Improving a lathe spindle head.

For some time I have been planning to change the single phase, single speed motor on my lathe for a larger 3 phase motor driven by a VFD (Variable Frequency Drive). The infinitely variable motor speed would allow me to eliminate the multiple pulley speed change system, but there is a problem when running a motor at low speed with a VFD. For a given spindle speed the spindle torque from a motor running at slow speed is less than achieving the same slow speed by mechanical gearing, for a given motor. A way around this problem is to use a vastly oversize motor.

This post is not about fitting the motor but is about some changes to the lathe head that the motor and drive system enabled me to do. My motivation was to make the head stiffer and less prone to vibration and chatter. Not that I had particular problems in that area but I find it hard to resist when I see a better way of doing something.

Although I show how I did this on my lathe, a 1240 JET, the ideas are certainly applicable to those lathes which do not have a closed in spindle head.

Attachment 30072 Attachment 30073 Click images for full size.

I have prepared a PDF file with all the reasoning and details at Inflicting GBH on a lathe

I plan to put together a video about this but in the meantime here is a link to a 2 minute video showing some stiffness measuring.
https://www.dropbox.com/s/7t7ho5x1qq...Lathe.mp4?dl=0

Warning: The pdf shows images of severe GBH to a lathe and those with extreme sensitivity to mechanical abuse are advised to look away now.

Thanks for yet another excellent and humorous writeup, Tony!

2 personal cents:
-Having the headstock trapezoid-shaped (rather than just box square) should (at least in my empirical theory) make it more stable,
and filling it with concrete will also add to that (apart of being over the top - but that goes without saying).

Q: -What about you workshops temp differences thru the year, combined with the "linear thermal expansion" of CI vs steel and concrete?
Any thoughts or hypotheses on that?

-Please keep up all your different projects and good spirits!

Cheers
Johan

Quote:

---

Originally Posted by DIYSwede

Thanks for yet another excellent and humorous writeup, Tony!

2 personal cents:
-Having the headstock trapezoid-shaped (rather than just box square) should (at least in my empirical theory) make it more stable,
and filling it with concrete will also add to that (apart of being over the top - but that goes without saying).

Q: -What about you workshops temp differences thru the year, combined with the "linear thermal expansion" of CI vs steel and concrete?
Any thoughts or hypotheses on that?

Cheers
Johan

---

thanks for the nice words.
I did think about thermal expansion differences. For the range involved I don't see much of a problem between the steel and CI. The concrete is an unknown quantity and my fingers are crossed as a preventative method. The internal shape will not allow significant movement even if the bond get broken but fretting could cause dust.

GBH indeed Tony but I like it. Could be the first ever lathe to suffer from concrete cancer (joking). Good move to eliminate the back gearing and I am with you on the merits of poly V belts. Thanks for another suburb write up.

Thanks for a quickie reply, Tony!
EngineeringToolbox re: Coefficients of Linear Thermal Expansion:

Cast Grey Iron: 10,8
Steel: 11- 12,5
Concrete (unreinforced): 13-14
Concrete Structure (assumed Rebar): 9.8

unit: micrometer/ meter x deg C diff

-How steel reinforced concrete can have a lower thermal expansion than any of its constituents has, was above my paygrade.
Then - checking into that* - aggregate/ filler material makes this difference - quartz/ gravel has the highest and granite the lowest LTE coeff.

Paper below measures differences in shrinkage in just-poured concrete, as well as LTE with different wetting methods applied,
this could perhaps be of interest?

*Shrinkage of Concrete and Development of Cracks | Concrete Technology

Cheers
Johan, "also a Reader - and I'm PROUD of it!"

