- Joined
- Feb 5, 2015
- Messages
- 662
Lathe Dead Centers - How Rigid Are They ?
(A discussion on another forum a few years ago produced varying opinions about how much rigidity a tailstock center adds to a long workpiece chucked in the headstock. The varying opinions resulted from different interpretations of beam deflection, specifically how the tailstock center should be represented when calculating deflection. I made the following brief experiment and posted the results.)
This may be an odd thing to provoke disagreement - if you need a center then use a center (if it is possible to do so). But since this is a fairly easy situation to set up and it was a slow Saturday morning, I decided to measure workpiece deflection under a couple of different conditions. The test setup was in a small Emco "Compact 8" lathe, configured as follows:
Here are some of the test details:
Here’s the little 4-jaw clamped to the 3/8 square lever.
The “control” configuration is with the test bar clamped in the headstock chuck and in a 5/8 tailstock chuck with pressure applied from the tailstock to make sure that the tapers are locked. With the load applied to this configuration, the static deflection is .008 inches. (Note that the small home-built steady rest in front of the tailstock chuck isn’t touching the test bar.)
As a sanity check, the calculated deflection of the “control” configuration, assuming the length of the test bar is 16 inches, instead of 13-1/2 inches, is .009. (The length is the distance from headstock chuck jaws to front of tailstock - that is, the length of the tailstock chuck is added to the shaft length.) The error is around 11%, which would be unacceptable in a laboratory but is OK in my shop, LOL.
I tried two dead center configurations, 1/8 center diameter and ¼ center diameter. I didn’t observe any difference in test bar deflection between the two different center diameters. Neither did the deflection seem to be dependant on the axial load applied through the dead center - in other words, tightening the tailstock handwheel past "snug" didn't appreciably influence deflection. (As above, the steady rest bearing surfaces are not touching the test bar.) The deflection of this configuration is about .010
The third configuration employed the steady rest alone to support the end of the test bar. The steady rest bearing surfaces were adjusted to a snug fit while the test bar was centered, then the center was removed. The deflection in this configuration was .012/.013 inches which surprised me because I didn’t expect the steady rest to be this rigid.
The significance here is that the dead center is ALMOST as rigid as a fixed end support like the baseline configuration. And the steady rest isn't too bad either.
What have we learned ?
Although I didn't perform deflection measurements with a live center, I would expect one that is not well-worn to perform about the same as the steady-rest, if not slightly better (for high-quality units).
Cheers, hope this has been of interest -
(A discussion on another forum a few years ago produced varying opinions about how much rigidity a tailstock center adds to a long workpiece chucked in the headstock. The varying opinions resulted from different interpretations of beam deflection, specifically how the tailstock center should be represented when calculating deflection. I made the following brief experiment and posted the results.)
This may be an odd thing to provoke disagreement - if you need a center then use a center (if it is possible to do so). But since this is a fairly easy situation to set up and it was a slow Saturday morning, I decided to measure workpiece deflection under a couple of different conditions. The test setup was in a small Emco "Compact 8" lathe, configured as follows:
Here are some of the test details:
- Test bar is 7/16 dia ground steel shafting, 13-1/2 inches long from the front edge of headstock chuck jaws to the end of the test bar.
- Load is applied to the middle of and UNDERNEATH the test bar, a DTI is positioned on top of the bar to measure deflection. (A small rectangular aluminum block is secured to the center of the test bar to insure that the indicator point is bearing on a flat surface rather than on the slippery, uncertain diameter of the test bar, see photos.)
- A weight of 5.4 lbs (a small 4-jaw chuck) is clamped to a 3/8 square steel bar that is 27 inches long (from the center of the test bar to the approx center of gravity of the weight). (see photos)
- The 3/8 x 27 inch square bar is supported, 5 inches from the centerline of the test bar, by a 3/8 toolpost. The 3/8 bar is free to move vertically but not horizontally.
- The load is about 119 lb/inch applied at the end of a lever 22 inches from the toolpost fulcrum. The test bar centerline is located 5 inches on the other side of the fulcrum. (The 119 lb/inch accounts for the weight of the 3/8 lever acting with and against the weight of the chuck.) The estimated result is a load applied to the centerline of the test bar of 24 lbs
Here’s the little 4-jaw clamped to the 3/8 square lever.
The “control” configuration is with the test bar clamped in the headstock chuck and in a 5/8 tailstock chuck with pressure applied from the tailstock to make sure that the tapers are locked. With the load applied to this configuration, the static deflection is .008 inches. (Note that the small home-built steady rest in front of the tailstock chuck isn’t touching the test bar.)
As a sanity check, the calculated deflection of the “control” configuration, assuming the length of the test bar is 16 inches, instead of 13-1/2 inches, is .009. (The length is the distance from headstock chuck jaws to front of tailstock - that is, the length of the tailstock chuck is added to the shaft length.) The error is around 11%, which would be unacceptable in a laboratory but is OK in my shop, LOL.
I tried two dead center configurations, 1/8 center diameter and ¼ center diameter. I didn’t observe any difference in test bar deflection between the two different center diameters. Neither did the deflection seem to be dependant on the axial load applied through the dead center - in other words, tightening the tailstock handwheel past "snug" didn't appreciably influence deflection. (As above, the steady rest bearing surfaces are not touching the test bar.) The deflection of this configuration is about .010
The third configuration employed the steady rest alone to support the end of the test bar. The steady rest bearing surfaces were adjusted to a snug fit while the test bar was centered, then the center was removed. The deflection in this configuration was .012/.013 inches which surprised me because I didn’t expect the steady rest to be this rigid.
The significance here is that the dead center is ALMOST as rigid as a fixed end support like the baseline configuration. And the steady rest isn't too bad either.
What have we learned ?
- a dead center is about 36 times more rigid than a chucked but unsupported bar
- a steady rest is about 30 times more rigid than a chucked but unsupported bar
- tailstock pressure didn't make a measurable difference in deflection (.0005 indicator) once the hand-wheel was snugged; it follows that over-tightening produces no benefit but CAN cause problems (lubrication, heat, expansion)
- center drill variation from 1/8 to 1/4 diameter made no difference in deflection, at least for the small shaft on which the measurements were based - may as well use a smaller, less obtrusive center-drill (unless the tip of your dead center is damaged)
- static measurements may not represent what actually occurs under cutting conditions, there are many contributions unaccounted for in the previous thirty-minute exercise
Although I didn't perform deflection measurements with a live center, I would expect one that is not well-worn to perform about the same as the steady-rest, if not slightly better (for high-quality units).
Cheers, hope this has been of interest -