Remember that old article by Base Circle ?

This guy takes us on his own ‘voyage of discovery’

**LINK**

https://youtu.be/so4O_yxUfCE

MichaelG.

I often find these "oops, that didn't go quite to plan" you tube videos more interesting than the" look how I did a wonderful job" ones.

Especially when the oops, as in this case, turns out to be exactly the reason why I've not tried myself.

The indexing issue is rather glossed over in the original article.

Seems to me that a better way would be to self index on the gear being made by registering "180° + one tooth" opposite the previous cut. That way any errors get spread around and you don't need another gear of the same tooth count for indexing or have to drill a hole circle in a dividing plate. Doesn't have to be 180°, pretty much any angle will do so long as the offset is one tooth space.

I wonder if a section of steel band drive belt would work in place of the wire?

Clive

The index locking pin system is really critical or you get errors that mess things up

The whole process is highly dependent on virtually no appreciable errors when using a shaper and is not easy to achieve when being set up from scratch

.

Some previous discussion on this forum covered that, Clive

... If I can find the relevant thread, I will post a link

MichaelG.

Accurate indexing is always the key to making good gears. Most especially avoiding cumulative errors.

Consider a 36 tooth gear. For practical purposes each tooth is 5° wide so a 0.1° indexing error on each shift means 3.6° by the time you get to the last tooth. Which will be very thin.

Spread the 0.1° error out evenly around all the teeth and all will be a touch thin but the gear should be fine for undemanding applications like change wheels. The method does inherently produce a correct involute form within the inherent limitations of being generated from many small line cuts. So I'd be unsurprised to discover that the result of evenly spread indexing error corresponds to the sort of addendum and dedendum modification used to adjust the centre to centre distance when standard gears won't fit. If so the gear should run just fine against a properly sized one.

Clive

.

Start here, Clive : **LINK**

https://www.model-engineer.co.uk/forums/postings.asp?th=124817

MichaelG.

Thanks for the link Michael. I now remember reading that at the time.

Clive

These Guys got the 'Base Circle' method to work some three years back - as part of an "Emma's Spareroom" challenge. They had similar issues with indexing and PCDs at first if I remember correctly.

They used a more robust (purpose made) indexing head that probably didn't have as much give in it as the other guys angle-plate setup. Interesting use of 3D printed parts too (and they even managed to give a tongue-in-cheek plug to Stephan G.)

Shaper Cut Gears from Oz

Regards,

IanT

Edited By IanT on 28/10/2020 16:24:57