<!-- BEGIN /var/www/html/homemadetools/protected/modules/zeus/views/tool/postUpdate.php -->
Thanks tonyfoale! We've added your Lathe Spindle Head Modification to our Lathes category,
as well as to your builder page: tonyfoale's Homemade Tools. Your receipt:

<div id="blocks"> <div class="block b1 pngfix"> <div class="bimg"> <div> <a href="http://www.homemadetools.net/homemade-lathe-spindle-head-modification"> <img src="/uploads/224458/homemade-lathe-spindle-head-modification.jpeg"/> </a> </div> </div> <div class="head pngfix"></div> <div class="left pngfix"></div> <div class="right pngfix"></div> <div class="blockover b1 pngfix"> <div class="title"> <a href="http://www.homemadetools.net/homemade-lathe-spindle-head-modification">Lathe Spindle Head Modification</a> <span> by <a href="http://www.homemadetools.net/builder/tonyfoale">tonyfoale</a></span> </div> <div class="tags">tags: <a href='http://www.homemadetools.net/tag/lathe'>lathe</a>, <a href='http://www.homemadetools.net/tag/motor'>motor</a> </div> </div> </div> </div>
<!-- END /var/www/html/homemadetools/protected/modules/zeus/views/tool/postUpdate.php -->

I think this is why many modern folks have gone to a granite-filled epoxy resin for this purpose...(and reading quickly I see that Tony has mentioned this in the PDF!)

Major modification to your lathe. I applaud your lathe improvement.
I have a 1976 Jet 10x24 lathe. Very similar head stock construction. I found when I had pulled the spindle many years ago (because of an oil leak), that the oil level sight glass, had a single hole going straight back to the area between the taper bearing cup, and the seal (annulus). If the level dropped below this hole, the sight glass indicated there was oil, but the bearing was dry. I modified my head stock by hand drilling a 1/8" hole from the bottom of the oil sight glass reservoir to the bottom of this bearing cup and seal annulus space (both front and rear bearings). I do see now when it runs that the bearing pulls oil in from the reservoir, and shows a low level, but returns when the spindle is stopped.
I too have installed a 3phase motor on my lathe, and a VFD, but still use the belts. I did change the standard v-belts with the segmented type, which greatly reduced vibration and noise.
This lathe was originally on a wood bench, and then a sheet metal stand. I could never do parting operations as it would go into a harmonic oscillation of the headstock and ways. I purchased a 2 inch slab of steel to bolt the lathe onto. I put that slab on my Wells-Index vertical mill, and faced the hot rolled slag off. Then drill and tapped the 6 mounting holes that mate with the mounting ears on the lathe way casting. This fixed the rigidity of the lathe to where I have no more issues with cutoff operations.

Your removal of the back gears puzzles me. They lock the spindle when trying to change the chuck or faceplate on a threaded spindle. They may not be needed for machining because of the VFD, however certain operations such as tapping would have much more torque with the back gears engaged with a higher motor speed.

Quote:

---

Originally Posted by metric_taper

I found when I had pulled the spindle many years ago (because of an oil leak), that the oil level sight glass, had a single hole going straight back to the area between the taper bearing cup, and the seal (annulus). If the level dropped below this hole, the sight glass indicated there was oil, but the bearing was dry. I modified my head stock by hand drilling a 1/8" hole from the bottom of the oil sight glass reservoir to the bottom of this bearing cup and seal annulus space (both front and rear bearings).

---

I was preparing a post on this very topic. What a crazy design feature that was, I wonder how many other lathe models have the same problem. The degree of incompetence that exists in some design offices never ceases to amaze me.

Quote:

---

Originally Posted by wa0dw

Your removal of the back gears puzzles me. They lock the spindle when trying to change the chuck or faceplate on a threaded spindle. They may not be needed for machining because of the VFD, however certain operations such as tapping would have much more torque with the back gears engaged with a higher motor speed.

---

No need to be puzzled, as I mentioned I almost never needed to use the back gears. I never use them to lock the spindle, my brutality to lathes does not extend to that. I made a tool to hold the spindle for chuck removal.
I do not need more torque for tapping, that could lead to broken taps. For that reason you can buy torque limiting tap drivers.

looks like an impressive mod.
left wondering why you didn't just attach the motor in such a way as to use the existing gear train.

