View Full Version : Machining Crowned Pulleys

Shuswap Pat
06-03-2012, 03:38 PM
I have machined a number of Crowned pulleys over the years, always flat tapered ~~1/8" per foot - to the center of the pulley. Has anyone made an attachment that will machine an actual radius on the pulley face. I have a number of confirations of radius (ball) turning attachments for the lathe, but none will generate a LARGE enough radius. I am guessing ~~ 20 inch / 500mm radius. On a CNC machine it is a piece of cake, but sometimes I like to prove a point, that manual machines can do amazing things.


Forrest Addy
06-03-2012, 07:54 PM
In the lathe '60's I needed a large lathe for a short project. Gig Harbor, a town about 30 mile from my home, used to support a thriving fishing fleet dating from the 1800's meaning a few machine shops were needed. I borrowed a large lathe in the remaining line shaft shop. The place was deserted except for me and the owner. When the owner left on an errand, I wandered around like I owned it looking at stuff. It was interesting, full of artifacts from the past like an un-curated museum. I found large collection jigs and fixtures on lumber shelves on the wall under the line shaft motor.

One of these was an odd gadget. It was intended t go into a standard lantern tool post maybe for the same lathe I was borrowing. It had a stubby tool worked from a double slide; one straight and one slightly curved. In the end of the slide was a tapped hole If you worked the slide the tool slid in and out following the curved path dictated by the cam.

I never understood its function until I read the OP and that image from years ago flashed on my consciousness. It was a pulley crowning tool with a about a 5" face width capacity. Pretty simple really. Even now in my mind's eye I can turn it over and examine its features. It was crude, an obvious shop made expedient but very compact. I can visualize how it was held in the tool post, how a rod went to a bracket attached to the headstock to operate it as the carriage fed, how the flexibility of the rod allowed limited in and out adjustment of the cross slide, how the tool would contour the pulley crown being turned, etc.

The straight slide waa merely a piece of flat stock sliding in a fitted opening. The curved cam was a piece of square stock bent around something with a large diameter maybe 5 ft like a large pulley. The best segment of the bend was hacksawed off to make the cam and screwed to flat stock. The follower was adjustable to minimize clearance and a stiff spring held the tool slide snug to the cam. There were a few crumbs of cast iron stuck in the grease. On the shelf next to it was a collection of flat bar lengths with drilled holes, rods, bolts in a box etc; maybe part of it, maybe not. I sketched out the details I remembered and it seems very workable.

What one old line shaft machinist cooked up in 1910 in one shop I'm sure necessity dictated similar solutions in every other shop in the world where pulleys needed to be crowned on a regular basis.

This pulley crowning gadget and thousands of other crude but efficient devices became junk after their need passed and were sent to the foundry. This is how small details of lost lore never came down to us. Lore that never made it into books or incorporated into manufactured tooling illustrated in catalogs. It all passed on with the last man who remembered them.

Years later the need to crown a pulley emerged leading the machinist with no connection to the past to muse "how the hell did they crown all those pulleys anway?

Maybe the gadget I recalled was one way. How they did it in a production setting in the early days of machine shop technology I don't know but I bet a cam and follower figured into the picture somehow.

06-03-2012, 09:26 PM
A while back I played with an arrangement which pivoted in a fixture held in the tailstock. The other end of the fixture held a cutting tool, and there was a level surface on the carriage on which the fixture could ride. As I pushed it back and forth it would cut on the face of the workpiece.

Eventually the machined part became a concave lens, the curvature of which depended on the length of the fixture between the pivot point and the cutting tool.

Such a method could be adapted to cut a crown by pivoting the fixture somewhere behind the lathe, with the pivot point directly behind the center of the roller being machined.

For a crown of 1/8 inch in a foot, the distance from pivot point to tool would be at least 10 ft, maybe more. I haven't done the math.

Seems unwieldly, but it depends on your shop layout. I have only 2 ft from my lathe to the wall, so it would be impossible to set up here. I'm not saying this is a good method- just that it's one way it could be done.

A more practical method would probably involve making a guide piece with the right curvature on it, and attach that behind the crosslide. Then you make all the other arrangements to have the cutting tool holder follow that guide.

Here's another idea- set up the roller between centers. At the tailstock end, the center is made to move front/rear as the carriage goes left/right. At the center of carriage travel, it is the furthest rearward, causing the center of the roller being machined to be fat. Good luck figuring out the geometry for this :)

Or- how about exploring the ways in which a mis-aligned lathe can turn a taper, or more specifically a fat center on a length. It's easy enough to do when you are trying to turn a well-fitted morse taper :)

06-03-2012, 10:21 PM
I simply rough them out in the form of square steps, then use a file to smooth them out - the same method as prescribed for other oddball profiles. In this instance, youre not really concerned with a specific tolerance, simply getting an approximate diameter and crown on the pulley.

06-03-2012, 11:01 PM
there is also a form called a trapezoid that will be easier to make. Also another for consideration is a surface sleeve. Have personally used the first with success.

06-04-2012, 01:48 AM
Easy way is to use a taper attachment and simply turn two opposing tapers that terminate in the center of the crowned face using the 1/8"per foot rule of thumb.Then round and fair in the center with files and sand paper.It's not rocket surgery and many a wooden pulley has been done by eye with hand held turning tools.

I've used that method on flatbelt and sanding belt pulleys over the years and never had an issue.

J Tiers
06-04-2012, 01:56 AM
Or just set the compound if the travel is enough..... most of the pulleys I have seen "new" had the double straight taper, and a small area of untapered "flat" in between them.... maybe 1/8" or a bit more on a 2" wide pulley.

