Problem with 4th axis toolpath

[ArtF]

Greg:

 hey, no pain at all. If it wernt for you Id have missed it. I didnt try differing pressure angles, they shouldnt in my mind affect the result.. but I better check it now.. Ill let you know what I find.

Art

[ArtF]

Greg:

  Hmm, I just did a mod5, 300 tooth, and the result at the end of the tooth where any error should be highest is shown here..
  Not sure why yours woudl be different but Im still checking. .

[Greolt]

Art

Here are my results for the mod5 300 tooth example at 15 and 25 degrees pressure angle

Also note the 10 degree example, which my result differs from yours.  This suggests to me I am doing something wrong.  Take this into account. 

In all cases the Y position is correct to align with tangency on the last cut.  It is the A position where error is showing.

And as a minor note, the tool looks like it needs to go lower.

Greg

[ArtF]

Hi Greg:

  Heres what I get on 300tooth Mod 5 , Pressure 25 ..

and on a mod5 300 tooth pressure angle 20 ..


  It may be we're checking things different ly. ( Must be really. :-)

  Here's what I do, but I use vectrix. I load the 2d dxf, then I draw a polyline from 0,0 to the angle of 90 - angle specified in that final tooth.
I subtract it from 90 just so it will operate at the top for ease of doing this.
I then take the base circle diameter and divide by 2 and call it y, I then start a new polyline at a postion of 0,y and draw for a length in the x
direction as indicated by the final tooth Gcode.

  In both cases here the tool in final tooth position is exactly 1/2 tool diamter from the tooth surface and at the tip of the tooth. Whcih is where it should be by my calculati on. I cant cut a 300 tooth either, but this seems to show no loss occuring even over a very large gear surface.

  Let me know if you have any thoughts as to where you  ( or I ) am going wrong. I suspect its OK, otherwise my calculati on should be off, but Im showing it dead on in every test now.

[ArtF]

Greg:

 I shoudl mention of course, that after drawing the angle line, I rotate till any tooth is in dead center of the angled line ( but I see you do the same.) .
 Are you subtracti ng the angle from 90 as I do?

Thx
Art

[ArtF]

Greg:

  Oh, one other point, as to the tool going lower, this is easy to check. The tool mathmatic ally MUST always be to the depth of the base radius, so if you look to the outside diameter and subtract from that the base diameter and divide the result by 2, you will have the Z coordinat e of the tool during an involute shaving operation .

  In the case of a 300 tooth mod 5 pressure angle 25 degree, we have 1510 - 1359.46 = 150.54 / 2 = 75.27 ( approxima tely) and the Gcode shows a Z of "g1Z -75.26" ..

  SO the tool depth at least must be correct. I suspect we simply have an error in test proceedur e. ( Be fun to find out if its you or me though.. :-)  ) , and no, I wouldnt take any bets..

Art
 

[Greolt]

OK I figured out my mistake.  

"Here's what I do, but I use vectrix. I load the 2d dxf, then I draw a polyline from 0,0 to the angle of 90 - angle specified in that final tooth.
I subtract it from 90 just so it will operate at the top for ease of doing this."

Yes this part is exactly the same as I was doing.

Then I draw the tool to scale  at TDC (zero A, Y and Z), then move this tool the value of the last cut Y and -Z

Now where I made my mistake was that I took all those values from the end of the tooth one section of code. (largest negative A value)

What I needed was those values from about middle of the code.  (largest positive A value)

Being a bit slow and dull, I did not realise these were different .  Still not sure why they are different, but they are.

Why are they different?  Three degrees different in the case of 25 degree pressure angle.

Greg

[ArtF]

Greg:

  Actually , its a good thing you did take them from the end. I was taking from the biginning, and a typo in the leftseg code was making the end wrong..

   I have two routines that control this, RightSege ment( ) and LeftSegme nt(), and while working with you on this, Ive been changing both routines, but I had a typo in LeftSegme nt() that made the differenc e. In actual fact they should be identical . The only real differenc e between left and right is direction after all since we start with 0,0 being center of the tooth.

  So your quite right, they should be the same, and they now are in 1.64.4 which is what is in the download now. Thanks for testing that, it may not have come to light real fast otherwise . I think if you test it now youll find it works in all pressure angles, and all tooth sizes.

  I really have to automate a test facility so I can verify its ok without all the futzing around .. I shoudl have noticed the differenc e, one side is simply the inverse of the other..
 
  Im still verifying all this, but the download is online just in case.

Art


   

[ArtF]

Hi Greg:

 Im finished verifying it, and it seems correct in all cases. Forward and back. Let me know if you see any error , sorry for how long it took to get it runnign properly, it was a fairt amount of redesign from the original code that was a bit messy and made it more complex than it needed to be.

Art

[Greolt]

Art

The A axis positions now look correct.  

The next problem I see is the amount we are asking the tool to cut on the first flank cut.

Cutting prior to this point would not cut proper involutes, ( if it cut anything at all), only cuts made after the initial 16 degrees should have any involutio n to them, which is why I eliminate d them in this case.)

Because you are now skipping moves until meeting tangency of the root or lower part of the involute, this means the first cut is aggressiv e and may be too much for the tool in some cases.

Approxima tely 50% of the remaining material is taken with this first cut and the remaining 50% shared amongst the next ten cuts (in this case).

Greg

[John S]

This will account for the odd tool I have broken when testing.
I thought it was me using too aggressiv e feeds on tiny cutters I'm not really used to and look forward to seeing what can be done.

[ArtF]

Hmm, Ill have to give some thought to that one to see what the best way of doing that woudl be without slowing the operation down too much..
   Its true the "First cut is the deepest"... Give me some thought time on this one..

Art

[ArtF]

Greg..

  Looking at it, that first move is a plunge cut, it woudl probably at least help if I changed it to plunge feedrate. . Since the first tangental position is almost always too high at position time, it seems plunging it properly would fix it up for most while I give it some thought as to a better solutions ..
    I mean I could start in center of tooth and move back to the edge, but Im not sure the load woudl be much better than if I plunged down the stock, in fact teh plunge would likley put less lateral pressure on the bit.. as long as the plunge feedrate was used instead of the cut feedrate. .
 
Art

[Greolt]

Art

I don't see it as a plunge cut.

Here is a snippet of code from the example posted above.  Tooth one.


 
  G1 Z-5.3925 F1000.000 0   'final Z depth while taking out bulk material at or near TDC
  F1000.000 0
  G1  A 0.1730   
  G1 X -4.0000  A 0.1730
  G1  A -0.1730   
  G1 X 0.0000  A -0.1730
  G1  A 0.0000   
( Tangental Sweep )
  G1 Y -7.3936 A 12.4825 
  G1 Z-6.0134    'this plunge move from Z-5.3295 to Z-6.0134 is made with the tool in free air
  G1 X -4.0000  A 12.4825  'this is the relevant move where that first aggressiv e flank cut is made
  G1 Y -8.8291 A 14.9025 
  G1 Z-6.0134
  G1 X 0.0000  A 14.9025

[ArtF]

Hi:

  lol, sorry, right I see what you mean. I was thinking in 2d not 3d..

Art