Question about machining shafts and holes

[swirnoff]

As a beginner, I have a chicken-and-egg type question.

In general, if I have a shaft and matching hole to machine, is it better practice to make the hole first and machine the shaft to fit, or the other way around? The shaft will be a loose press fit (I think that's the right term...or maybe interference fit).

This is for a hand operated device, so nothing in it spinning at blinding speeds which might cause a catastrophic failure. In case that matters.

 

[Firstram]

I find that it’s easier to fit the shaft to the hole. What equipment are you using?

 

[markba633csi]

Unless you have a set of machine reamers which produce precise holes or you can bore the hole to a precise size, then it's usually
easier to machine the shaft to fit the hole

 

[francist]

I also find fitting the shaft to the hole is easier (you can always run a file over a shaft to take that last little bit but that’s kind of hard to do in a hole), but I’d also consider how you’re able to check for fit. If both pieces are small and easily handled, great. But if one is awkward or difficult to manipulate it may influence which operation you do first.

 

[swirnoff]

Thank you all. I have a Sherline benchtop mill and a 1939 Atlas 6" lathe. It'll be exciting to figure out how to make the 2 machines play well together.

 

[RJSakowski]

Considering that you are making an interference fit, you won't be able to use one part to test the other. I would make the hole first and a test plug that is slightly smaller than the shaft diameter. Make the hole to fit the test plug as a tight slip fit and then turn the shaft to the required slightly larger diameter. You are talking about a thousandth difference or less.

When making tight tolerance mating parts, I find it useful to turn a test plug with a range of diameters varying by a half thou or so which allows me to sneak up on the final dimension without overshooting.

 

[WobblyHand]

If the hole is blind, consider drilling a weep hole at the bottom to allow air to escape. Or score along the length of the shaft to allow the air to escape. The score depth obviously needs to be deeper than the interference.

I learned the hard way that this was necessary to get all of my pins to have uniform stick out. I pressed them all in using a turned spacer, but a day later, some of them moved outwards. This is because some of them had fully trapped air and had compressed it to very high pressures. Over a single day, the pins had random stick out. Since this was a 6 pin pin spanner, I had to start over again. Here's the one I made that worked out well. 11mm pin circle. Body and pins are 12L14. Pins are 2mm diameter x 8 mm long. Body is 3/4" (19.05mm) x ~1.25" (32mm). Pins stick out 3.5mm. The spacer I used is shown below. Pressed 1 pin at a time using an arbor press. The spacer held the pin in place and acted as the stop for pressing. Drilled the interference fit holes with quality 1.95mm drill. Spacer drilled with a "bad" 1.95mm drill that drills over 2.05mm! Oh yeah, I bought 10 of those lousy drills from "Drill America" and they were off-size or slightly bent. I had made a test piece and was astonished that the pins just fell in - rather than being an interference fit. So much for junk drills. Anyways, it was a fun little pin spanner to make.

Word to the wise - measure everything you buy - don't trust your suppliers, especially if on eBay. Measure your stuff with a micrometer as soon as you receive it. Process a return for non-conformant product if off size. That being said, I have bought some very good stuff there, and a few clunkers. Overall, the eBay experience has been mostly favorable.

 

[markba633csi]

You need an accurate micrometer if you are going to do this regularly- something that will read to ten-thousandths of an inch = 0.0001
And it is all too easy to undershoot the diameter when you are sneaking up on a dimension- BTDT
Don't be surprised if your Atlas cuts taper- mine does, badly. To cut short, accurate diameters I don't use the carriage.
Instead, I carefully adjust the compound at exactly 90 degrees and use that to cut longitudinally with the carriage locked.
-M

 

[Jake M]

As a beginner, I have a chicken-and-egg type question.

It's a very valid question.

In general, if I have a shaft and matching hole to machine, is it better practice to make the hole first and machine the shaft to fit, or the other way around?

The best way, for interchangeability, conformity, and just plain being proper, is to make the shaft to the nominal dimension, and the hole would be over or undersized to get the proper oil clearance, free fit, snug fit, press fit, whatever you're after.

In practice, if none of the above is going to be an issue, it's easier (more certainty, less chance of scrapping a part) to do just the opposite. Make the hole to size, and then "fit" the shaft to it. By doing it that way, the hole can be "gauged" by a simple plug (of about any material), which is far more approachable than having the tooling to make a polished bore and the metrology to nail that down to tenths. Doing the shaft at that point becomes a simple addition of outside measurements. (Whatever the press fit test plug comes out to be, plus the desired interferance).

The shaft will be a loose press fit (I think that's the right term...or maybe interference fit).

Interference fit is the proper name. Press fit is the common name. Same/Same.

This is for a hand operated device, so nothing in it spinning at blinding speeds which might cause a catastrophic failure. In case that matters.

If the press fit is solid, it really doesn't care (within reason) about whether the actual finished diameter of the finished inside/outside dimension (which will be the same at that point), as a few thousandths above or below nominal won't be anything that (in the home shop) could ever be worked out and counted on, as it'll be lost in the noise with all the other inefficiencies or home shop production. So if it's 3/4 inch nominal for example, it doesn't care if the finished "joint" comes out with a 0.748 diameter or a 0.752 diameter, as the material probably wasn't properly pedigreed anyhow, so you can't bank on it's material properties to enough decimal points to have that make a difference. If it was a 1/4 inch nominal assembly, and the hole came out undersized at 0.200 instead of 0.250, and you undersized the shaft to fit that hole.... Now you're talking about a problem. But so long as you keep the percentage of deviation small, zero mechanical worries and so long as it's a one and done part, with no interchangeable parts involved, there's no social or moral issues with doing it "backwards". It's much more efficient (in a home shop) to do it that way.