Optics Curmudgeon
06-04-2012, 02:11 AM
The way I've done this in the past was to make a sheet metal template with the correct radius and clamp it to the tailstock ram (just a convenient place). With an indicator mounted to the cross slide you can trace the radius well enough to finish with a file.

06-04-2012, 03:21 AM
I read in an old book I have somewhere, a method of facing a radius.
A punch mark is made on the headstock, centered under the spindle.
A second mark is made on the cross slide, the moving part of it.
A rod with pointed ends is prepared with length calculated for radius desired.
The points are placed in the punch marks.
Hand pressure is placed on the long. travel handwheel to trap the rod in the punch marks.
The cross slide is then traversed across the face of the workpiece to create the radius.

There is an equation for calculating the length of rod needed for radius desired.
I need to remember which book and look it up.
An adjustable length rod makes this scheme easier.

One difficult thing about this scheme is that the work can not always be mounted close to the headstock. For instance a large radius would require quite a long rod, requiring either a long tool to reach the work or mounting the work away from the headstock, so the tool bit will reach it.

This same method can be used for creating a radius on turned work.
The setup would be different. Possibly a bracket would have to be fastened to the bed, to accept a punch mark.

If you have a taper attachment, you may be able to replace the straight bar with a curved one of the desired radius.

Forrest Addy
06-04-2012, 04:06 AM
I've fake cutting radii countless times by manually moving two axes to rectangular coordinates. I used a calculator for this since 1970 but the technology has moved on.

Rectangular coordinates are easy in these days of computers with spreadsheet software. Run formulae for sine-cosine values at regular incremets, print the list and cut the radius or whatever. I used two dial indicators for years until I modernized and instralled DRO's. My newest DRO is a Newall. I use the hole pattern function set to drill a couple hundred holes. Then I use the hole offsets on the display to plunge into the radius with a suitable tool. You dont have to drill a hole to use it. The same feature works for tool positioning too.

Once you get the radius roughed in you can dress off the cusps.

But it would go easier if the tool followed a continuous path like a template, cam, or a radius bar. There are four bar linkages whose motion over a restricted path approximates a circular are - close enough for a pulley crown anyway. Google "4 bar linkage" I think engineering schools offer extra credits for posting linkage crap.

Shuswap Pat
06-04-2012, 04:28 PM
Okay - Plan 'B' - does anyone have a slick plan for a simple tracing attachment - to 'Trace' the crown on the pulleys?


06-04-2012, 05:22 PM
It wouldn't be too difficult to rig up a mechanical tracing attachment.

A weight attached to the cross slide and hanging down on the backside of the lathe. Disconnect the cross slide screw anchor and let the weight pull the cross slide up against a suitable roller that bears on a suitable template. Make the template wider than the pulley so the tool/cross slide doesn't run for home when the tool pressure isn't there. Cutting depth otherwise regulated with the compound.


06-04-2012, 05:47 PM
A few years ago HSM, PIM or MW had an article about machining crowned pulleys using a form tool. Sorry, but I can't sight the specific magazine of issue for you, but perhaps someone here can.

06-04-2012, 06:10 PM
A long time ago I had a job making lens ring joint gaskets, or "lens rings" as they were called. The faces were spherical, anywhere from 6-16" radius, depending on size. We had some large homemade radius gauges, ground a tool to match and then cut the face with the large radius tool. Once it was close, checked symmetry with the radius gauge and then polished by hand. Probably made 1000's of them that way. The stainless ones were tough as you would always end up with some chatter that had to be polished out. Wow, was that primitive or what! I know that's not what you have in mind but that's how we did a large radius.

06-04-2012, 06:55 PM
You might be able to do it on a mill with a fly cutter. and spin jig or rotary table. You could get a large radius by setting the fly cutter on center for thickness and off center radially, then turning the pulley around underneath the cutter. The resulting geometry would not be a true radius, but I doubt if the belt would be able to calculate the difference. It would be a quick easy setup and nice and even all round.

06-05-2012, 12:06 AM
Clamp a (hand-made will do) template to the back of the lathe bed.

Set the top-slide over the cross slide (both parallel and at right angle to bed axis.

Put an indicator in the cross-slide and a turning tool in the top slide.

Zero the indicator on the template using the cross slide.

Set the tool using the top-slide.

Use the carriage hand-wheel to traverse the cutter and the indicator.

Feed the tool in using the top-slide.

When traversing, use the top-slide to keep the indicator on zero to replicate the template form on the pulley.

06-05-2012, 07:34 AM
Okay - Plan 'B' - does anyone have a slick plan for a simple tracing attachment - to 'Trace' the crown on the pulleys?

What you want is a profile attachment - a variant of the taper attachment, except with a replaceable cam plate instead of the straight track of the taper attachment. A properly-built profile attachment would do ordinary tapers too, of course.

Once you have the profiler, any sort of crowned pulley, gun barrel, large bowl shape, chair leg, etc. that you want to make simply needs a cam made out of whatever is at hand.

If you just need to git-r-done, nibble the basic shape with a narrow tool or cutoff tool, finish with a file.

06-05-2012, 08:13 PM
Here is how I do it.

Another shot


Mount the pulley on a arbor ,use a rotary extend the radius with a heavy bar and mount your tool post . now you can just about cut any radius you may need.

Ohio Mike
06-06-2012, 07:46 PM
Bonus points for clever use of a Van Norman. :D

06-06-2012, 10:02 PM
You have a ball turning attachment- maybe you could extend the handle and ride the end of the handle against a template.

06-06-2012, 11:41 PM
Lane wins hands down---easiest, most accurate method yet (as long as you have a horizontal!) :D