Skew bevel gears were mentioned on another forum and the discussion sparked my interest. There seems to be very little in the literature regarding skew bevel gears, other than their use on some logging locomotives. I've found one reference to them in one of my gear books, a third of a page and one figure!

Normal straight tooth bevel gears have intersecting axes often, but not always, at right angles. The idea of skew gears to use the straight tooth form but with axes that do not intersect. The basic idea is that the tooth is formed by extruding the involute form to a point along a line that is tangent to a circle within the gear, rather than a point at the centre. The axis of the mating pinion is also tangent to the same circle. Here's the basic construction:

And tidied up:

There seem to be a number of issues with these gears. First, since each tooth is translated as well as rotated from the previous tooth the root of each tooth does not match the previous one, either linearly or in slope. Of course this can be fudged, and as it is at the root shouldn't affect the gear operation.

A more interesting issue seems to be that each tooth not only rotates pass the mating gear but is also translated. So as well as the normal rolling contact there must be sliding contact. This also implies that as the gear rotates the effective DP of the tooth changes as it passes the mating gear. This hardly seems conducive to smooth running.

It is obviously possible to make skew bevel gears that work, but they would seem to be very inefficient. Is that why there is so little about them in the literature?

Andrew

Thanks for the replies. I am aware of the work by Kiraoka. If I understand his work correctly he provides detailed numbers for cutting the gears as parallel depth bevels using three passes. All fairly standard, notwithstanding the offset of the teeth. If one looks at pictures of his skew bevel gears the involute form on both gear and pinion are very distorted, and not symmetric. I suspect that is a function of the way they are cut. Essentially the spaces are being formed rather than the teeth, and the process generates different curves on each side of the real tooth. I've also found pictures of skew bevel gears by Britnall (no mean model engineer himself) and they are the same – the involute form is quite distorted. Clearly these gears work well, but don't seem to tie up with the figures in my gear design books.

I have found a youtube video of home made skew bevel gears where the involute form seems to be symmetric. They look very different from those made by Kiraoka et al. In the video they are shown running, so clearly work as well. Although they have a few more teeth than shown on my gear above the involute forms are almost rack like on both gear and pinion, which is odd.

I've also found a lithograph of the wheel truck for a Climax logging locomotive showing the arrangement of the skew gears. Although not terribly clear the tooth forms do look rack like and are clearly symmetric. So while the Kiraoka method produces working gears they don't seem to represent full size. What interests me is how the full size gears were designed. If I can draw it I can probably make it, as I am not limited by the machines available.

All of the skew bevel gears I have seen have rather more teeth than my CAD model. With no evidence to back it up I'd surmise that the technique works best for relatively large numbers of teeth on the gear, where the mismatching from tooth to tooth is small.

On a discussion on another forum I found confirmation of my thought that the gear teeth slide as well as rotate when passing the pinion. A sliding as well as rotating action combined with the exposed gears and the crud associated with logging operations must have tested the life of the gears on full size engines. Even lubrication may not have been advantageous, it might just attract the dust and dirt!

Neil: Are the involute tooth shapes on your 'accidental' skew bevel gears symmetric or not?

Andrew