RobCox
Member
- Messages
- 402
- Location
- Cambridge, UK
If you've seen several videos on youtube of Heath Robinson creations for cutting spur gears on a shaper and liked them, you might like this. I watched videos by Geoffrey Croker and Paul Hopewell ("The Shed Dweller") and had struggled to come up with an improvement to their ideas, until a couple of months ago.
I wasn't keen on the wire and cylinder technique as it looks like a) it could slip and ruin the work and b) you need a different cylinder for every pitch diameter gear you cut. The idea that Paul came up with of replacing the wire and cylinder with a lever was better but a) it doesn't quite generate an involute (not far out though) and b) the mechanism for setting the rotation of the gear blank as it travelled didn't look that precise.
For the uninitiated, the idea is to use a simple HSS cutter with a shape matching a rack tooth (straight sides) to cut involute gear teeth. I can't be the only one who detests spending £25+ for a cutter each time I need to make a gear size I haven't done before.
The light bulb moment came when I was messing about with the vertical feed mechanism that I'd made for my Elliott 10M. The same mechanism is used to traverse the table side to side, but the gear wheel is moved from the vertical feed screw to the cross slide screw:
As the aim is to rotate the workpiece exactly proportional to the linear travel, I figured I could tap off the cross slide gear with another shaft as that achieves this. The next problem is to hold the work, index it but also rotate it as the table traverses. Well, I have an Elliott universal dividing head which achieves just that! So all that was needed was an attachment to join the cross slide gear to the dividing head with a gear train with the right gear ratio. I designed it initially on a 3D package (Designspark) because its quicker and cheaper than making mistakes in metal. One over-riding aim was that there should be no permanent mods to the shaper, no drill holes for example, just clamp on:
Gears here are just drawn as discs. The gear on the right picks up the drive from the cross slide gear. It has an internal spline to drive a shaft that runs side to side. The left hand end of this shaft is bolted to the table and moves left and right with it, hence the spline. A pair of 45 degree helical gears turn the drive through 90 degrees onto another shaft bolted to the T slot in the side of the table. The end of this second shaft will carry the first gear in a train that drives the dividing head universal drive spindle.
After a couple of months of cutting metal it's finally done and here it is bit by bit going on the shaper.
Pick-up gear assembly, clamped to the ways:
Splined feed shaft before its bracket is bolted to the table:
This is then bolted on so everything runs smoothly:
The next shaft has a bracket with slots for T nuts and studs so the mesh of the helical gears can be adjusted:
And here it is bolted to the table:
Finally the gear train is fitted to the dividing head banjo to join everything up:
Sorry about the blurred images. Note the spindle locking lever has been removed from the dividing head and replaced with a bolt to keep the chips out. Clearance is tight and there is a possibility of it rotating into the toolholder - the shaper wont care but it would do a lot of damage somewhere!
And finally the video of it chewing its way through a delrin blank:
At the moment this is as far as it's been tested. I have an aluminium blank and an en3 steel one lined up next, but this gear has already been ruined by me not indexing properly (lots to think about to run this) and starting with the gear train on the banjo ****-backwards thanks to an error on my spreadsheet.
At the mo this is a 16DP 26 tooth gear being cut with a tool I originally ground to cut an acme thread (the 14.5degree angle each side was very obliging!) Just the tip needed grinding wider.
If anyone's interested I can go into much more detail (yawn), like for example the helicals are 14T to 11T to approximate 4/pi as that unpleasant little irrational number crops up in the maths. I'll post a vidoe of the ally and steel tests when I finally get there, but I've been working on it since the beginning of November, and it all takes time.
Rob
I wasn't keen on the wire and cylinder technique as it looks like a) it could slip and ruin the work and b) you need a different cylinder for every pitch diameter gear you cut. The idea that Paul came up with of replacing the wire and cylinder with a lever was better but a) it doesn't quite generate an involute (not far out though) and b) the mechanism for setting the rotation of the gear blank as it travelled didn't look that precise.
For the uninitiated, the idea is to use a simple HSS cutter with a shape matching a rack tooth (straight sides) to cut involute gear teeth. I can't be the only one who detests spending £25+ for a cutter each time I need to make a gear size I haven't done before.
The light bulb moment came when I was messing about with the vertical feed mechanism that I'd made for my Elliott 10M. The same mechanism is used to traverse the table side to side, but the gear wheel is moved from the vertical feed screw to the cross slide screw:
As the aim is to rotate the workpiece exactly proportional to the linear travel, I figured I could tap off the cross slide gear with another shaft as that achieves this. The next problem is to hold the work, index it but also rotate it as the table traverses. Well, I have an Elliott universal dividing head which achieves just that! So all that was needed was an attachment to join the cross slide gear to the dividing head with a gear train with the right gear ratio. I designed it initially on a 3D package (Designspark) because its quicker and cheaper than making mistakes in metal. One over-riding aim was that there should be no permanent mods to the shaper, no drill holes for example, just clamp on:
Gears here are just drawn as discs. The gear on the right picks up the drive from the cross slide gear. It has an internal spline to drive a shaft that runs side to side. The left hand end of this shaft is bolted to the table and moves left and right with it, hence the spline. A pair of 45 degree helical gears turn the drive through 90 degrees onto another shaft bolted to the T slot in the side of the table. The end of this second shaft will carry the first gear in a train that drives the dividing head universal drive spindle.
After a couple of months of cutting metal it's finally done and here it is bit by bit going on the shaper.
Pick-up gear assembly, clamped to the ways:
Splined feed shaft before its bracket is bolted to the table:
This is then bolted on so everything runs smoothly:
The next shaft has a bracket with slots for T nuts and studs so the mesh of the helical gears can be adjusted:
And here it is bolted to the table:
Finally the gear train is fitted to the dividing head banjo to join everything up:
Sorry about the blurred images. Note the spindle locking lever has been removed from the dividing head and replaced with a bolt to keep the chips out. Clearance is tight and there is a possibility of it rotating into the toolholder - the shaper wont care but it would do a lot of damage somewhere!
And finally the video of it chewing its way through a delrin blank:
At the moment this is as far as it's been tested. I have an aluminium blank and an en3 steel one lined up next, but this gear has already been ruined by me not indexing properly (lots to think about to run this) and starting with the gear train on the banjo ****-backwards thanks to an error on my spreadsheet.
At the mo this is a 16DP 26 tooth gear being cut with a tool I originally ground to cut an acme thread (the 14.5degree angle each side was very obliging!) Just the tip needed grinding wider.
If anyone's interested I can go into much more detail (yawn), like for example the helicals are 14T to 11T to approximate 4/pi as that unpleasant little irrational number crops up in the maths. I'll post a vidoe of the ally and steel tests when I finally get there, but I've been working on it since the beginning of November, and it all takes time.
Rob
