Gear cutters for clock wheels

[Jouke van der VeenParticipant @joukevanderveen72935]

Sorry for my double post, I did not see my first post, therefore.

Huub, I faced in the 4-jaw. If I remember well, with the backface against the lower steps of the jaws, not against the chuck itself. So that I could machine the whole frontface. That was wrong.

The Emco Compact 5 has a MT2 in the spindle. But with at the entrance a cylindrical hole with the result that many standard tapers do not fit. You need a special taper to fit. I found this information somewhere on this forum. Otherwise I would have used a taper with a machineable front for the excentric mandrel to cut the teeth.
Now I clamp a mandrel in the 4-jaw and support it with the lunette to minimise pushing it out of center during cutting.

I think I will go for the faced super glue chuck to improve parallelity. Perhaps with a central button to exclude detaching of the blank. I can not use the revolving center for pressing because its diameter is just a bit smaller than the center hole in the blank. We will see.

[John HinkleyParticipant @johnhinkley26699]

Forgive my brief intrusion into your thread, but, as a complete numptie with regard to clockmaking, it seems to me that what you are looking for is covered by Clickspring on YouTube here, for making the buttons and here, to make the cutter.

Apologies if I've misunderstood, but my attention was drawn by the similarity to the drawing supplied by John P on page 1 of the thread.

John

[Jouke van der VeenParticipant @joukevanderveen72935]

Concerning the bearing in the lunette. I did not like the sliding of the lunette pins on the mandrel. Always at the same circle. So, I thought: let’s have a try. It is a bearing with ceramic balls normally used in bicycle bottom brackets.It is pressed and glued on the 24mm mandrel and has a 37mm outside diameter. The lunette has an max opening of 40mm.

[Jouke van der VeenParticipant @joukevanderveen72935]

Starting with the mandrel according to Dean Williams’s design I asked him if it would be possible to introduce a center support there where the excentric cutter blank is clamped. Not so easy I think, but I got no answer. Therefore, I decided to use the lunette and added the bearing to it later on.

[Jouke van der VeenParticipant @joukevanderveen72935]

Hallo John H,

I watched several Clickspring videos but not these! Very nice but hat looks all to be very precise machining. I have to deal with a rather sloppy Emco Compact 5 milling cloumn.

I did not feel myself able to make a double button cutter. You have to drill two holes at a very precise distance from each other and the two small buttons must have precise diameters. I did not even try. My single button “ground ball”method seems to me much easier but I realise that my precision work comes at the end when cutting the teeth of the gear cutter itself.

Regards, Jouke

[HuubParticipant @huub]

Posted by Jouke van der Veen on 11/02/2022 09:02:20:

Huub, I faced in the 4-jaw. If I remember well, with the backface against the lower steps of the jaws, not against the chuck itself. So that I could machine the whole frontface. That was wrong.
That should not make such a difference in thickness (if nothing is worn out) and absolutely a good way to do it.

The Emco Compact 5 has a MT2 in the spindle. But with at the entrance a cylindrical hole with the result that many standard tapers do not fit.
Maybe a standard taper is "to short" but if you can machine the front and add a (threaded) bar, it will do the job.

I think I will go for the faced super glue chuck to improve parallelity.
You could also glue some spacers on the claws and machine them. This should give a nice parallel reference. If you do, clamp the mandrel to compensate for any play in the claws.

You can also just use one cutting edge, that would also solve this problem.

[John HinkleyParticipant @johnhinkley26699]

Jouke,

Fair comment about your "sloppy" machine, but surely at some point you have to make precise movements to make the cutter, whether it's spacing the two buttons or moving a single button to the "other" side of a blank to make the cutter?

No DRO? Use a dial indicator.

John

[Jouke van der VeenParticipant @joukevanderveen72935]

Huub,

I think the flanks of teeth are parallel since they are cut in two parallel operations. Only the radii have a clearance since that operation is excentric. Am I wrong?

