Home‎ > ‎

Gear Cutting

The following represents a series of experiments on gear cutting using a vertical milling machine. 



I have spent far too much time making a Rack Cutting Attachment for a
vertical mill.  But at last it is 'finished' (useable is a better word).

It's one attachment I have always coveted since cutting a longish rack -
especially on a vertical mill - is interesting to say the least.

The sleeve which fits the quill is intended to take various different bolt
on devices as necessary.

I very nearly made it with a quite long thin universal jointed drive shaft
leading to a seperate motor or drill for power.  The cutter could have been
mounted under the centreline of the quill with this arrangement which would
have been a lot better.

But a set of nice hardened gears from an outboard enticed me to do it this
way (see pic) using the existing mill spindle for power.  Because of the
offset, it imposes large rotational  forces on the quill.  But it is for
light duty anyway so I hope it will be OK.  Just have to remember its
limitations.  Somehow I think the original idea would have been better.
Small racks and other small Horizontal Milling cutters and slitting saws in
arrangements where this is more suitable.

The spindle uses 20 mm ID/42 mm OD ball bearings but the vertical drive spindle
is steel in a bronze sleeve.  Maybe this would have been better in ball
bearings too.  It warms up in use more than I expected even when well oiled.

The rack in the picture is 20 DP.  Well, it was cut with a 20 DP cutter.
But the teeth are spaced at 4.0 mm.
That represents a 0.2 % error and seems to run nicely against a normal 20 DP
gear.  I suppose I could expand the diameter of the pinion by 0.2 % but I
think this is splitting straws.  The 4 mm spacing was too good to pass up on
a mill with metric table screw and little wear which made cutting the rack a


  At last I have my botched up gear hobbing machine working.

Not sure if it is the most sensible way to make a gear hobbing machine.  Its
different from most approaches.

It will be very versatile - easy to change the helix angle and all
movements are catered for.  I had another idea to make a worm driven gearbox
with change gears attached held in the mill vice - easy to adjust angle

But I had the dividing head although its almost too big.

The dividing head is of course unlocked so it drives via the gears,..  The
handles are dangerous flying around and one needs to keep well clear.
It would be better to remove them.

The gear on the vertical mill nose is on a ring which grub screws on.  Wide
short well fitting rings are a curse.  It would have been better clamped
with a pinch bolt.  An alternative would have been to attach to an ER32
collet holder - maybe even permanently since they are cheap enough.
A subsequent article on hobbing small numbers of teeth below shows a gear
shrunk on to a normal NT30 - ER32 collet holder which is worth considering. 

I have built up a collection of smallish hobs from eBay over the years.  96,
64, 48, 32, 24, 20, 16 DP.  The 96 is probably finer than any needs I have.  The set up
above is for 60T of DP 32.  Diameter just under 2 inch.  All the hobs are held on arbors
which fit into a 3/4 inch collet.

I tried some gears in PVC and brass.  Works well but if you get the
calculations wrong the result is a weird pattern rather like a rose engine
produces.  Also too fast a sideways feed produces a peculiar effect.

Tried hobbing both a spur gear and some worm gears using some common taps
which are a equivalent to makeshift hob of pressure angle = 30 degrees.
Small number of teeth are however inefficient and require very slow feed.

The bevel gears I made previously are seen on the quill attachment.  They
work well.

While the hobs can produce any tooth number, I don't have a sufficient range
of timing gears  for many.  Would have to produce them first.  In this
respect, for odd gears of strange tooth number cutting with normal cutters
has an advantage.  But one needs a whole lot of cutters.

Something obvious.  Its not very sensible to produce a long gear then part
off smaller gears.  The burr is very hard to remove without some damage to
the gear teeth - depending on material.  Better to produce blanks first,
stack, then hob the lot.

A 40:1 standard dividing head is well suited to hobbing gears with many teeth but
for small gears eg. 12 teeth, the work has to be geared UP many times.
This places enormous strain on the gear train and universal joints and is not recommended.
In this respect it would be better with a different drive system. 





Since I wanted to make very small gears with few teeth I made the device shown
below.  The ratio of crown wheel and pinion is 4:1 ( 80/20 teeth) and the chain drive adds
another 2:1 so this is equivalent to a dividing head of 8:1.
Previously I had made some bevel Mitre gears which were very successful (20 teeth).  The
4:1 Crown wheel and Pinion was not so successful.  Something moved during the cutting
 of the Crown wheel, and the calculations for offset and roll are not perfect and I should make
 another perfect gear.  The gears when cut show slight errors because of this.  
If one could obtain a precision hardened Crown wheel and Pinion as used in a car differential
of simple ratio this would be a perfect solution.  It would be possible to place strange gears of prime numbers
somewhere else in the drive train but this is messy.
I have since obtained a pair of precision helical cut spur gears of 4:1 and am considering a rebuild using them.
But that needs more time which is lacking.
The bevel arrangement, allows the pivot point of the mechanism to be close to where the gear to be cut is
 mounted which makes a useful simplification when tilting the blank in the vice for different helix angles.

