Trouble with 4th axis

[Stephen Fornelius]

I'm trying to cut a 8-tooth #25 chain sprocket with a .25 hole through a 1" blank.  The reason I want to do this is to make a sprocket with integral collars on both sides to stabilize it on the shaft. 

I can generate the project, but when I go to the cnc portion of the program, it refuses to work with 4th axis and defaults me back to 2.5D.  I've tried this a few times but no luck.  Attached is the file if you want to try it on your machines.

Any ideas?

Steve

[ArtF]

Hi Steve:

  Ill look into it and report back. Im working today on increasing the collision sensing in the
4th axis toolpathing so you may want to wait a day or two before doing one anyway as the toolpaths
may be better and more efficient tomorrow. The problems Phil found with low tooth counts are similar
to a problem with sprockets where the profiles are very flat so Im working on a collision safety function
to watch over all 4th axis paths. Should have something within a day or so..

Art

[ArtF]

Hi Steve:

  I can load it into 4th axis no problem, BUT there was a version a couepl ago I think where there was trouble with that. I just uploaded a version I used to test with, a new version, so redownload and you should be fine.

Id regen the sprocket, I have changed them so the waveform appears pretty much proper on the 4th axis screen.. ( was 1/2 tooth off..).

  (By the way, you can design that as a normal think sprocket, just enter -1 in the MaxDepth box and youd still cut it as a 1" thick sprocket.

Art

[Stephen Fornelius]

Thanks Art,

Sounds good to me.  I'm nervous about how much I have to take out of the sprocket for the shaft, so this will make things a lot easier vs. 3 different parts screwed together. 

Let me know what to do when you've got your current project completed.

Steve

[ArtF]

Steve:

  I just finished uploading the newest version. Download and install it. Redo your sprocket and try the 4th axis on it, it should be fine. Make sure the depth per step is low enough to be possible.

  I dont know how your going to cut that one mind you..takes a real small bit...

Art

[Stephen Fornelius]

downloaded the development version.  Set stock to 1.5" face width.  On 2.5D it shows the part properly, but when I go to 4th axis, it looks like its back to .25 inches thick (along the a-axis).  When I try to change the stock size, it asks if I want to delete the part.

If I say yes, I get a "Error found, Gearotic:1202, Attempt Recovery".

Steve

[ArtF]

Steve:

  Cant seem to repeat that. I go in and use facewidth 1.5" 8 tooth #25 sprocket. Looks good on creation. Send to 4th axis it is
not 1" as my last depth. Set depth to -1.5, sprocket is now 1.5" thick. Toolpath looks good.

> it looks like its back to .25 inches thick (along the a-axis)

  You mean on the X axis? ( The a axis is rotary ..cant have thickness there..)

Art

[Stephen Fornelius]

Art:  Yes, you're right.  I wasn't using the final depth parameter correctly, and you're also right about the "thickness" being the x-axis too.  Now it looks ok when I generate the toolpath.

Ran the last one on my mill just to see what it looked like and there was a LOT of movement - I'm assuming that all cuts on a specific tangent are done before the a-axis rotates to the next tangent.  It doesn't cut one gullet and then rotate and cut the next one, etc.?

Thanks for the help with this.  So far the only limitations I can see with the program are my limitations on experience and knowledge.

Steve

[ArtF]

Steve

  Thx, good to know its working as planned. As to sequence.. its probably a good idea I write something
that can be referred to by others in future. SO as to what its doing and in what order..

How GT Machines in 4th axis...
   
   
      GT uses a different thought for cnc work. While GM tried to figure out a toolpath based on the
      involute gear formula's, it meant that things like pulleys and sprockets couldn't be done
      in 4th axis because firstly they normally have no formula for their profiles, but secondly
      their shapes may or may not be appropriate for tangential shaving.
        If one uses a bit of the correct shape pretty much any shape can be done in 4th axis,
      but my experience tells me 90% of people do not use such shaped bits, they want to cut
      pulleys , sprockets, and gears with a straight flute. Fair enough. In GM, as I say it
      was done by very complex calculations at time of path generation.
          GT goes by the theory that if one had the tooth shape of anything, pulleys, sprockets
      or gears, one could cut anything by simply cutting to that profile. Simply is a bad word
      here I suppose, its still complicated, but less so than GM's routine.
     