Concerning the MT2. An option would be two use a taper with drawing bar and M14x1 thread at its nose. I prepared a 13mm hole already in the excentric mandrel, still internal thread to be cut.

I do not understand your: “You can also use on cutting edge …”.

John,

My feeling is that the milling column is more sloppy than my lathe. At least in the way I am working with it. 😉. Yes, I have a dial indicator and some kind of a removeable DRO.

 

Edited By Jouke van der Veen on 11/02/2022 11:56:46

[HuubParticipant @huub]

Posted by Jouke van der Veen on 11/02/2022 11:30:49:

I think the flanks of teeth are parallel since they are cut in two parallel operations. Only the radii have a clearance since that operation is excentric. Am I wrong?
Yes you are wrong. Due to the eccentric turning, the total tool profile is shifted towards the centre. You will see that the tip of the gear cutter is flat, curved and tapered.
1 mm down the cutting edge, the cutter is moved more towards to the button (eccentric turning effect). This results in a smaller tip at this point and creates clearance.

I do not understand your: “You can also use on cutting edge …”.
If you have only 1 cutting tooth the actual width or flatness of the cutter blank doesn't matter.

My feeling is that the milling column is more sloppy than my lathe.
On all small desktop turning and milling machines, you have to lock all unused axis to improve rigidity.
When I mill using the X-axis, all other axis (Z, Y, Quill) are locked.

Check the play by locking all axis except 1 and than rock this axis to feel how much play is has. It should be unnoticeable.

On my small lathe, the set screws for adjusting the gibs are not locked. I adjust them each time I use the lathe. Once you get a feeling for it, it only takes a few seconds to setup.
The drawback is that when you turn a lot of parts, these set screws some times loosens.

[HuubParticipant @huub]

Posted by Jouke van der Veen on 11/02/2022 11:30:49:

I think the flanks of teeth are parallel since they are cut in two parallel operations. Only the radii have a clearance since that operation is excentric. Am I wrong?

Youke, you are right. The part of the gear cutter tip that has parallel sides doesn't has clearance.

I am sorry if I confused you.

Regards, Huub

 

p.s I can't edit prevous post

 

 

Edited By Huub Buis on 11/02/2022 13:45:47

[Jouke van der VeenParticipant @joukevanderveen72935]

Huub, you left me in a lot of confusion 😉! But you are forgiven. I learned a lot from you on the other items. I am sure I am keeping progress with support from this forum.

[Jouke van der VeenParticipant @joukevanderveen72935]

But…

If the form tool is given a very small non-zero curvature behind the “frontradius” then you will already introduce a clearance for the “flanks” of the gear cutter.

This can not be achieved with a button tool but with a tool with one varying radius. Perhaps with some kind of tangential cutter with a round toolbit, as shown in my other album. A tangential tool gives a varying radius which can be analytically calculated. You choose an angle and the radius is known all along the edge of the elliptical cross-section.

However, as soon as the tangential tool cuts further than half its long elliptical axis then you get parallel flanks again. Meaning that you can not cut high enough teeth? Not practical to my opinion.

Edited By Jouke van der Veen on 11/02/2022 21:18:37

[HuubParticipant @huub]

If the form tool is given a very small non-zero curvature behind the “frontradius” then you will already introduce a clearance for the “flanks” of the gear cutter.
Only if this flank is wider than the tip, otherwise the tip will remove the material.

Perhaps with some kind of tangential cutter with a round toolbit.

However, as soon as the tangential tool cuts further than half its long elliptical axis then you get parallel flanks again.
I aggree

Meaning that you can not cut high enough teeth?
The Clock tooth has a pretty long parallel part. Basically, this part will be cut by the tip of the cutter. The tip has clearance so that is not really a problem. As long as the tip is sharp, cutting wil not be a problem.

I think even without clearance these cutters will do the job because all cutters made using an eccentric mount have the same problem and are still used with good results by a lot of clock makers.

The cutters I make for broaching a keyhole on the lathe, by design, have only clearance at the tip, not at the parallel sides and therefore they work pretty good for this task.

