Question On How To Make An Mt3 Socket?

I cut this this evening and it seemed to work well.
I did have the cutter just a hair below center though -- it seemed to generate a socket that has decent contact over most of the taper (and is much more inline with the center line of the part). In fact, it seemed to work so well that I'm thinking the reamers really weren't necessary for this...

I am, however, having a tough time envisioning how having the cutter off center messes up the taper?
It would seem that 1) every 'cross-section' of the taper must be a circle, 2) the radius of each cross-section should reduce by tan(compound-angle)*distance-compound-advanced?
The cutter is moving in a straight (and horizontal) line...

I'm just not seeing it... any clarification appreciated...
 
Ah - I think it just clicked...
If the cutting edge was 90 degrees off, pointing straight up, the center of the (circular) cross sections would be moving towards the operator at the same angle as the compound.
So effectively the resulting removed shape would be a cone with each 'section' slightly shifted (so like a leaning cone - sort of...)
Ok, so then, since I had the cutter slightly off center I would think the resulting taper (even if it seems to hold a tool well enough) would still be off-axis from the centerline of the workpiece - yes?

I am still seeing a .007 'runout' at the tool tip as the workpiece is rotated, which I wasn't able to figure out. I'll bet it's the socket being off-axis by just that little bit...

Sounds like I need to rechuck the part and try this once again with the cutting tool at dead center.

Mucho thanks to Dave for pointing that out!!!
 
The transverse axis (the line connecting the foci) is perpendicular to the lathe axis. The inside of his taper would look like half of an hour glass.

[Edit] The term I was looking for is "hyperboloid of revolution of one sheet". Follow the link below and run the animation to see how a straight line can generate it.

https://en.wikipedia.org/wiki/Hyperboloid
 
Last edited:
That's the one! Plane sections parallel to the axis should be bounded by hyperbolae then.
The animation on that Wikipedia page is a pretty clear illustration of how a straight toolpath not intersecting the lathe axis would produce one, too.

Thanks!
Dave H. (the other one)
 
You guys are blowing my mind!!!

Ok, it's not a "leaning cone". The wikipedia page helped a lot, but still had to noodle on it for a while today...

I *think* I'm starting to "see" how the geometry falls out... since the tool is slightly below the center line, as it moves the change in the diameter of the cross-section does not vary linearly (based solely on the angle of the compound), instead there is an additional "error" component (the distance the cutter is off-center) that, while constant, is being overlaid on the varying circumference of the cross-sectional circle (I envision this as imparting the "twist" to the bottom section of the animation). I'm sure that my description isn't of much use to anyone other than myself, but that's fine...

As I see it, there are three major take-aways from this:
1) The cutter needs to be on dead-center
2) I'm definitely going to try this with a piece of scrap, cutter on the outside of the part, a larger compound angle and a significant height off-center to see if I can get a cool shape!
3) I owe both you guys a beer for pointing this out! I wouldn't have even considered how whacky this geometry falls out if you hadn't pointed it out.

Thanks!
 
One more thing to do, when you indicate in the compound with the sample taper. Indicate the compound in without the tailstock being used. Also make sure the contact point on the indicator is on vertical center of the part as close as you can get it.
 
4gsr said:
Indicate the compound in without the tailstock being used.
Click to expand...

:dunno: Why would that be?
I used the tailstock to ensure that the I had the center I was using well centered on the lathe axis. I could turn the spindle and get minimal movement of the indicator needle at any point along the length of the taper. I get having the indicator as close to center height as possible. Not using the tailstock when setting up has me scratching my head...
 
kizmit99 said:
:dunno: Why would that be?
I used the tailstock to ensure that the I had the center I was using well centered on the lathe axis. I could turn the spindle and get minimal movement of the indicator needle at any point along the length of the taper. I get having the indicator as close to center height as possible. Not using the tailstock when setting up has me scratching my head...
Click to expand...
Lack of runout doesn't tell you that the taper is centered on the lathe axis. It just tells you that you placed your tailstock center on the axis of the taper. Not surprising since that center hole was probably used to grind the taper. Think about what happens when you offset the tailstock to cut a taper.

Set up again, without the tailstock. Using your DTI rotate the taper until the point of maximum runout at the far end is on the top. Then indicate your angle off the side.
 
I second John's advice, with a caveat: if the runout's more than a couple o' thou", either try to eliminate it to the point where you can't measure it or use an "elephant's foot" on the DI - these are flat pads, usually 1/4 - 1/2" diameter, that screw into the plunger in place of the usual ball end, and allow for a slight variation in height of the contact point where you measure.

Tracing the taper from a line that isn't parallel with the axis of the cone will also result in an error, now THAT is a parabolic section! I remembered Geometry!

If you have a parallel test bar, turned between centres, you could centre that up and test the tailstock centre's on-axis easily enough, run the DTI / DI from end to end side and top and make any adjustments (this assumes the tailstock ways are parallel to the spindle axis and carriage ways though... it can feel like you're chasing your tail sometimes, getting a lathe correctly aligned!)

Dave H. (the other one)
 
I'm sure I'm probably way of base on this, but wouldn't it make sense to start with an oversize o.d. and then once you have created and lapped in the socket remount the piece on a known good male taper and correct any run-out/concentric issues at that point? Mike
 
You must log in or register to reply here.
Back
Top