Run out query
I have a Myford Super Seven and a Myford ML10, just made a collet holder for ER collets, really made it for the ML10, both lathes have identical nose thread and registar, I have carefully checked the registar on both. I bored the taper for the collet on the ML10 whilst it was screwed on to the mandrel so on checking obviously zero run out, thought lets see how it is on the Super seven and to my suprise 4 thou total run out of the tapered bore, how can this be ?
Both mandrels are running absolutely true (I checked them) the register in the holder I made is spot on, I'm baffled. David
Is the run-out axis parallel to the spindle axis or angled. I'm not sure what the design looks like but have you tried rotating components one by one and checking how the run-out moves.
Last edited by philbur; 10-12-2010 at 01:21 PM.
dust, grit, swarf, bent spindle.
Mandrel ??? I assume that's the term used for what we call "spindle" on this side of the pond?
Originally Posted by David S Newman
Have you rechecked after removing/reinstalling multiple times on the ML10? ...always get a dead true indicator reading each time?
Likewise, do you consistently get the 4 thou TIR on the other lathe?
If so I'd recheck the registers on each, at different radii from the center.
...then I'd do the same as you, scratch my head, roll my eyes, curse the fates, etc. etc.
I'd think there must be some sort of cumulative error combinations creeping into the mix, but I have no more brilliant thoughts to offer. I'd first suspect the culprits RB211 cited, and then go from there.
The "Register" you refer to has nothing to do with the accuracy of the spindle/mandrel, or chuck. ( The shoulder behind the mandrel threads are sometimes called the register.)
I know this is contrary to what most believe, but it is the truth
A chuck is controlled on the vertical and horizontal axises by the threads
The third axis (L &R) is controlled by the face of the spindle/mandrel
[ the face, is the perpendicular surface behind the threads, and adjacent to the shouder (next to headstock)]
You need 3 axises of control (6 directions) in order to stabilize any component.
Trying to match the "register" to the threads produces nothing but stuck chucks if you have not faced the part with the threading operation
The reason for the runout, is that you did not cut the threads AND the face at the exact same setup.
This introduces error in the chuck and is angular in nature
Fixing a chuck or face plate after having this manufacturing error is very hard to do.
Last edited by Rich Carlstedt; 10-12-2010 at 06:36 PM.
Is this true for all threaded machine spindles or just Myford.
Originally Posted by Rich Carlstedt
Yes Phil, all spindles.
My Boxford has a 1 1/2"-8 Spindle thread.
I have two Pratt Bernard chucks and a Original 6 Jaw Buck Chuck.
These are some of the most accurate chucks made.
The Pratts came with the Boxford and have about .018 (.5mm)clearance from the shoulder diameter. The Buck came with a .025 (.6mm) clearance. yet each have tremendous repeatability when removed and then replaced on the spindle. my P&B is less than .001 and the buck is dead nuts on.
To better understand the function of helical threads, take a thread that has a loose nut on it, and push it radially, and note that the nut is not restricted totouching the crest or root and can move. Now add some force with your fingers such as seen when you load the nut in a normal application. As you apply a axial load, the nut will automatically center itself on the thread, and no longer allow radial shifting. You may note that it even resists twisting as well. You may recall that when ever you strip a bolt from overload, it is always a complete concentric failure. You never strip a bolt or screw on one side ?
The helix angle and V thread is extremely effective in centering the nut, and your chuck threads....accurately, every time (no dirt or chips please).
Now as the spindle loads the chuck, the rear face of the chuck meets the face of the spindle. If the chuck face was turned at the same time as the threads were cut, The face will be perpendicular to the load and be in harmony. If the face was cut during another setup, it will not be true as it will touch in one point first,and it then will attemp to cock the chuck, resulting is error and maybe a stuck chuck. If the shoulder clearance ( ie register) is tight, it will jamb the chuck before the "faces" are in total contact
Last edited by Rich Carlstedt; 10-14-2010 at 12:00 AM.
Hi Rich, then I’m confused as to why DIN 800 has a specific tolerance requirement for the "register" but not for the thread.
Also the test protocol for my lathe takes great interest in the run-out of the "register" but no interest in the thread. Its clear when I screw the chuck on that the “register” is forcing the chuck to align with the spindle.
What am I missing?
I don't agree. What is the purpose of the register then, if it isn't to centre the backplate? My school installed 2 Boxfords new in 1964 (My only ever exposure to new lathes) and on both of them you could feel an increase in resistance when the register started to enter the chuck backplate, so there obviously wasn't significant clearance between the two as supplied by the manufacturer. I have always been taught that the thread is there to hold the chuck in place, the register and the flange face behind the register are there to align the chuck. My Raglan lathe has a square thread nose, which can't be self aligning as described above, and 2 registers, one in the normal position and onother smaller one at the front of the screw. The handbook says 'The spindle nose has two bearing positions for accurately locating backplates, faceplates etc, the function of the square thread being merely for screwing these attachments on and off'
"the function of the square thread being merely for screwing these attachments on and off" Square thread? I thought the previous poster only mentioned V threads. Of course square threads won't locate anything.