Lathe Modifications  

this includes anything based mostly on the gingery book series.

modifications go here

fenn's mods:  

vibration damping concrete pedestal
compound slide/radius cutter/milling attachment
boring table cross slide
brass split nut w/adjustable ball detent lock
ball bearing feedscrews (made after lathe was functional)
height adjustable tool post/boring bar holder
ACME leadscrew (1/2"-10 w/delrin nut)
127K9720 pancake servo motors for computer control
planned/in progress:
lever feed micrometer tailstock (oops - left castings in galesburg, illinois)
dividing headstock & spindle brake & spindle clutch
MT2 inside 1"-8tpi outside 1/2" through bore spindle
longer crank handle brackets to take advantage of full travel
mouse-wheel DRO (with AVR and 7 segment displays from electronic goldmine)
spindle encoder for cnc threading
not planned but wish I woulda:
steering-wheel style handles (not yet)
offset leadscrew crank bracket to clear tailstock removal
wider carriage and wider cross slide ways
concrete-filled steel tube bed
headstock bracing to compensate for taper roller bearing preload
design headstock around large taper-roller bearings
bad ideas/gave up:
bicycle "change sprockets" - highly dangerous, unreliable
stepper motor control - stepper motors suck
micrometer collars - doesnt work so well with odd number of thousandths per rev, from 16TPI leadscrew


The following are comments I clipped from the gingery_machines archive for my own use.

There are three major problems with the Gingery Lathe. First for the size and strength of both the bed and the bedway, it should only be able to turn 3.5" stuff. The bed just isn't strong enough when made in aluminium to deal with a 7" lathe. My research into this shows that the bed casting should be at least the same width and height as the distance from the center of the lathe down to the ways.

The ways need to be 120% of this width. That means they really should be made from a piece of 5" CRS that is 1/2" thick and the bed casting needs to be 4x4.

The kicker here is that you will have to adjust the design of nearly every casting to accomodate the larger ways plate. However the advantage is very much worth the effort.

One other thing to look at, since you'll be re-sizing all of the castings anyway:
Make the cross-slide correspondingly wider. I forgot to do this when I cast the pieces, and had a horrible problem with the tool bit bouncing sideways during most cuts. I finally gave up and made a completely new cross-slide which is working MUCH better.

The way that the CRS is mounted to the bed also allows significant flex. Like in the Mill Book, the ways should be screwed down with two screws every two inches or so, instead of just one stud. This will prevent the way plate from lifting up off the bed.

The bedway needs to be shifted forward a bit so that the leadscrew sits more directly underneath the bedway. The way it is in the stock book is that it creates a twisting moment on the carriage which makes it harder to pull/push. Even sillier was that the half-nut was positioned to the right of the carriage. I mounted mine so it was underneath the carriage. My carriage is a bit difficult to move by hand but goes effortlessly with the leadscrew. Again, why apply a twisting force to it?

  • Finally, and this is a big issue that I have yet to solve, is that the compound shaft and motor assembly is a simple but unfortunately poor design. Vibration from the motor, slightly rough belts and pulleys that are ever so slightly out of round can make the whole thing vibrate big time causing poor cuts. With a dial indicator, the can see the headstock moving back and forth by as much as 0.002" and it's even worse if I tighten up the belt tension.
    Obviously, I need to make rounder pulleys but realistically, changing the countershaft so it's braced against the headstock will go a long way to getting rid of some of the harmonics in the cuts. I intend on making a new base and belt/pulley assembly.

Someone described how Gingery improved the bed design on the miller [from that of the lathe]. I do not know if it is easier to pattern and/or cast, but it does look sturdier. I have also not seen a picture of a lathe made with the miller's bed design.

See: for a constructed Miller used primarily as a lathe.

Do not make the Headstock support foot the same as the tailstock support foot. It needs to be as wide as +15% of the headstock.

Commercial bed designs mostly flare toward the bottom. The way the miller bed is cast, the sides and ends taper inward toward the bottom. Would it not be better and easier to cast the sides/ends separate from
the top, and then screw the bed top on so the sides taper outward at the bottom? The bed could be made larger/beefier because the top and sides were cast separately (due to limits on amout of metal melted at one time). Also, if this is done, one could simply just omit the bed top, and just use a thicker piece of CRS for the ways.

See: for an example as applied to the lathe.

"The Milling Machine, Chapter V -- The Spindle Head" begins with:

"You'll notice that the photos of my miller show the bearing support screwed to a tongue on the shank of the spindle head slide. That was a poor design, and it proved much easier to cast the shank straight, and screw the support casting to the shank rather than cast it as a tongue."

This same tongue exists on the slides for the lathe. Would it not be wise to use this improved design on the lathe, because: 1) it is easier to pattern and cast; and 2) the screwed-on screw bearing support provides a large bearing surface?

I was examining the [lack of] straightness of the feed screws I had just cut for the lathe, and thinking I was glad there were only two because I have learned that I lack the ability to start threads on a rod without sweating and swearing [and, of course, the lack of straightness when I was done]. I thought to myself that surely it must be easier to take some threaded rod and epoxy a sleeve over the threads on the end that you want smooth so it will turn in a bearing.

The miller has a much better design though, with a spindle running through the bearing support and attached to a pre-threaded rod with rivets. This is a better design because there is no threading to be done, the srew-rod to spindle attachment with rivets provides a sort of universal joint to help with any manufacturing inaccuracies, and because acme threaded rod can be used.

It is a problem when making the lathe, though, because the spindles must be turned. They could not be turned on the makeshift lathe we use to file the lead screw bearing journals for the lathe. I have not figured out how to overcome this obstacle yet.


The compound is probably the weakest point on the entire lathe. It sits high on relatively flat surfaces. A much better design is the compound for the milling machine, featured in the chapter on the "Universal Compound." This one is much more rigid and is much less prone to flexing. The only downside is that you need a lathe to machine this version.

A proposed work-around: build the lathe as shown in the book (with some of the modifications listed above) and then make a new compound. The cross slide will have to be redesigned as well in order to use the improved compound, but if you have just built the lathe, making a new cross slide and compound will be easy.

I stripped out the four tapped holes that hold the compound at a particular angle on my lathe just a couple of months after I finished it.
What I did was to drill and tap the holes to a larger size (either
5/16" or 3/8"). I put thread locking compound on four bolts and put
them in the holes. After they set up I cut the part sticking out off
and filed them flat. Then I drilled and tapped the bolt for 1/4"
threads. It's worked well for ten years now.
I also had the same problem with the holes for the toolpost and
fixed them the same way. *1

links go here part gingery lathe, part monstrosity! tube bed construction CRS bed construction

*1 John Schwytzer

Front page   Diff Backup Reload   List of pages Search Recent changes   Help   RSS of recent changes
Last-modified: Sun, 04 May 2008 18:13:33 GMT (982d)
The Gingery Machines wiki is published under the
Creative Commons Attribution-Share Alike 2.5 License Creative Commons License