45 degree worm M2.5

[philbur]

Hi Art, I'm not sure what you mean by project file? The gth file is not on the list of acceptable files for attachment.

So here's the project summary and the gcode file:

Phil :)

PS: The error (undercut) gets less as the number of teeth increase but I don't think the error is ever zero

[ArtF]

Phil:

  >>gets less as the number of teeth increase but I don't think the error is ever zero

  It doesnt. I think I found the error. Its in the tooth profile stored for tool shape. I use that shape to actually plan the
cut and Im wrong on the helicals. The tooth shape IS correct, from a transverse point of view, but not for the
machining planning in 4th axis. The problem is the shape needs to be identical to an equivalent non helical shape.

  I will fix this and put out a fix. Thanks for pointing this out, a the tooth count increases youd notice it less and less
but its always there as an error. Once I fix the machining tooth shape, I think it should be OK..

Art

[ArtF]

By the way, Ive fixed it so that gth files may now be put in threads..

Art

[ArtF]

Phil:

  Ive made a fix...

  The problem is that to do toolpaths, and in order to allow 4th axis epicycloidals and such, I changed the way I do things. I now take
  the tooth shape which is shown on DXF's and such for tool grinding..and I wrote a routine that traces it with a tool. The theory was that this would
  work for multiple tooth shapes and was self correcting in that it wont allow a bad cut as nothing can exceed the shape of the tooth shape used
  to make the tooth. The problem with that, is that helical's in particular have their pitch diameter increased by the secant of the helical angle.
  This means the tooth shape was never correct as shown, the shape shown was what youd see looking at a helical directly sideways( tranversely),
  but the shape you NEED was the shape looking at the gear at helical angle ( perpendicular to gear normal direction).
 
    So..what Ive done is change the tooth tool shape, ( not the gear..its the same..). Ive done this by changeing the tooth count used to make
  the tools shape. The equivalent spur tooth shape is given by a spur gear with more teeth. ( Toothcount of equivalent spur is the helical tooth count
  divided by the cosine of the helical angle cubed. ). This fixes your rooting trouble..BUT Ive noticed an inconsistancy in the output, the model doesnt
  match the toolpath. This tells me somethign is wrong. ( The beauty of math is that it should match when looked at from any angle. ). Something
  has gotten ugly in there. Im suspecting I have inadvertantly not made the chordal helical tooth arc correctly, so it "may be" your teeth will be
  slightly too thick. Im investigating, but Id be interested in whatever you find in testing. This is one of those complex situations where I know either
  the model OR the toolpath is slightly incorrect, but Im unsure exactly which.. I suspect its the toolpath being slightly thinner than required, but
  Ill make sure before I change any more. A second issue with this is that a helical really isnt a worm mathmatically. Helicals are speced to be 12
  teeth  or larger mathmatically, though this is often ignored, so the fewer teeth you have the toth equivalency count formula's blend off a bit. 
  Undercut in a helical is supposed to be verboten, though I do allow it in GT and GM. SO a 4 tooth is pushing it ..in math terms, personally I like 
  seeing it though. 
  :)
 
  The new version with this fix is now online.. and I continue to investigate.,..
 
  Art
 

[philbur]

 
  The new version with this fix is now online.. and I continue to investigate.,..
 

Hi Art, I'm glad to hear you found the problem, Which one has the fix:

- Current Release 4.67 (Click Here)

- Current Development 4.xx (Click Here)

Phil :)

PS: are the version numbers correct ...... 4.67 and 4.xx?

[BobL]

Phil;

It would be GT in current development, that is Ver4.XX


Cheers
Bob

[philbur]

Thanks.

Phil:)

[ArtF]

Hi Phil:

As Bob said... The numbers only rarely match up, Im terrible at numbering and I turn out fixes or adjustments too frequenctly.
Whenever a release is very stable I update the top one, and then always update the bottom one.. :)

Art

[philbur]

Hi Art, I've just been looking at the gcode generated for an involute helical gear using an involute saw (I assume this is a B&S type involute cutter). It looks like  a series of slots cut by passing the cutter across the gear blank with the blank angled at the helix angle. If this is the case then I do not think that it will generate the correct form. The error is easiest to visualise if you consider a large helix angle and a very wide gear blank. As the cutter passes across the gear blank it will have reduce depth of contact either side of the centre point of the blank. If the blank is wide enough the cutter will completely lose contact at the two edges of the blank. To work correctly I think maybe the cutter needs to be angled at the helix angle and the gear blank rotated as the cutter passes across the blank.

Phil :)

[John S]

I agree with Phil.

Not looked at this module yet but what he says makes sense.

[ArtF]

Phil:

Thx, Ill take a look at it. Its my plan to totally recode the Gcode section, at the moment its all one class, Im going to subdivide it so
each can work independently. I suspect I screwed up whatever was wrong with saw even more with the new tool profile, an angled
axis woudl require the previous transverse I think as a planning shape.. ( Perhaps not, Ill be looking into both..

Art

[philbur]

It seems that the involute bit (I assume this is single tooth involute cutter) cuts the helical gear correctly. This is strange because I think the cutting process for the B&S cutter should be the same, the only difference is the number of cutting teeth!!!

Phil :)

PS: I see that the cutter type is not as I thought so ignore the above.

[ArtF]

Hi Phil:

  We'll look at each routine one by one in near future.. :), I do think theres a thickness limit
for that type of helical cutting due to cylidrical wrap on larger widths... anyway..later..

  I have had to redact yesterdays changes, I jumped the gun. The tooth
profile which I thought had to be corrected..was correct. I should have known this as I cut
several helicals with it myself. I changed it to correct for profile , but that was wrong and
wasnt your trouble. I have uploaded the program with those changes removed.
  This happened because I duplicated your problem and could see the same shape..
surprisingly , when I redacted my changes and tried again, the paths were correct.
  Thats very weird, and always worries me. Hopefully you didnt waste time or material
on testing that version. I need now to try to reproduce your paths for that worm..
( well..long helix anyway.. :) )

  Ill let you know what I find and perhaps ask you to verify it. You should probably redownload..and see if
that path still looks screwy..sometimes a touch error in the compiler allows a modification to not compile till
I make another change, this may have happened here. I cant explain why the path looks OK to me today.

  Last night I tried it and the rooting and toothing matched the profile in the center of the gear. (This is one way
I verify, when viewed from various angles one can see the toolpaths touch the profile when dia of the tool is zero
in the toothing, and for rooting ( where 0 wont work), I enter a small diameter and see if the root paths just miss
the center profile). Last night they matched perfectly for some reason, yesterday early they matched yours..
something is afoot....now to find it..


Ill comment more later..

Art

[philbur]

Hopefully you didn't waste time or material on testing that version.

Hi Art unfortunately I just came back in the house from trying yesterdays revised version and yes it doesn't generate the correct form. I appreciate that using the helical form to produce a worm is probably stretching it a bit and I'm not even sure if it is the correct profile. A worm cut on a lathe is done with a straight sided "v" cutter!

However the worm allows for very large ratios in a relatively compact space and gives many possibilities.

Phil :)

[ArtF]

Phil:

  Todays version does appear to generate the correct form, at least differetn from your original test, but
Im not sure why its different. Redownload and see if the toolpath looks different.. I wouldnt cut anything till
we at least figure this out in theory. Although it may not be a proper worm, its my opinion that it shouldnt overcut the way it did. Your fine pathing looked and seemed to work great, so its just the rooting that something seems screwy on... and while Helicals shouldnt by spec have undercuts, I fail to see why GT
couldnt make them anyway...

Art