[How-To] Theoretical question about lathe chucks

[Bi11Hudson]

I was watching a video on https://www.homemadetools.net/ regarding a "loose" Logan. The size and the brand have no real bearing on the question. I was taught many years ago by a ("lifer" military) machinist. I didn't have the craft knowledge to learn all that much from him. But what I did learn has stayed with me for 50 years. From his teaching, there is a shoulder at the end of the spindle. This shoulder is the registration surface for the chuck. The threads are merely to hold the chuck against this surface. They have little bearing on alignment. My tutor emphasized the point about never allowing chips or dirt when I was mounting a chuck, for this reason.

I have seen examples of users having a gap at the back of the chuck. It seems to me to be a "too shallow" chuck mounted on extended threads. That is the most likely answer, to me, there may be another. The chuck bottoming out on the threads before closing with the shoulder. The problem was usually presented as "loose bearings" in a friction bearing machine or a misaligned chuck. In this video, the operator had a video camera pointed at the back plate where I saw movement. Thus the question. . .

I have swapped chucks around between an Atlas (Craftsman) and a Taiwanese 9x20 having the same (1-1/2-8) thread on the spindle. The Chinese machine (Grizzly 1550) has a longer spindle than the Atlas. It was necessary to make a "shim" plate, ~1/2", to make up the difference. It seems to have worked properly, both chucks run reasonably true(as in < 0.002) on both lathes.

The question is: Is this a valid perspective or was I lucky when I made the "shim"? All responses are welcome, the older professionals more so.

Bill Hudson​

 

[Ulma Doctor]

just old not a professional....

you are correct in your thinking.
the chuck registers on the flat of the spindle, not the threads.
the shim made it possible to indirectly register the chuck against the spindle

 

[Lo-Fi]

A mate of mine bent the threads on his spindle by running it with a chuck that the previous owner had mounted that didn't register on that shoulder you're talking about. Managed to save it thankfully. Was his first time with a lathe and he didn't know. The clamping force between the threads and the shoulder are where any rigidity in that setup comes from.

A shim seems entirely the right way to go in your case.

 

[JRaut]

To add just a little:

On a typical threaded nose spindle, the chuck with a properly dimensioned back plate will register on two faces -- (1) the back of the chuck will register on the flat of the spindle nose, and (2) the unthreaded portion of the bore in the back plate will be a tight fit to register on the boss of the spindle nose. Both of those registering surfaces are required to ensure repeatable accuracy. **​

​

On a typical taper nose spindle, the chuck needs only to register on that tapered face to sufficiently locate it axially and radially.​

** I currently have a very nice 6-jaw chuck that I need to make a new back plate for. The bore is oversized about 50 thou with respect to the boss on my spindle nose, and as a result, I get very poor repeatability when mounting/re-mounting my chuck. As you suggest, the threads alone are insufficient to ensure repeatability.