Problem with 4th axis toolpath

[Greolt]

You are going to have to forgive me if I am wrong.... .but I think there is an error in the 4th axis gcode.

It all looks good on smaller gears but the larger tooth count, the larger the error.

There is no way I can understan d the math behind all this, so the only way I can SEE it, is to draw it up.

I have drawn up a lot of examples and always get this type of error in the larger ones.

Hopefully the picture is self explanato ry. 

Showing the cutter and A axis (gear) at the position of the last cut of the involute.  As is shown they do not align as you would expect.

Greg

[Greolt]

Here is another example.

I will never want to cut a module 5, 300 tooth gear, but it is just to show how the error creeps up.

[ArtF]

Hi Greg:

 Ill do an analysis and let you know what I find. How was the first cut??
The photos make it hard to tell, at zero degree's, the tool should be in the tooth center of the hole.. but Ill see what the numbers say..

Art

 

[Greolt]

The photos make it hard to tell, at zero degree's, the tool should be in the tooth center of the hole..

Yes at zero degrees the tool is middle of the gap between teeth.

The pictures above represent the gear (A axis) position at last cut.

This picture (below) is the tool and gear position at first involute cut.

Greg

[ArtF]

Hi Greg:

 Yup. Sorry about that. I have found an error in my numbers. Surprisin gly, I also found a smaller error in the numbers for smaller tooth gears that appear to work properly. So Ive fixed one, and Im still fixing the other. The smaller gear numbers are basically OK, I used an asin() instead of an atan() function, but that only means they are a small percentag e inaccurat e. In your example, the root inversion gears, the numbers ar emuch more off, and woudl get worse as it rotates more. Im fixing that now, hopefully Ill have it done by tommorrow at some point..

Thanks
Art

[Greolt]

I used an asin() instead of an atan() function,

All double dutch to me. 

I am just glad I did not embarrass myself by saying there was an error and missing something basic.

Greg

[ArtF]

Hi Greg:

  I never want ANYONE to feel bad about putting in a big report. Its true that many such reports are misunders tandings, or inexperie nce, but I nver get upset at such things as the ones that are correct help me a great deal such in this case.

Thanks
Art

   

[ArtF]

Greg:

  I may have spoken too soon on this error though.
Ive recrunche d the code and cannot find an error in the numbers,
what I thought I screwed up on, was corrected fiurther down in the code.

  Can you tell me what your using to similate this? Or what size this gear is, Ive tried to generate one I can make to test it, but my 4th axis isnt
large enough to make one that a reasonabe l bit size will work in.

  Im intereste d though in how your figuring the rotated angles vs the final position, is there an easy way to do that? I could write something to grahicall y show it, but Im curious if theres somethign out there in cad that woudl allow me to experimen t graphical ly to check it. I take it somehow you put a gear on a cad screen with the hole in top dead center, then rotate it by 27 degree's or whatever, then draw the tool in position?

  Ill try VCarve and see if it will let me do that.. Id like to see where Im screwing up as the numbers all say it should be perfect..

Thx
Art

[ArtF]

Greg:

  Ahh, found the trouble. Its fixed and if you redownloa d youll find your now on version 1.64.2 . Ive found a way to graphical ly verify the numbers given and I verified tooth 1 of a mod 100 tooth gear. It showed the tool would cut the tooth properly without loss.

  Let me know if you run a test, mine is too small with that number of teeth, but at least I feel better kmnowing graphical ly it simulates properly now. I didnt before I corrected the error I found in the end..

Thx
 Art

[Greolt]

Art

Looks like I am not getting the new version.  Still says 1.64.0

Reran the same file and got identical gcode

Where do I get the new 1.64.2 version?

Greg

[ArtF]

Hi Greg:

  hmm, I just downloade d and installed and mine says 1.64.2 , are you downloadi ng the version that says 1.64? It may be your provider has cached the downloads page, try refreshin g then downloadi ng. Works from here.

Art

[Greolt]

Tried everythin g I can think of.  Even another computer. Still same version.

Have to go to work now.  

Greg

EDIT:  OK I have tried everythin g I can.

Refreshed browser.  Google searched on refreshin g proxies.  Installed on another computer.  Still getting 1.64.0

Is there somewhere else I can download it from?

[Greolt]

Ok this morning I tried the download again and now it is Ver 1.64.2.  Go figure.  

Unfortuna tely when I compare the old and new version's output of my Mod 5, 300 tooth gear, there is only a very small differenc e.

As shown in the two pics.

Just to make clear,   I have no need nor ability to make a gear of this size.

I saw a problem with a much smaller gear that I cut and used this large example gear to demonstra te.  To make the error easily seen.

Or I should say, the error I believe is there.  I could be wrong. 

Greg

[ArtF]

Greg:

  Nope, your not wrong. A part of my code didnt carry to the release properly.
Sorry about that, it has now been updated to version 1.64.3 , and Ive verified both numerical ly and graphical ly that its correct right up to
300 tooth mod 5 gear. In fact at 300 Mod5 its accurate enough that it probably doesnt lose anything no matter how large the gear gets.

 Let me know if you see any error, Ive reverifie d down to 10 teeth as well just in case.

Art

[Greolt]

Yes that now comes out as real good for the example gear.

Problem is, if I just change the pressure angle to 30 or 10, it blows it out again. 

Sorry to be a pain.

Greg