Quote:

---

The degree of incompetence that exists in some design offices never ceases to amaze me.

---

me too, it starts at the top and filters down. Try explaining to someone you cant physically make something as there is no access for tooling into a sealed void. when the Boss wants a word with you about for your attitude and behaviour with the full intention of firing your arse.
I like to think managers are like politicians, promise the earth to get you to work for them then find reasons to discredit you and cut your wages with and end goal of getting you to work for nothing. Wont listen wont learn, its obvious why they like yes men, migrants and robots - people who wont answer back or make them look stupid.

Quote:

---

Originally Posted by tonyfoale

I was preparing a post on this very topic. What a crazy design feature that was, I wonder how many other lathe models have the same problem. The degree of incompetence that exists in some design offices never ceases to amaze me.

---

I purchased this lathe new, I was age 20, and was the first tool of this size ever purchased by me. I've learned that "JET" was an importer back in the 70's-80's and my lathe was shown manufactured in Taiwan. I don't know what they were thinking as they must have gotten lathes to copy. When it comes to the tail stock, cross, and compound feed screws, they are all Imperial (as well the main screw). The lathe they copied must have been metric, so they looked at an Imperial tape measure and saw 1/8 inch as the increment vs millimeter, so all four feed screws are 0.125 feed per turn. I can't think of any other reason for this odd screw pitch, I guess it could be worse if they used 7/64" pitch.

Quote:

---

Originally Posted by metric_taper

I purchased this lathe new, I was age 20, and was the first tool of this size ever purchased by me. I've learned that "JET" was an importer back in the 70's-80's and my lathe was shown manufactured in Taiwan. I don't know what they were thinking as they must have gotten lathes to copy. When it comes to the tail stock, cross, and compound feed screws, they are all Imperial (as well the main screw). The lathe they copied must have been metric, so they looked at an Imperial tape measure and saw 1/8 inch as the increment vs millimeter, so all four feed screws are 0.125 feed per turn. I can't think of any other reason for this odd screw pitch, I guess it could be worse if they used 7/64" pitch.

---

I was 15/16 when i bought my first lathe, a Myford Super7. I cannot believe that was 60 years ago.
Do not quote me on this but I think that JET and other Asian manufacturers can supply their lathes with either metric or imperial feed screws, depending on the country of destination. Like you my JET has imperial threads but I bought it 11 years ago in the US when I lived there for 3 years. It will not matter soon because I plan on fitting a motor driven ball lead screw. I may even go completely mad and do a full CNC, it would be easy enough.

Quote:

---

Originally Posted by MeJasonT

looks like an impressive mod.
left wondering why you didn't just attach the motor in such a way as to use the existing gear train..

---

I explained that I wanted a smoother drive, and the whole point of fitting the variable speed motor/VFD combination was to get away from the original drive train.
Also I would not have been able to make the stiffening mods if I kept the original.

Mr. Foale, amazing write-up! I was actually considering a similar modification to a Logan 1875 that I've been restoring, and this might be the impetus to acutally get it done. I've been "restoring" it for over two years now, and every time I walk by it, I'm reminded why the warden calls it my "wall-o-shame." Most of those kinds of projects started ending up at that wall in the shop, and it was the first big one to land there.

Are you able to do any vibration testing, such as with an accelerometer, or a device similar to vibration sensing devices used in the automotive industry to diagnose engine vibrations? If so, I'd really like to see data on how well your modification has performed.

Quote:

---

Originally Posted by tmoore4748

Are you able to do any vibration testing, such as with an accelerometer, or a device similar to vibration sensing devices used in the automotive industry to diagnose engine vibrations? If so, I'd really like to see data on how well your modification has performed.

---

I could easily put accelerometers in different places but it would not tell me very much because I have not done that with the lathe in its original format. I wish that I had now that you mentioned it.
My warden has no idea what happens in my workshop.