In the video I am making, I will show an easy way to make these clock gear cutters, having clearance on all cutting edges, simply by manual milling. I am very anxious to see the difference between this type of cutter and the once made the old way.

 

Edited By Huub on 11/02/2022 22:24:07

[John HaineParticipant @johnhaine32865]

If you use superglue as a "chuck" it's easiest to remove by either boiling the glued assembly in water for a few minutes, or heating in an oven. Boiling seems to work best, and any glue comes off the surfaces in thin sheets. I often stick down thin plates for profile milling by applying masking tape to each surface then glueing together with superglue, then boiling after cutting. Works fine, no glue left.

[Jouke van der VeenParticipant @joukevanderveen72935]

Gentlemen, thank you.

Huub, what do you think of my elliptical calculations? Perhaps, you did the same for milling your gear cutters with end mills?

[HuubParticipant @huub]

Youke,

what do you think of my elliptical calculations?
I haven't seen your calculations. Mine (involute gears) are based on the Ivan Law gear cutting book, Mikes work shop and probably some more information I have gathered from the Internet.
I will send you my spreadsheet by PM. The documentation is in Dutch but that won't be a problem for you. Once I have translated it to English, I will place it on Git so that all can benefit.

I check my calculation by drawing the involute gear in cad and then add my own tooth profile. That makes it easy to see the differences.

hessmer spurgear builder

Mikes workshop

Ivan law gears and gear cutting

[Jouke van der VeenParticipant @joukevanderveen72935]

Huub,

You will find it in my other album “Tangential cutter”. I added a photo of a sheet with handwritten text in Dutch about radius calculations of an elliptical cross-section of a round toolbit. I made this about 8 year ago and it should be improved by explaining wat the intentions were, and in English of course.

Regards, Jouke

[HuubParticipant @huub]

Youke,

I have found your calculations and this is an interesting approach.

My first thoughts are that you want to mill an ellipse by placing the mill (button) at an angle. This way you can approach the desired tooth form more accurate than by using a mill or button the common way.

Question:

How do you know the dimensions of the ellipse that best fits the required shape.

Making such a cutter is not so simple but doable:

If you rotate the head of the mill, you will lose your position and setting the right position is very difficult as I experienced some time ago.

Placing one button at the right angle is doable. Placing the other button at the same angle is also not a problem, just reverse the tool 180°. Getting the right distance between the two buttons is very difficult, but using 2 separate tools is not a problem. So I think it is doable.
You could also mill one side of the two seperate tools until the distance between the buttons is OK and than bolt them together.

The benefit:

I expect that the extra accuracy you gain by making an ellipse will be lost once you cut different tooth sizes (12,13) using the same gear cutter. So the real gain will be only for the first gear in the range. Because making a cutter on the CNC lathe is so easy, I make a cutter specific for each tooth count.

If I look at the tolerances the cutters a made (D+f = 2.16) I expect that in the end the gain is little. The tolerances on clock gears are even greater.

How to check the result:

A way to check the real benefit is to design the gears in CAD made in different ways and check the differences.

Regards, Huub

[Jouke van der VeenParticipant @joukevanderveen72935]

Huub,

I had a quick look at your spreadsheet. Nice work. It needs more attention from my side.

Years ago I thought it was not such a good idea to make a cutter by grinding a radius by hand and then to check this radius against a known radius in a thin plate. The latter is made by drilling a hole in a plate and cutting a quarter away. Many clockwheel makers made their radius tools in that way I believe.

The idea was to make a well defined radius by cutting a round toolbit under a known angle. The part of the ellips with the highest curvature (direction of long axis) could be used as a cutting edge. The radius of that edge is surprisingly constant over many degrees. So you do not use a round button for cutting but the elliptical shape with a rather constant radius.

