[SEL] semi-OT Graham in Oz's antique engine gear cutting

[Richard Allen]

To: Graham in Oz

Hi, there, Graham!

For 56 teeth you will be fine using a cutter for 54-55 teeth. I am very
surprised that the cutter is specified for only 54 and 55 teeth because
with that number they are in the range of being able to cut at least
four consecutive ranges of teeth acceptably, in this case it would
probably be 53, 54, 55, and 56. The cutter possibly could be for high
speed performance gears, which is of little concern for your
application. Or it is for helical gearing and won't work at all. You
are quite right about it being correct for only one pitch and all the
others being close approximations, but at that great number of teeth
the differences in the individual involute curves between 54 and 56
teeth are getting to be very slight. The biggest problem you might have
is that the teeth will be a little too fat and may not have enough
clearance between them, but I doubt it because the crank gear probably
has made a few turns by now and is worn a bit! In case the new gear
teeth are too fat, that can be fixed pretty easily by either slightly
re-indexing the newly cut gear on the index head BEFORE loosening or
removing it from the collet or chuck so it cuts an additional 0.125
millimeter (~ 0.005-inch) from the contact faces of one side of the
teeth, or by lapping on a babbit gear that has been cut with the same
cutter. Don’t lap it using your original crankshaft mating gear, it
probably has enough wear on it already! If you have enough babbit or
lead wheel weights lying around, cast a gear blank a bit over the size
needed to cut a 56-tooth gear and make all your experiments on it (in
my shop I call mistakes my “set-up experiments” {:>). Before casting
your babbit, flux it by stirring it up with rosin, candle wax, or even
some sugar or sawdust that is floating on top and burning, and then
skim the dross off to get any dirt and grit out. Don’t have anything
damp in your mold because it will explode with quite surprising
violence all over the place, or make it from sand because it will embed
and ruin your cutter. For casting round babbit biscuits I use a piece
of iron banding strap from a shipping crate rolled up into a circle
lying on a thick steel plate. Machine your soft lead alloy blank to
size and then try cutting a few teeth on that so you get all the
surprises out of the way. After cutting a few teeth, if it comes out so
that the cutter makes teeth too fat to mate with the crank gear, then
re-index it and re-cut those few teeth until it mates, and then when
cutting the real gear follow the same procedure. 

If you want to lap the new gear to make it fit better or smooth it up,
re-cast the set-up babbit gear and machine it for 55-teeth instead of
56. This is because when lapping gears with the same number of teeth,
the timing MUST be changed one tooth at a time until all the teeth have
been gone through, otherwise each tooth becomes mated with only one
particular notch! To avoid that, make the lapping gear with 55 teeth
instead of 56 so it has a floating or “hunting tooth”. That is why gear
trains always have every-other intermediate gear with a prime number of
teeth, or the sum of the numbers of teeth in each pair of gears in a
train is a prime number, or each pair of gears has a number of teeth
that do not share a common prime numerical factor with the number of
teeth on the other gear. All that is to make sure that all the teeth
eventually come into contact with one another and spread the wear
evenly. Gear shift transmissions where the gear pairs are constantly
disengaged and then re-engaged with random timing, engine camshaft
drives with only two gears, and mechanical calculating machines (like
old-fashioned cash registers) that rely on specific ratios to work, are
the exceptions. Camshaft drives almost always have a third drive
component, either an idler gear or drive chain, with prime number
indexing so as to avoid uneven wear. Where there is no “hunting tooth”
in a drive it MUST be re-assembling with the original gear tooth timing
to prevent unfamiliar gear tooth mating, which causes galling from very
high point contact pressure from wear-misshapen surfaces. Valve
camshaft drives of course already have timing marks.

Too bad babbit isn’t a little tougher, you could make the real gear
from it. For very light duty at slow speeds you can, actually. There
are some pretty good grades of aluminum around now that machine easily
but have good wear characteristics that you could make the working gear
from, unless you want it to be made from the original material for
authenticity. I melt down old zinc Zamac die cast metal (don’t
accidentally get any aluminum die cast in it, it will ruin everything)
from junk to use for a lot of things like this since it works great for
light duty gears. The gears in my 65-year old Atlas lathe are Zamac,
which is pretty good testimony to the durability of that alloy.

Rich Allen

Re:
Date: Mon, 31 Jan 2005 14:23:00 +1100 (EST)
From: Graham Harris <ozengine at yahoo.com.au>
Subject: [SEL] re: Gear making in brief - a question

Hi Rich and Tommy

N-i-c-e practical info on gears mates. I's even saved
me a copy.

Got one question that no one seems to be able to
answer for me. I would like to have a go at cutting a
cam gear for my Novo 4. Needed is a 56 tooth gear, 10
pitch and 14.5 PA I'd assume. I have a involute gear
cutter here that cuts the 54-55 tooth range (or
whatever is the standard teeth count for the numbered
cutter). So the question is...can I use this cutter
indexed correctly to do the 56 teeth? I mean, how much
"out" would it be? My understanding is that the
cutters are only exact for one teeth count and all
others in the range are close approximations.

Thanks much

Graham in Oz