A hob with helical teeth would just cut away all your gear blank if it were not rotated in time with the hob. But you can make straight tooth hobs with just a Rack form that can be used by indexing the blank and they will form close to the correct curved profile on the teeth.

Isn't that the hat from the first Harry Potter film?

Wow! To both the machine set up and what must surely the the world's most comprehensive set of Myford change gears.

Phil P,

beautiful machine. Nice work. Kudos.

Makes me wonder who had the most patience; your Dad; or your Mom. Nonetheless a beautiful piece of kit.

Just to put the subject back on track.

If you make a hob with circular grooves which conform to the rack tooth dimensions, that is a hob with NO helix angle.

You can take one pass through the blank mounted on a dividing head and with the hob centred, then index one tooth and take another pass until you have gone all the way round.

You will get a good approximation of an involute tooth form, the difference being that it will be made up of a lot of tiny flats as opposed to being a generated curve, for a lot of slow rpm jobs they will perform adequately.

Phil

John, the cutter will cut several teeth at one setting so you get a sort of faceted shape to the tooth,here is one I prepared earlier

Edited By JasonB on 02/01/2019 09:57:06

Incorrect; it's basically the same technique as the "Sunderland" planer shown by Phil. As illustrated by Jason you end up with the correct tooth form, but approximated by a series of straight lines. If you index by a fraction of a tooth width and offset the hob you simply increase the number of straight lines, and they're shorter, so you end up with a better approximation.

When free hobbing a worm wheel the last operation, once at full depth, was to move the worm wheel along the axis of the hob by a fraction of the worm pitch:

If I recall I did four increments; at each increment cutting noise could be heard, until the worm wheel had made one rotation.

Andrew

You would need to use less table traverse per stroke to keep the facet-count down and reduce chip load on smaller tooth counts but the shape would always be correct IMO.

Yes, I can confirm the parallel hob approach works well:

A major advantage is that it makes  cutting gears with non-standard centre distances very easy, although if overdone the tooth shapes start to suffer.

Edited By Neil Wyatt on 02/01/2019 16:30:04

Edited By Neil Wyatt on 02/01/2019 16:32:33

That could only be possible if each stroke indexed one tooth, which isn't how it works. The tooth form is generated by moving the table and rotating the blank a tiny amount between strokes, not a full 1-tooth index.

No, you just index 1 tooth and don't move the hob, as Neils box of gears confirms

You get a better curved profile if both are moved like a Sunderland machine, as JS's post part way down this page shows

Edited By JasonB on 02/01/2019 18:39:37