You start with a horizontal tool holder in which a round toolbit is clamped under a known angle from the vertical. You can always use a toolbit with the constant diameter D.The top part of the toolbit is machined down to a smaller diameter d which will become the short diameter of the ellips you need. This ellips is made by cutting the top of the toolbit parallel to the tool holder.
The curvature (or 1/r) of the ellips in the direction of the long axis is now fully determined by diameter d and angle of toolbit from the vertical. As said above: radius r is rather constant along a considerable angle of the “tip” of the ellips.

Depending on the radius r you need you machine a toolbit down to a corresponding diameter d.

The elliptical cutter was intended to be used feeding perpendiculer to the gear cutter blank axis. I agree that moving away from this direction may give undefined shapes.

A lot of words about a shape tool I never really used and in principle the same as a tangential cutter with a round toolbit I was not familiar with at that time. So “invented” by myself 😉.

Jouke

[Jouke van der VeenParticipant @joukevanderveen72935]

Gentlemen,

I tried to improve parallelism of a gear cutter blank with diameter 30mm having a thickness variation of 0.05mm by applying the super glue chuck method. It decreased to 0.03mm, so not so much improvement. I bonded the blank to an aluminium plate, faced after mounting in the independant 4-jaw, by applying 4 droplets of super glue (between the 4 small holes, I will add a picture to the album). The blank was quickly pressed against the chuck by means of a mandrel in the tail stock. After debonding in boiling water I found two thin disks of glue back on the aluminium. I measured thickness of these and found 0.09 and 0.08mm, respectively. Rather thick in my opinion. I wonder how an a constant and minimum thickness of the glue film can be achieved. I used Ever Build Stick 2 Super Glue GP Industrial Grade, set time 5-15s. I think it is so called medium viscosity.

Regards,

Jouke

[HuubParticipant @huub]

Jouke,

Gentlemen,
May women not reply

The blank was quickly pressed against the chuck by means of a mandrel in the tail stock.
If you do this, you have to use very little pressure because the raw blank it self is not flat. Have you removed the burs from making the mounting hole? Have you removed the burs from making the chamfer?

It decreased to 0.03mm, so not so much improvement.
From 5 to 3 is 40% better, you are on the right way

My procudure for this would be:

• Face the blank, drill the hole (- 0.1 mm), chamfer the blank, chamfer the hole, do a last facing pass.
• Turn the blank to do the other side
• Face the blank, chamfer the blank, chamfer the hole, do a last facing pass.
• Ream or drill the hole to remove the burs from chamfering and get right the diameter
• Give the blank on both sides a few manual grinding passes using an old worn 400 grid diamond grinding plate (1 mm thick)

I have checked a M20 washer 40 mm diameter, 4.83 mm thick, that I faced using the above procedure a few months ago. Thickness spread 4.82..4.83 mm. This was done using a taper in the spindle bore as I explained before.

• Have you faced both sides of the blank using the super glue method
• Did you use a HSS tool or Insert (Carbide)
• RPM
• Cutting depth
• Thickness of the blank
• Has the blank a really smooth finish
• Have you checked the cross slide play. Any play will lead to thickness variations.

Regards, Huub

Edited By Huub on 14/02/2022 22:26:50

[Jouke van der VeenParticipant @joukevanderveen72935]

Huub,

Your first point: I realised it too late to correct it. I apologize for it.

Of course there was some 40% improvement and you can improve it to below zero 😉 but you can make it worse.

The sequence of operations you describe should be the one I should apply next time.

I am not sure if a low pressure during bonding to the face place would work for me. The small blank is in vertical position when it is placed in position and it should in center I suppose. Perhaps I should use a thinner glue.

I used a 4.5mm thick blank that had already the final central hole and the 4 3mm holes. Burrs were removed from all edges. I did not polish before facing. One side was faced in glued position. I still can do the other side but I can also start again from the beginning on this blank.

Carbide insert with radius .6mm for this facing (the glued side was with HSS).

Rpm was 550.

Cutting depth appr. 0.05mm. Not a really smooth surface.

I did not lock the saddle and dit not correct cross slide for play.

