Tangential Saw shaving 4th axis

[David]

Hi Art

Thanks for all your help.
Have set up a gear to the same parameters, and attached a screen shot (using gearotic screen shot gave only a white rectangle) so this shot done another way.
All seems the same except the material size boxes which are different and no wavy black lines for the shaving.
Not had a chance to try a cut to see how this goes
Attached gth file which now can be reloaded and screen shot.
Thanks again
David

[ArtF]

Hi David:

  Something isnt matching up well with this gear, to be honest Its much smaller than I anticipated in design, so Im seeing if I can track down the error,
doesnt seem too deep to me though, seems to be shaving outside the disk..

Art

[ArtF]

David:

  As  I try larger modules I see the problem diminishes, so its an accuracy issue, It likely hasnt been noticed as
most 4th axis cuts are larger modules, so Im going to have to trace the math on thisone to figure out why the
shaving path is actually out of the disk.. Doesnt look to me worth cutting till I figure it out..

Art

[David]

Hi Art

Sorry this is causing you so much trouble.
I tried cutting an 8 tooth 4 mod gear just as a trial, given that you think the small size might be a problem.
I used a 2" blank, a 2.75 diameter saw 37 thou thick, X and Z axis zeroed on rotary table centreline, Y zeroed just touching blank at X=0, Z=0. Table facing X--.
The first side of first tooth seems OK, though some air cut, as soon as the A axis turns cut is applied via Y the saw plunges and bends under the A motion.
Could you just tell me if I have set this up incorrectly?
Attached gth file, screen shot, tap file and a picture of cut.
Many thanks
David

[ArtF]

David:

  Dont be sorry, its really my fault. I have searched this week and found an error. The toolpath I think is incorrect.
Ill release a temporary fix tomorrow so you can generate the file.

  When you use a saw though, you reallymust specify it as wider than the balnk really is, the rotation should never
occur while the blade is in the material, but the toolpath will rotate when the tool is on one side or the other, so
when using a saw you have to make the system think the blank is very wide so the saw will move all the way out of the material
on either side when cutting. That will fix one error I think, and the other error should be better with tomorrows
release.

Art

[ArtF]

David:

Just to explain that abit more, the .25 for depth/width you have entered means the blade will go .25 to one side of the blank  , plus the diam of the tool , this was really meant for flutes, so to make the blade miss theblank , youd have to enter a number large enough to ensure the the blade is clear of the blank on either side of its cut.
  We'll see how tomorrows toolpath looks though to see if it all looks better..

Art

[ArtF]

David:
 
  I just uplaoded a new version with a fix for the 4th axis. Your toolpath should be much more exact now... BUT,
your experience has shown me I miscalculated on the settings for saw blades.
  The program is meant to have the tool centered int he center of the blank. If the gear is 4mm thick and the tool is 1mm is radius,
the path cuts from 2.5mm on one side to -2.5mm to the other side. BUT, a saw must move much further to clear itself from
the blank. So judge that carefully. You need to tell the system to make the gear much wider to give yourself clearance.

When the code is rewritten Ill take all that into account, for now though, if using a saw you need to make the gear wide enough to
clear, just use the gear you have, but in the "depth/width" box, enter a large enough number.. play with it and it will become obvious I think..

Art

[David]

Hi Art

Loaded new version and tried the same 2" diameter 8 tooth approx 5 module gear as before.
X and Z zeroed to rotary table axis Y to edge blank. A full 2" width giving a large saw clearance well past full cut depth.
First side first tooth seemed OK, noticed that direction cut reversed with each change in Z axis, so some normal and some climb cuts.
Another odd thing as Z down increased the saw cut the blank at Y zero! Z was accurately zeroed to plus or minus a thou.
Disaster struck around line 500 plus on the code, after a small A rotation (Cleared cutter) there was a larger rotation (cleared cutter) then then a 0.25" or more Y movement then an X cut breaking the saw and bending the blank.
Attached gth and tap.
Could this be a shaving cut with the A axis rotating in wrong direction rather than moving to next tooth?
Sorry and thanks again
David

[ArtF]

David:

  Lets look at the code around that spot and see if we can see whats up.

Heres the code around that position with some comments..