Quote:

---

Originally Posted by tonyfoale

I was 15/16 when i bought my first lathe, a Myford Super7. I cannot believe that was 60 years ago.
Do not quote me on this but I think that JET and other Asian manufacturers can supply their lathes with either metric or imperial feed screws, depending on the country of destination. Like you my JET has imperial threads but I bought it 11 years ago in the US when I lived there for 3 years. It will not matter soon because I plan on fitting a motor driven ball lead screw. I may even go completely mad and do a full CNC, it would be easy enough.

---

The more I think about issues with this lathe, None of the feed dials can be zeroed, as they are part of the hand crank thrust bearing tightening assembly. Thinking back, I've done many improvements on this "pre-assembled kit". Nothing like what you've just done.

The best add on was a DRO back 30 years ago, which cost almost as much as the original purchase price of $1800 in 1977. The price of DROs has come down to where like VFDs they are commodity items. As well making my own backplate and purchasing a new Bison brand 3 jaw chuck.

The spindle thread was a real bastard at 50mm diameter, 8TPI. I recently found 2 backplates on eBay that fit this thread. They were slightly too tight of a thread fit, so I got them centered in a 4 jaw, and retimed the thread so I could use the lathes power feed with an internal threading bar. I manually rotated this assembly to scrape the interference fit. So much for mass manufacturing with interchangeable parts.

Quote:

---

Originally Posted by metric_taper

The spindle thread was a real bastard at 50mm diameter, 8TPI. I recently found 2 backplates on eBay that fit this thread. They were slightly too tight of a thread fit, so I got them centered in a 4 jaw, and retimed the thread so I could use the lathes power feed with an internal threading bar. I manually rotated this assembly to scrape the interference fit. So much for mass manufacturing with interchangeable parts.

---

Mine came fitted with an imperial spindle thread but also with a metric thread spindle. I use the imperial thread because the chucks and face plate which came with it fit. To change to the metric one would entail new back plates etc for no benefit. Without going down to measure I do not recall the dimensions. I have an assembly drawing of the complete head and that shows two different part no.s for the spindle labelled metric and imperial. The previous owner bought the second spindle.
I am not a fan of screw on chucks, I much prefer a taper fitting like my old Colchester Student had.

Quote:

---

Originally Posted by tonyfoale

...My warden has no idea what happens in my workshop.

---

Good gawd, if only. She doesn't understand most of what I try to do, even though I explain. She just sees the pile of uncompleted projects and makes fun of me.

Quote:

---

Originally Posted by tonyfoale

I am not a fan of screw on chucks, I much prefer a taper fitting like my old Colchester Student had.

---

My other lathe has a D1-8 spindle nose. I like that I can power turn in both directions.

I had to build a bridge crane as the 12" chucks are now getting beyond my strength limit of back damage.

I also like keyless chucks, but they have this same issue.

I hope you document your CNC lathe conversion. I've thought of the same, I'm curious about methods of attaching a shaft encoder to the spindle. Keeping it out of the way for long stock though the spindle hole appears to need some sort of cog belt and timing gears. Then a quality encoder with ball bearing and such is needed, and there must be wear to worry about in all this.

Quote:

---

Originally Posted by metric_taper

The spindle thread was a real bastard at 50mm diameter, 8TPI. I recently found 2 backplates on eBay that fit this thread. They were slightly too tight of a thread fit, so I got them centered in a 4 jaw, and retimed the thread so I could use the lathes power feed with an internal threading bar. I manually rotated this assembly to scrape the interference fit. So much for mass manufacturing with interchangeable parts.

---

I just checked. My fitted imperial spindle is 2.25" with 0.125" pitch thread, the metric one has a 62 mm mounting diam and a thread diameter of 60 mm with 3 mm pitch.