So many things to improve. I could repeat an improved procedure on the same blank but could also start from the real beginning by cutting new blanks from silversteel bar stock.

Regards,

Jouke

 

Edited By Jouke van der Veen on 15/02/2022 09:30:20

Edited By Jouke van der Veen on 15/02/2022 09:58:48

[Jouke van der VeenParticipant @joukevanderveen72935]

Of course there was some 40% improvement and you can not improve it to below zero 😉 but you can make it worse.

That was what I meant.

Jouke

[HuubParticipant @huub]

Jouke,

I still can do the other side but I can also start again from the beginning on this blank.
Just use the same blank until you get it right.

Not a really smooth surface.
If you are targetting the 0.01 mm range the surface should be really smooth. That is the first to achieve. Start by using a softer steel to get a feel for it. Check the tip of the tool using a 10X magnifier. Some damage is hard to see by eye.
You should play with the RPM, cutting depth and feed. Depending on the insert, you can't do a shallow cut. Depending on the rigidity of the lathe, you can't do a thicker cut. If there is an unbalance in the chuck (or bad bearings), you can't run high RPM.
Listen to the sound the cutting makes. You can hear if the cutting is smooth or rough!!!
Make notes of the metal type, turning conditions, tool data and the result. Buildup experience!!!

I did not polish before facing
I just give it 2 passes in 2 directions to remove any burs or high spots (from impurity's).

Rpm was 550.
That is pretty slow for carbide. I have a Vc (Velocity at the outer diameter) of 180 mm/min!!!

My test on the small lathe
I faced an other 40 mm washer on the small lathe (HB300) using the glue method (cheap Super glue from the Action store). I faced a 20 mm register to loosely centre the washer. I gently pressed the washer against the mandrell using my fingers.

I faced both sides CNC using a carbide insert, nose radius 0.4 mm, RPM 1400, Feed 60 mm/min, 0.05 mm cutting depth (final pass).
Spread 4.045 .. 4.055 mm, the same as on the bigger (HBM BF290)

Then I faced one side again, but now manual (on the same lathe)
Spread 3.925..3.920

So I can repeat the results on both lathes manual and CNC.

I boiled the glued washer for 6 minutes (microwave) but couldn't separate the washer from the mandrel. It took a bit force (screw driver) to get is lose. So I couldn't measure the thickness of the glue but I estimate around (0.02..0.05 mm).

Normally I heat up the mandrel using a torch!!!

 

 

Edited By Huub on 15/02/2022 13:07:36

Edited By Huub on 15/02/2022 13:08:06

[Jouke van der VeenParticipant @joukevanderveen72935]

A lot of improvement here!

I started with polishing the “faced” gear blank (see earlier) on a 800 diamond coated steel disk. Of course this does not improve parallelism (0.03mm) but does show all features on the surface.

Thereafter I faced the other side of the blank by bonding it to a freshly faced alu disk in the chuck. Again I applied 4 droplets of super glue to the blank surface, between the 4 small holes. Then I put the blank on a 8mm mandrel in the tailstock and pressed it against the chuck. Facing was done in two steps of approx 0.05mm with a carbide insert with radius .4mm (not .6mm) with 950rpm.

After appr. 15min in boiling water the blank came off. I measured thickness between the 4 holes a close as possible to the outer edge, with following results: 4.30(3); 4.30(5); 4.30(7) and 4.30(5)mm.

To my opinion this result is more than good enough to proceed to the next operation: profiling of the blank on the excentric mandrel. When starting with new blanks I will follow a more consequent sequence of operations, as Huib Buis proposes.

The glue remainings are still on the alu disk. I tried to remove it for thickness measurents (as I did earlier) but I could not peel it off. Perhaps I should leave it longer in water but that may increase thickness by swelling. It should be dried before measuring. It is for sure that this thickness must be very constant, otherwise the above reported parallelism is impossible.

With thanks to the forum and Huub especially.

Regards, Jouke

[Author]

Posts