Comparison of Mounting Methods for Turning Slender Spindles

[Dennis J Gooding]

Comparison of Mounting Methods for Turning Slender Spindles



If one attempts to turn a spindle down to a smaller and smaller diameter, eventually one reaches the point where chatter marks begin to develop and further turning leads to increased chatter and possibly whipping of the spindle. The minimum diameter that can be achieved before this happens depends on several factors including the method used to mount the spindle. There have been posts from time to time comparing various mounting methods, but I have not seen any that give quantitative results. As an engineer (long retired), this intrigued me to the point that I decided to see if I could shed some light on the subject. I carried out a mathematical analysis of the problem based on standard engineering equations for beam deflection and followed it up with controlled experiments to try to confirm the calculations. An article, giving full details of the project can be found in the “ How-tos, Tips and Techniques” forum, but the main results are summarized below.

I compared five mounting methods, which are listed in order of increasing performance:

1. One end in chuck, the other free. 100%

2. Between centers—the most common method, and the standard of comparison here. 50%

3. One end in chuck, the other supported by a center. 59%

4. Both ends in chucks. 50%

5. Both ends in chucks and the spindle under tension. 50%

In each case, it was assumed that no other support was provided. (No steady-rest nor steadying hand wrapped around the spindle.) All results are based on a simple cylindrical spindle and will not apply exactly to arbitrary spindle profiles, although the ranking order will remain as shown. Mathematical analysis of Method 5 is impractical so only experimental results are presented and they are limited to one particular case. For each mounting method, there is a critical point along the spindle where deflection is greatest for a given tool force. The locations of these points are noted in the list above as a percent of the distance from the drive end to the outer end of the spindle.

The premise underlying the analysis was that for a given type of wood, spindle length and cutting tool and possibly other factors, chatter begins when the deflection of the spindle caused by tool pressure reaches a certain amount, independent of the mounting method used. On this assumption, mounting methods that yield higher stiffness will perform better than ones that yield lower stiffness. Furthermore, with this assumption, one can use well-known engineering beam deflection equations to calculate the relative minimum diameters that can be achieved for a given spindle length by the various mounting methods and, conversely, the relative lengths of the spindles that can be turned to a common diameter. The experiments were carried out to test this premise and, in fact, the test results seem agree quite well with the calculated ones as shown in the following figures. Note that these are relative results, with Method 2 being the standard.

Comparison of Relative Diameters Achievable


Comparison of Relative Lengths Achievable

 

[Dennis J Gooding]

Oops, I forgot to include the following picture in the above post.

 

[john lucas]

that's intriguing but so much is dependant on tool sharpness, tool handling skills (lack of pressure on the bevel) and in most cases using your fingers as a steady. I remember working for years trying to reduce chatter using all sorts of methods. Even tried putting spindle under tension using a chuck in the tailstock however start up inertia simply ripped the spindle apart since I didn't have the tension device geared to the headstock. ONe day I just started on a spindle and apparently had developed the "touch" and have been able to do it pretty well ever since.
ONe method you missed was taught by John K Jordan at a demo at our club last month. Using a short morse taper on the headstock end of the spindle instead of a chuck. It has a couple of advantages. One is if you remove the spindle it goes back in accurately. John cuts his morse tapers with a little relief in the middle so that only both ends are touching the inside of the spindle. This makes accuracy less critical and when you drive it in the spindle the compression of the wood also helps make the taper fit better.

 

[robo hippy]

Two food for thought points here....

One local turner, Rudy of Wooden Apple made the comment when demonstrating spindle turning about a 10 to 1 ratio for spindles to prevent chattering/flexing when turning, so 1 inch diameter you can go 10 inches long with pretty much no vibration problems.

The other was from watching Ashley Harwood when she was in Salem this summer, turning her long dainty finials for her sea urchin ornaments. She was going 6 or so inches long, and the tips were down to less than 1/8 inch diameter. She was using her 40/40 gouge, and didn't use a finger on the back side to support it when turning after it was roughed out. This, to me is the apex of 'the bevel should rub the wood, but the wood should not know it!' I told her that I needed to go back to the shop and practice my 'dainty' skills.... I did figure out a long time ago that when you use your finger as a steady rest, tool pressure = finger steady rest pressure, so it make sense if there is no bevel rub pressure, you don't need any finger pressure...

robo hippy

 

[Bill Boehme]

Thanks for taking the time for your analysis and posting the tutorial.

I have considered methods 4 and 5, but had some concerns that misalignment would actually introduce whipping for very long thin spindles. Also, it looks like the inertial mass at the tailstock end should be considered in a dynamic analysts.... but that is probably taking analysts way too far.

Woodturners have some tricks to help deal with chatter such as using supporting wood to keep the drive end stiff.

I am wondering how much torsional whipping in long thin spindles contributes to chatter. Even with string steadies and finger support eventually vibration will win.

