with a much finer finish. In the case of the arbor it was the sizes after a light filing then that stoning trick.

[boslab]

Interestingly the first gauge blocks aka johaneson blocks aka jo blocks were made using a converted sewing machine!, sounded far fetched to me but the instructor was well up in the metrology world so there we go.
mark

[Mcgyver]

Don't know about you, but I frequently have to measure small parts to a tenth. It is easy to do if you have a good tool. I would not use a 100 year old Starrett with .001" graduation for such jobs. I have a very good digital Mitutoyo micrometer, which reads to a .00005" and is accurate and repeatable to .0001". If I really need precision, I would set this micrometer on a correct size gauge block and then use it to measure the part. This method never failed me. You need to watch the temperatures of course.

exactly. yuck yuck, you can't measure tenths in the home shop, . I use an Etalon or mit indicating Mic and mit gauge blocks....How could you not measure to a tenth? I can grind things on my grinder, which I scraped to a tenth myself, that come to a tenth end to end on diameter, its not hard, you just need the right equipment....either that or they rezoned my property to industrial when I wasn't looking lol.

[Richard P Wilson]

Sigh. All you 'homeshop' guys with your precision grinders and tenths measuring equipment might find it hard to believe that theres some of us don't have these things, but somehow, we still manage to get things done.
I don't guarantee that a 2" cylinder bore that I do will be closer than +/- 0.002" to size, but I will assure you that the piston that fits it will fit well.
Even though I can't measure better than 0.001" accuracy, I can get a bearing race (+/- 0.0001") to be a press fit in its housing.

I've seen my grandfather acheive similar results when all he had was a steel rule and callipers, didn't even have index dials on his lathe.
Yes it takes longer, and no, it wouldn't satisfy Boeing, but I'm not working for Boeing, I'm working for me, and it satisfies me (as the guy with the 3" d--k said to his girlfriend)

[Doozer]

Size is a means to achieve a fit.
Fit governs function.
If you control the fit, size does not matter.

-Doozer

[Rich Carlstedt]

Size is a means to achieve a fit.
Fit governs function.
If you control the fit, size does not matter.-Doozer

Exactly ! Bravo !
Generally size is a guess or a general standard ( Like 3"bore ...meaning 3.000 "or whatever)
The fit makes it work
Rich

[J Tiers]

Sigh. All you 'homeshop' guys with your precision grinders and tenths measuring equipment might find it hard to believe that theres some of us don't have these things, but somehow, we still manage to get things done.
........................
Yes it takes longer, and no, it wouldn't satisfy Boeing, but I'm not working for Boeing, I'm working for me, and it satisfies me................

Totally agree.

That was the point of the comments on making a fit even if it isn't to exact "print dimensions and tolerance" ("satisfying Boeing"). The piston and cylinder is a perfect example.... A few thou is neither visible nor important, so long as both parts are made to fit each other (assuming that you do not need interchangeable parts).

I rarely make a project where everything would "satisfy Boeing". Even when I think I did, I am probably wrong.

Oh, yeah.... that grinder thing.... yes, I DO have a grinder. But...... it did not come from G&L, Norton, or even KO Lee. Nope, it is a 120 year old Greenfield T&C grinder base and table, with a Dumore tool post grinder doing duty as a head. I have under $100 in the whole thing, but it sharpens cutters for my small horizontal mill just fine. I need to make a better head for it, and also scrape the ways. Meanwhile I use it as-is, it works for tool and cutter sharpening, and even cylindrical grinding. You don't need to spend a bunch of money.

[Mcgyver]

Sigh. All you 'homeshop' guys with your precision grinders and tenths measuring equipment might find it hard to believe that theres some of us don't have these things, but somehow, we still manage to get things done.
)

Why would you think that? Refuting that it can't be done hardly implies it must done or needs be done (in any particular instance). Its just a point of logic,i.e. saying it can't be done is erroneous.

