Dog Clutch disaster…

[Iain DownsParticipant @iaindowns78295]

I’m in the process of making a dog clutch for the next iteration of my mill power feed.  However, it’s come out wrong and I can’t work out why.

This is the current state of play…

The 3d printed item on the right is (more or less) what it should look like.  The metal is the wrong’un.

You will see that it doesn’t fit together.  According to the design, the lands should be about 6.85 mm at the circumference and the gaps should be 7.05.  giving about 0.2mm for breathing room.

Instead they are roughly the opposite way round.

The diagram below indicates how this was meant to work

To make it I took a piece of 40mm bar and put it in 4 jaw vice on a rotary table.  The rotary table was zeroed from the table against the spindle (indicator on rotary table) and then the bar was trued up in the 4 jaw against the spindle (indicator on spindle).

There were a few intermediate steps to work up the final cuts, but in essence the table was moved 1.9 mm to the left and a series of 9 x 5mm deep cuts were made with a 4mm carbide cutter each 40 degrees off from the previous one.

This should have resulted in each cut been 0.1mm left of centre as per the drawing.  The first cut started at the bottom and went up.  The fourth cut started near the top and came down.  The full lines are the left of the cut and the dotted the right of each cut.  This shows only the two cuts which formed the first land and gap of course.

You will see that the land (at the bottom) should be 6.85mm and the gap at the top 7.05 not the reverse which is what I seem to have.

Both parts of the clutch have the same problem so it’s (probably?) not a set up error.

I can’t work out how I mucked this up, so if anyone has any thoughts I’d be grateful.  I’m loathe to cut more metal without knowing what I’ve done wrong!

Many thanks

 

Iain

[JasonBModerator @jasonb]

From your description it seems to work, I got 6.78 for the solid and 7.18 for the cut at the outer edge and the inner has similar clearance. Can only assume the 1.9mm offset was set wrongly or cutter cutting undersize.

Edit your sketch does not actually show the 4mm slots being offset by 0.1mm as the lines all pass through the middle, if one side is meant to be tangental to the small central circle than that should be 0.2mm dia to give the 1.9mm offset to the ctr line of the cut.

[David JuppParticipant @davidjupp51506]

But your drawing doesn’t show cuts as being ‘left of centre’.

Not sure where the dimension values come from – I can’t match them either with cuts touching the diameter, or 0.1mm offset from it.

Surely if table were moved to the left, cuts would be offset to the right – though I still can’t match your numbers.

[Iain DownsParticipant @iaindowns78295]

Jason – you’re right it should be 0.02mm diameter whereas I put 0.01 in.

The circle doesn’t show well as it’s tiny.  Here’s a close up with the offset circle corrected

That does change the theoretical land to 7.15 and the gap to 6.75, which means my results are  even worse.

David.  The cuts are left of centre by a tiny amount (0.01mm out of a 20mm radius).  The diagram above shows the working.

What I did was to centre a 4mm cutter over the part and then the table it 1.9mm to the left, resulting in a cut which would end up 0.1mm to the left of centre.

Not sure what I can do except try again.  I will probably go and make sure that the rotary table is correctly centred.  That could explain the issues if I was some few tenths of a mm over.

 

Iain

[JasonBModerator @jasonb]

This is how it is

[JasonBModerator @jasonb]

Your drawing showed 0.1mm now changed to 0.2mm but you text is now saying 0.01 and 0.02. If you only moved it 0.01 instead of your intended 0.1 that could be the problem.

[David JuppParticipant @davidjupp51506]

But if you move your table to the LEFT, the cuts move to the RIGHT, which does explain why teeth are larger than gaps.

I checked that in CAD.

[Iain DownsParticipant @iaindowns78295]

Jason, the 0.01 was a mistake.  As per the description above I moved the table 1.9mm to the left of the centre of a 4mm cutter which means that there is still .1mm over the centre line.

David – same applies.  I’m moving the table 1.9mm to the left but the centre of the 4mm cutter was 2mm into the left to start with so 0.1mm remains across the centre line.

I think the only thing I can do is to have another go setting everything up from scratch and see if it works – or possibly see if I can set up the current pieces and take a bit more off, though that sounds like a few too many degrees of freedom (left, right and rotational) to succeed.

Thanks for the feedback

Iain.

 

[JasonBModerator @jasonb]

Works OK for me David, Grey and red are one each offest to left and right, Grey/Grey posted earlier was offset table to left, red/red is table offset to right

[David JuppParticipant @davidjupp51506]

If cut is displaced 0.1mm to the right of centre, then the following arises (and note the very small ‘pip’ in the centre).  Teeth are hatched.

[David JuppParticipant @davidjupp51506]

I obviously didn’t read the alignment / traverse bit carefully enough.

Was table locked after the 1.9mm traverse?  Could the traverse somehow have been 2.1mm rather than 1.9mm?

[JasonBModerator @jasonb]

Pip is not a problem as coupling will have a hole in it. However as the cutter should still overlap the ctr by 0.1mm be it on the right or left how can it leave a pip?

Proof is in the Pudding

First cut with table moved 1.9mm to the left which puts cutter to the right of ctr part marked R

remaining 8 cuts to complete one part.

To show right of left makes no difference the second one was done moving table to the right which puts cutter 1.9mm to the left marked L

After doing teh other 8 cuts try the two together and they fit

Just to further proove that left/right doe snot matter, imagine the green lines are a cut to the left

Now have I done a cut to the right in this last photo or am I simply looking at the part the other way round. I know it is simply being viewed the other way round.

 

 

[David JuppParticipant @davidjupp51506]

The pip can only be formed if the cutter doesn’t overlap the centre line – so if the traverse was larger than intended or cutter smaller than intended or table moved due to cutting forces….