You will not have a problem with the steel and concrete on the lathe, the thermal coefficients are so close and in the order of 10 to the power of -6/degree C. This combination is used on a lot of the very large CNC machines that I install, build and maintain.
It is the closeness of these Thermal Coefficients that makes it such a good match for reinforced concrete.

Quote:

---

Originally Posted by metric_taper

I hope you document your CNC lathe conversion. I've thought of the same, I'm curious about methods of attaching a shaft encoder to the spindle. Keeping it out of the way for long stock though the spindle hole appears to need some sort of cog belt and timing gears. Then a quality encoder with ball bearing and such is needed, and there must be wear to worry about in all this.

---

I sometimes need to fit an encoder for a cam measuring setup but that method does not allow long work pieces.

Attachment 30077 Click image for full size

I am not convinced that you need a small resolution encoder for most lathe operations. Screw cutting is probably the most severe test. The Mach3 CNC control software only accepts one pulse per spindle rev. That sounds too coarse to me, although when screw cutting the tool load over a revolution should be close to constant and so there should not be much speed variation over a revolution. It would be very easy to mount a pickup to get pulses from the spindle gear.

Attachment 30078

If Mach3 gets away with a single pulse per rev then the number of pulses from the gear should be pretty good. Then the question is finding compatible software.
My mill has servos with linear encoders but if I do the lathe then for simplicity it will use open loop steppers which limits software choices. There is Mach3 and 4 (I do not know whether M4 accepts an arbitrary number of pulses), then LinuxCNC but I have no idea what that accepts. The Centronics Acorn looks like a good controller and software option but once Centronics have you hooked they seem to pile on the charges for the extras needed to make it useful. It is hard to know what the final cost will be.

Quote:

---

Originally Posted by Tonyg

You will not have a problem with the steel and concrete on the lathe, the thermal coefficients are so close and in the order of 10 to the power of -6/degree C. This combination is used on a lot of the very large CNC machines that I install, build and maintain.
It is the closeness of these Thermal Coefficients that makes it such a good match for reinforced concrete.

---

Thanks for that confirmation. I had no real worries about it, I have used the technique before to stabilise various measuring devices but never before in a machine tool. I did not know that it was common practice.

Quote:

---

Originally Posted by tonyfoale

I am not convinced that you need a small resolution encoder for most lathe operations. Screw cutting is probably the most severe test.

---

Tony, what I would like to do, is put an encoder on the spindle, and one on the main feed screw, and fix the timing as the thread dial has shown to be wrong either for metric or Imperial threads. What I do now, is keep the half nut engaged, and reverse the lathe. This gets tiresome. So I'm thinking I can set up an Arduino with the two encoder inputs, and have it give me an engage light when they are timed for the start of a thread. I have not thought this through with all the gearing and such, but this seems like it would enable reliable half nut engagement, so I don't start new threads in the part.

Quote:

---

Originally Posted by metric_taper

Tony, what I would like to do, is put an encoder on the spindle, and one on the main feed screw, and fix the timing as the thread dial has shown to be wrong either for metric or Imperial threads. What I do now, is keep the half nut engaged, and reverse the lathe. This gets tiresome. So I'm thinking I can set up an Arduino with the two encoder inputs, and have it give me an engage light when they are timed for the start of a thread. I have not thought this through with all the gearing and such, but this seems like it would enable reliable half nut engagement, so I don't start new threads in the part.

---

The problem only arises when you want to cut metric threads with an imperial lead screw or vice versa. Remember that you cannot engage at will, the half nut will only engage when the screw is lined up. It should be feasible but depending on the thread pitch and the lead screw pitch you may have to wait some time before everything is back in sync. It will also depend on just where the saddle is when you want to engage. You would have to withdraw the saddle to a stop. If you are serious about making mixed system threading easier I think that you should look into motorising the lead screw (changing to a ball screw is preferable but not essential).
There are a ton of ELS (Electronic Lead Screw) projects on the net. Some are excessively complex but there are some quite practical examples.
If and when I do mine, I'll share it here with Arduino code if that is how I do it.