 The work holding spindle is a standard ER32 collet holder with hardened 20 mm
shank.  This is a very easy way to get a precision spindle.  Cheap from Hong Kong and quite
 good quality. It has been mounted in standard 20 mm ID ball bearings
using a spacer to give ideal preload.  The Crown Wheel mounting ring was shrunk onto the collet holder.
So the workholding device might not look pretty but it is a precision item apart from an imperfect crown wheel.

 It is simple to change helix angle using the vice but a better fine adjustment would be nice.
 This setup for small gears has been thrown togeather for a few pictures.
The gear train is for a 16 tooth gear but the hob in place and its position
may not be very sensible.





 These have been cut with the attachments above.  They represent a series of experiments
embodying a learning curve.  Most have been made in either PVC, a few in Polythene, and Brass.
The skew gear in PVC is a bad idea because of heat/friction but it validated the process.
The worm gears used 3/16 WW and 1/2 UNC taps as hobs.
Proper proportioned worm wheels really need purpose made hobs since normal
gear hobs have too large a diameter to replicate a practical worm.
Correct calculations, setting of correct attitude for helix angle, and considerationsof the
 handedness of the helical gears (except skew gears) are essential.  This is a study in itself.
In general I am happy with the results and have learned a lot.
The universal jointed shaft is best made as long as possible - or even longer.
The timing gears need to be mounted in a very versatile way.   It is often necessary to 
introduce idler gears simply to facilitate lining up the universal jointed shaft. 
The arbor holding the blanks should be as long as possible consistent with accuracy
 and stiffness to remove the  hob from close vicinity to the crown wheel.
It would be very nice to have both Hob and Blank held on arbors supported at each end or at
least stiffened with centre at their far ends.  The hob could then be a larger distance from the
blank support system and dividing head.  But for small light duty gears, it is useable.
One needs an understanding of the mathematics of gears to arrange for the correct diameter, hand, depth of cut,
and helix angle - especially for helical gears.  Without correct setting of these parameters the results are
useless.  It is often worth doing a trial cut in Plastic first.  If one calculates incorrectly, the result is usually
horrible but how it will be affected is often not intuitive.  Don't ask how I know.......
The calculations are more complex than I expected and it is worth whipping up a small program or Excel
spreadsheet to minimise calculation errors.
An ideal  setup for producing large numbers of teeth quite different from that for producing very small
 teeth numbers.  The drive may very well be best applied to the cutter in one case and the blank in another.
This is especially apparent with the vertical mill setup.  In this respect, a dedicated machine would be better
using a dividing head or worm of maybe no more then 12:1.  Certainly 40:1 or 60:1 is a bad idea unless one
 is contemplating making gears with many teeth.  I wanted to make 8 tooth skew gears and this was in practice
impossible with the dividing head version.  But the dividing head version was perfect for from about
 60 - 120 teeth.  This is obvious on reflection but I had spent years collecting suitable high ratio worm sets
with the intention of making a hobbing setup without giving it sufficient thought.
The universals I made using cone pointed hardened screws working in bronze using a bandsawed Al fork.
  They are a weak point and need occasional readjustment.  It would be nice to have the very best universals
 possible.  But used with care they do work and are cheap.
The idea of using a milling machine vice to facilitate helix angle adjustment is simple and effective but one
needs some means of angle measurement.  A reference surface and clinometer, or some graduated scale
are candidates. 
The device which clamps around the quill in two halves has been very successful.  It is much easier to make and
 apply than the long sleeve used for the Rack Cutting device.
Shrinking a mounting ring onto a commercial parallel shaft ER32 collet holder has worked out really well.  This was
done in Al - partly because it is easier to shrink on and was available.  The shrunk on ring was then turned
very carefully true to the body.  The commercial 20 mm diameter running in standard ball bearings was easy and 
very precise.  Fairly cheap, stays cool,  and never needs lubrication. 
Tilting a standard dividing head to harge helix angles really eats up a lot of height fast.  In this respect, the crown
 wheel version has an advantage.
The traditional way of making gears on a mill using a set of gear cutters has merit - especially when an
unusual prime number of teeth is required.  But for large numbers of teeth and helical gears the hobbing
 process is far more convenient.