[Iain DownsParticipant @iaindowns78295]

It’s possible that I’ve uncovered the root cause of the issue – or at least a contributor.

I decided to strip the setup back to basics, take everything off the mill and check the tram and so on.

Underneath are 3 photos.  The first is with the indicator to the right of the table, the second to the left where you can also an indicator set against the spindle.

The third is the same setup as the 2nd but the spindle is locked.  Both indicators measure in microns.  I apologise for my photography.  I appear to have got the most important bits out of focus…

Right

Left

Left with Lock

From this the tram is out by about 50 microns (2 thou) across 300 mm, which is not ideal, but nothing I’m going to cry about.  I may try and improve it.

However the base of the spindle moves 80 microns to the right when the spindle lock is tightened, which I am concerned about.  There is also some indication that the tram gets quite a bit worse when the spindle is locked.

That’s only 0.08mm so it theory that’s not enough to distort the layout quite enough (it would need 0.2mm to reverse the 0.1mm offset), but it’s not helping.

Before I try cutting metal again I thought I would seek advice here.  Is the spindle movement abnormal and is there anything I can do about it (Mill is a VM32L from Amadeal)?

 

Iain

[JasonBModerator @jasonb]

Simple way to deal with it would be to setup your rotary table, chuck, workpiece and locate ctr of workpiece all with the quill locked. Also make the cut with it locked.

Your photos also show a fair amount of quill extended which won’t be helping, try and keep the quill in as much as possible, should really only need extending for drilling or a short amount if touching off on a horz surface.

[Iain DownsParticipant @iaindowns78295]

Fair point, Jason.  Makes sense.

The extension was the least I could use for the tramming.  there’s a fair amount of air between the spindle and the bed with the head is at the lowest setting.  I try and keep the stickout as short as possible.

 

Iain

[Tony Pratt 1Participant @tonypratt1]

Other people are dealing with the CAD issues but out of tram by 50 microns (2 thou) across 300 mm in reality is going to make no difference to your dog clutch, I’m presuming there is clearance between the teeth as size for size teeth and gaps will never go together.

Tony

[JasonBModerator @jasonb]

You can see the clearance in the second image of my first reply

The fact that Ian’s 3D printed one from the same geometry also went together shows that bit is right and my metal one also went together so must have been an issue when Ian’s was machined.

[Howard LewisParticipant @howardlewis46836]

Maybe I am simple minded.

If a 4mm cutter is set on the centreline, and then offdset to one side by 1.9 mm, the cutting edge will be 3.9mm from the cetreline. Having taken the cut and then offset the cutter by 1.9 mm to the otherside of the centreline (Not forgetting to take out the backlash, before making the second offset,) you should finish up with a slot that is 7.8 mm wide.

To my tiny mind, to machine the slots, the offset should be 2mm, each side, to produce a slot 8 mm wide.

(cutter radius 2 +  offset 2 = 4 each side = 8 slot width)

To machine the teeth, the 4 mm cutter would need to be offset 3.9 mm, each side of the centreline, to produce a tooth that is 7.8 mm wide.

Care would need to be taken, not to climb mill, in case that drags the cutter into the work.

Or have i got that totally wrong?

Howard

[JasonBModerator @jasonb]

Yes. Needs a different mind set

There a 9 teeth, each pass of the cutter machines the left side of one and the right side of the other, by the time you have done all 9 cuts the gap at the periphery will be approx 7.2mm and the teeth approx 6.8mm. I have marked on the sketch below how a cut right across the work cuts the left side of one tooth and the right side of another.

If I then repeat the cut represented by the red rectangle 8 more times this is what it looks like, the two 4mm cuts overlap at the periphery to leave the 7.2mm gaps and the green areas remain as the raised teeth

[duncan webster 1Participant @duncanwebster1]

Having the side of the cutter offset from the centreline results in the face being not radial. When you then mate with the other dog you’ll get line contact rather than face. I think you can overcome this by setting the side of the cutter on centre, cutting through one side, but rotating the job by a degree or so (anticlockwise in Jason’s sketch) before ploughing on through the other side.

[Iain DownsParticipant @iaindowns78295]

Howard the cutter is diameter 4mm so radius 2.0.  So when offset by 1.9mm the cutter is 0.1mm from the centre.

Duncan.  I suspect you are right, but I don’t think it’s going to make a lot of difference in my application.

 

Iain

[Tim StevensParticipant @timstevens64731]

An afterthought:

It would save time winding handles if the workpiece had a larger hole in the recess. The points produced on the inner ends of each tooth add nothing to the torque-resistance of the clutch, so cut them off before you start.

And whenever you find that teeth are too tight, remove metal from the bronze and leave it on the plastic.

Cheers, tim

[JasonBModerator @jasonb]

I suspect the “points” are only there because  David and myself did not add a central recess.

You still need to wind the work from one side to the other it just would not be cutting all the way.

[Iain DownsParticipant @iaindowns78295]

Well.

Yesterday, I had another go at making the part.  I followed the same steps as before and got this…

As you can see the metal bit now fits the plastic bit as it should.

I don’t really know what went wrong the first time.  I should say that I centred the rotary table at a different time than the actual work took place at.  My working theory is that I must have nudged the X axis whilst the DRO/Mill was off.  Alternatively I majorly cocked up the centring of the rotary table (by around 0.2mm).

Tedious though it is to redo the parts, at least I will have them!

 

David – there is a central recess made on the lathe.  This picture shows the (defective) parts split apart…

Thanks to all for input..

 

Iain

[Author]

Posts