 

[Dennis J Gooding]

Bill, I purposely tried to leave skill out of the study but instead emphasized taking care to try to get repeatable results that reflected the physics (mechanics?) of the problem. All experiments used a skew that was regularly ground and honed and the speed was fixed at highest speed of my Oneway 2435, about 3200 rpm. No hand or other special support was used. I regard the results as a starting point for applying skills such as hand support, alternative tools, speed variation, etc. I believe that in most cases skilled turners who repeat my experiments but use their off-hand as a steady rest and choose optimum tool and speed for each of the five basic mounting methods will end up the same ranking of mounting methods that I did, but with slimmer or longer spindles for all methods.

Regarding misalignment problems with Methods 4 and 5, I don’t see a scenario where it would be worse than with Methods 1, 2 or 3. Normally, the tenons would be turned between centers , so if there is no significant compression to cause an initial bow, the tenons will be co-linear. Of course wood stress release could lead to bowing during the turning process, but its magnitude would be least for Method 5 and greatest for Method 1.

Regarding torsional vibrations, yes I expect that they occur and may be the reason why the chatter marks tend to be spiral in nature.

 

[Bill Boehme]

Regarding misalignment problems with Methods 4 and 5, I don’t see a scenario where it would be worse than with Methods 1, 2 or 3. Normally, the tenons would be turned between centers , so if there is no significant compression to cause an initial bow, the tenons will be co-linear. Of course wood stress release could lead to bowing during the turning process, but its magnitude would be least for Method 5 and greatest for Method 1.

My thoughts really weren't about real world applications, but more along the lines of carrying things to the extreme such as the long thin trembleurs that some turners do as a challenge to their skills. But, as you stated, you didn't want to muddy things with unquantifiable things like skill level. I agree that bowing due to compression between centers is definitely more of a problem than holding the ends locked in a chuck.

Even pen turning on a mandrel can cause bowing problems. I'm wondering if holding the wood in a chuck (or a wooden drive center as John Lucas mentioned) at the headstock and having a live center something like the one used by pen turners that applies no axial force at the tailstock would be a good compromise between methods 3 and 4. It would mostly eliminate John's concern about the inertia of the tailstock chuck creating new problems.

 

[Mark Hepburn]

Dennis, I really appreciate your post. I find that, because I like to turn a variety of wood for my finials, and my skill level is such that I experienced a great deal of chatter sometimes. Also, you are dead on about compression and the resultant bowing it can cause.

I find this especially true with some of the softer materials or Greenwood that I'm turning for the finials I like to bend

The idea of holding it under tension really appeal to me, and makes the most sense to this non-– engineer.

Putting a tenant at either end, chucking it in a live center of some kind with the pin to retain it seems like that would really work.

Would not necessarily have to be focused on production, but just a standardized tendon that could be parted off.

Or did I misunderstand something here?

 

[Doug Rasmussen]

In my business I designed and we built a couple machines to turn pool cue shafts for a professional pool player. A typical shaft might be 30" long with a concave taper from around 7/8" down to1/2" at the tip. The machines had to be capable of going from a square blank down to almost finished diameter in one pass.

The basic machines were based on a Unisaw type table saw. There was a traveling carriage assembly with the equivalent of a light duty lathe that held the blanks between centers. The carriage traveled front to back of the saw (parallel to the saw's blade). The "lathe" was turned 90 degrees (as if you tipped a lathe over away from you). The blank rotated slowly under power centered over the saw blade. The lathe assembly pivoted on the carriage up/down into the saw blade with a follower riding on a pattern of the shaft. There were two DC variable speed motors, one to move the carriage, the other to rotate the shafts.

There were no steady or follower rests to dampen chatter. The 10" saw blade with it's large radius of contact cut smoothly and seemed to provide a steadying of the shaft during cutting. That was the one thing I wasn't sure of in the design.

No pictures available. People who make or aspire to make pool cues are very secretive, the last thing they want is for others to know how they make their cues.. I don't play pool so this was my first real contact with a high stakes pool player (gambler). It was an interesting experience, not my usual type customer. Everything was cash, when he gave me the go ahead after talking for a couple weeks he came over with a large wad of 100 dollar bills and started peeling them off, stacking them on my desk telling me to say when to stop.

 

[robo hippy]

I had a neighbor who made pool cues. He used a very slow speed lathe and a router on a template. Not sure which type of bit, but probably spiral type...

robo hippy

 

[Dennis J Gooding]

I just want to add a couple of points. First, regarding the inertial impact adding a "chuck" onto the live center for Methods 4 and 5, I would point out that the two examples shown in the figure above each weigh only 1.2 oz, and both could have been lightened by turning away excess material. Given their small diameters (1-inch max) and mass they would have a very small moment of inertia. Not having disassembled my Oneway live center, I can only speculate that it has a considerably larger moment of inertia.

Second, if one is turning spindles that require tenons on both ends, Methods 4 and 5 have another advantage over Methods 2 and 3, because with Methods 2 and 3 the total unsupported length of the spindle will be increased by the lengths of the tenons that are not supported by chucks. For example, if you are turning a 20-inch long spindle that requires 1-inch tenons and use Method 2, you would actually have a total length of 22-inches to deal with. By my calculations, this would lead to a 7% increase in the minimum diameter that could be achieved. Therefore, the comparisons of the performance of the various methods given in the original post understate the benefits of the higher-numbered methods in situations where tenons are required on the finished turning.