Ok, all the code up to this point has been rooting, with no rotation at all. The saw is zeroed to the top of the blank,
in the center and the rooting had been done. Both side of the tooth should be visable and shaving is about to start.
Are you saying only one tooth side is visable at this point?


line 503:          X -1.0370  Y 0.0648 Z -0.4200 A 0.0000
          X 1.0370  Y 0.0648 Z -0.4200 A 0.0000
          X 1.0370  Y -0.0648 Z -0.4200 A 0.0000
          X -1.0370  Y -0.0648 Z -0.4200 A 0.0000
G0        X -1.0370  Y 0.0000 Z -0.4448  A0.0000
G1 F10.0X 1.0370  Y 0.0000 Z -0.4448 A 0.0000
          X -1.0370  Y 0.0000 Z -0.4448 A 0.0000
          X 1.0370  Y -0.0000 Z -0.4448 A 0.0000     
         
//so down to here we are routing a spot at the maximum depth in the center.
// y== 0 , so dead center, We're at max X too, so the saw is outside the blank.
// we no rotate to 10 degrees. This is the "Root Wall" Rooting, where the system takes one slice
// with a roation so the base circle is tangent to the saw. Hmm, this looks wrong to me, looks like the system is taking the
//total thickness instead of 1/2 of it when positioning the blade.
         
G0        X 1.0370  Y -0.0370 Z -0.4429  A10.3960    //this may take too much but it shouldnt hit, the blade should be clear.
G1 F10.0X -1.0370  Y -0.0370 Z -0.4429 A 10.3960      //and we sweep to ther side again..to clearance..
          X -1.0370  Y 0.0370 Z -0.4429 A -10.3960
          X 1.0370  Y 0.0370 Z -0.4429 A -10.3960
G0        X 1.0370  Y 0.0370 Z 0.2000  A-10.3960
          X -1.0370  Y -0.1116 Z 0.2000  A-0.2727    //and here we clear the blad to safeZ before rotating.
         
      Do you know if it hit prior to this?
      Dave, how is this blade oriented. Top of blank, cutting on the X axis straight down?
     
      Art

[ArtF]

Daniel:

  I think I see the issue. Im afraid I introduced a bug in this code as we did some tests at the end of last year.
This code is not for a sideways saw, the Y and Z are interchanged.. The Y is going back and forth as it root, and
if your using a saw on its side, the Z should be going up , then down for the clearance. Im investigating to see
why..

Art

[ArtF]

David:

New version 2.010 is online for your problem. I found the bug and repaired it. The code is now for your side cutting saw.

Keep the clearances as we discussed, stop your cut if you cannot see both sides of the tooth before it starts to rotate.
For now, uncheck the "Root Wall" boxes, they shouldnt matter much anyway , I just want to verify them..

What should happen..

You saw shoudl cut to the back, then the Z goes up and youll cut back. The Z then goes down and this repeats with
the Y making the blade go deeper on each group of passes. This cutting will get narrower and narrower as you get
deeper and both involute flanks should become apparent. Shaving will then begin. Iv elook at the code thought the run and
it looks proper , and the saw should be withdrawn to a safe position before any rotations.

  The bug was reversing Z and Y, so it little wonder it hit. Sorry about the blade. I suspect your one of the few to run a saw
tangential routine, these things get debugged slowly as a result. Ill be trying very hard to get a simulator working on all this
during this falls development, the lack of testing of some of these routines is an issue, and Id feel better if we could all
run a simulation to see what material is removed..

You can redownload anytime..

Art

[David]

Hi Art

Thank you for your Sterling efforts and magnificent support.
It is midnight here in the UK, so will have another trial tomorrow.
At present I am using old saws and any scraps for blanks whilst I learn the program, so no real damage done!
I  will be intrigued if we can get this working for clock gears because the availability and cheapness of narrow slitting saws makes it a very attractive proposition.
This is only a hobby, going from large scale steam to full size steam cars to large scale steam to small scale steam to maybe clocks, oh and I broke my neck somewhere around steam cars  :)
it would not be fun if there was no challenge and we rely on people like yourself to make some of it possible.
You may like my steam car site locomobile dot plus dot com
David

[JustinO]

"...because the availability and cheapness of narrow slitting saws makes it a very attractive proposition."

Have you tried flycutter/gearcutters? With simple flat edges on a single point, you can generate involute toothed gears.

[David]

Hi Justin0

I know, but you do not need 4 axis CNC or this program to do that.
I can make and use single or multiple tooth involute cutters, but it is one or several cutters per module or DP
My idea is to be able to use essentially a simple saw and let the computer do the work.
Given my limited mobility it is easier and more productive to let the CNC work whilst I get on with another component  :)
Kind regards
David

[ArtF]

David:

Well, for my part Id just like to know the routines work. Weve all seen nice examples of the flute cut tangentals, but I havent
seen any by saw. This is a good way to get rid of bugs and to learn a bit more about changes I need to make in the next iteration
of the 4th axis code. The code I checked tonight made much better sense, I couldnt understand why the Y was moving in and out,
took me awhile after this hiatus to figure out it thought the saw was on top. :)

Anyway, we'll figure this one out and hopefully by start of next year we'll have a graphical way to simulate the damn things..that
will save us all a bit of head scratching, and will allow us to experiment more on bevels and such..

Thx
Art