In the round, a small bit of what I do (some tooling, spindles & P4 bearings fits, quills & tailstock bores come to mind) requires it and I've figured out how to. Most work doesn't and the work gets done (I suspect) the same as in your shop. As said in my first post, the craftsman strives for the tolerances the part or task calls for and no more. Why is that contentious or causing anyone to sigh?

Size is a means to achieve a fit.
Fit governs function.
If you control the fit, size does not matter.

Maybe obfuscation, maybe baloney. maybe just being Doozer lol

For one of's as I'm sure you know, SOP is do the bore, then make the mate to fit. That eliminates the need for the bore to be within tolerance to a nominal dimension. However to achieve the desired fit, the size of the mating part of course matters and you'll obviously have to know how measure the size of both parts and machine the second to size to get the fit you want. The nominal size might not matter, but the size of each part obviously does.

[Doozer]

Maybe obfuscation, maybe baloney. maybe just being Doozer lol
The nominal size might not matter, but the size of each part obviously does.

Really disect what you are saying.
You can have 2 parts that have a certain fit,
and neither one has a size that relates to any
measurement system. They just have to relate
to each other. Are you defining size as having units?

SIZE-the relative extent of something; a thing's overall dimensions or magnitude; how big something is

Fit does not need units.
A sliding fit.
A running fit.
A press fit.
A force fit.

You can make a piston FIT a cylinder to function in a specific application
without using units. Just use ID and OD spring divider calipers.

Size comes in with interchangable parts. Thank you Eli Whitney.
But fit is as old as people have been putting tab A into slot B.

-Doozer

[npalen]

I worked for a machine tool builder back in the late sixties and we would rough turn all the bearing lands on the spindles 0.020" oversize on a tracer lathe and then finish to 0.0005" OD tolerance on the cylindrical grinder. The size 30, 40 and 50 NMTB internal tapers were also left a little undersize and were only finish ground after assembly with the bearings. This produced runout as good as the class 7 bearings were capable of. Still remember using the prussian blue to check the fit and the taper gauge to check size. The spindles were all 4140 pre-hardened and 5-1/4" OD chunks for the 50 NMTB IIRC.

[Mcgyver]

You can have 2 parts that have a certain fit,
and neither one has a size that relates to any
measurement system. They just have to relate
to each other. Are you defining size as having units?

Of course size is expressed in units. In the machine shop, we don't have to parts and a fit, we make two parts to a fit. And to do so, their size matters.

I think what you are trying to say is nominal size does not matter, obviously, for a one of. You have to know the size of the bore say (ergo its size matters) and be able to measure the size of the mate (ergo its size matters) and machine it to size to achieve the fit. What doesn't matter (for a one of, if you don't have to work to a drawing) is whether one is 3.000, i.e. could be 2.995 or 3.004 etc, but whatever it's size is, it certainly matters else you'd have no change of turning the shaft to the right fit.

You can make a piston FIT a cylinder to function in a specific application
without using units. Just use ID and OD spring divider calipers.

Correct, I've mentioned here accounts of model engineers in the past, who lacking a micrometer, did just that. Try to do so for an interference fit, or .0002 clearance on a tailstock quill and bore. That example only works because the acceptable accuracy is doable with that technique. Close to a fluke. Were you required to achieve a specific fit, which of course changes with diameter, it would not be accurate enough, ergo the size of the parts would matter

[J Tiers]

You ALWAYS have a specific size to hit making a part to fit another part. An "absolute" size. And every one of them relates to any measurement system that has ever existed. You could measure in cubits if you like. (Decimal cubits?)

The units might only be "just enoughs".... general "feel" units of the resistance of the calipers moved over the part, and you aim for one "just enough". Or you make to the part itself, and it must be "one diameter" in size, But there is some measurement means.

So what? You still get it done. Not sure what the heck Doozer is driving at.

You need to agree on definite units and the means to measure them only if you need to make a part without reference to any other part. If you are making something to fit a part on your brother's tractor 500 miles away, you have to make it to measurement. If he lives next door, you may be able to borrow the matching part and work to it, and not a measurement.