      1)  GT firstly looks at a shape, any shape, and figures out a roughing path, this entails
        a routine that looks at depth and at the two intersection points there would be for any depth
        with that profile. It places the bit at center, and then makes a path striaght though the blank,
        at center position. If helical, the blank rotates while this path occurs. This is repeated
        on each side of center progressively until the side is hit. We then move down, figure out the
        intersection points on either side and repeat, telling the processor simply to never hit the
        sides. That pretty much concludes the roughing. Other than helical needing rotation, the blank
        will not rotate during primary roughing. Depending on your depth poer pass this can be many passes.
       
      2) Root Radialization. : When roughing is complete, two special passes of this name are done. This
        is done by figuring out for a given shape where the tooth shapes turn inward. Consider a normal
        spur gear, from the outside we follow an involute curve toward center, but when we hit the base circle,
        this involute turns inward, from an involute to a trochoid or straight radial to center. GT figures
        out where this switch over occurs. It finds that point , rotates the blank till that point is straight
        up, then moves the bit slightly to the side ( by 1/2 diameter of the bit) and takes a pass. It does
        this on either side. This makes a nice straight radial line for a root. In the case of a sprocket there
        may or may not be a point that allows a radialization pass, the system decides based on the shape.
       
      3) Tangential cutting.
     
        Now that the rooting has cleared as much as it can, and the root has been radialized as much as it can,
        the system will find all points above the "inward turn points" and for decide the lowest point that fits
        that description and the highest point ( outside of gear.). It them divides that distance into "Segmentation"
        steps. You set "segmentation" in the dialog. For each one of these depths ( typically 9 ), the system does
        similarly to the radialized root, it rotates the blank so that points normal is perpendicular to the tools edge,
        in other words rotate the point so its normal points straight into the tools edge, move the bit sideways by
      1/2 diameter of bit plus however much the normals tangency demands and take a pass.
       
       
    Of course any of the above passes can be helical or straight. If helical, we need to make sure all the above is done
    to the center of the blank, so we start off by rotating the blank 1/2 of the total rotation of the helical and end
    1/2 total the other way when we reach the other side. This means you may see quite a bit of motion. But it will always
    follow the above 3 steps. So when your cutting pay attention to what step your in. Root, Radializing, Tangential.
   
      If you get gouging in ROOT, you need to add more clearance. If you get gouging in Radializing, the bit is larger
    than you have entered. If you get gouging in tangency, the bit is too large. If tangency steps do cut, but seemingly
    not deep enough because previous gouge looking effects are not erased, thats a sign of too low a clearance as well.
   
    GT's Gcode IS experimental in that only over time will we see if allowing any shape to be 4th axis cut is
    preferable to GM's old method, and because of that I'm always here for advice, I love to see any photos, and often
    I can tell from a photo what you may have done wrong. ( OR where GT may have done YOU wrong.. :) )
   
      So always send us photos god or bad, I love to see them.
     
      Art
       

[philbur]

Hi Art, I assume from your last post that GT is now good for me to retry my 4 toothed helical. I am home tonight so should get a chance to try it tomorrow - Saturday.

Phil :)

[ArtF]

Phil:

  Yes. I have made changes to 3 things. First, the root radializing was a bit wrong in its math, this is now accurate
to many decimals, second I implemented a heuristic correction that tries to pull back from the edge if it sees the plane
is too flat.  Third I attached the clearance to the rooting. ( Tangential doesn't use clearance, it shaves the actual profile,
but roots can be pulled back from the side flanks with clearance. Check a clearance of 10%, then look at 100%.
( thats a % of bit diam.)

  So for something as small as that 4 tooth, Id advise a clearance of 25-40% or so, ensure zero is very accurate
( if not zeroed slightly above the part..), observe the toolpath onscreen, if it looks to close to the flanks for you,
then raise the path a bit by zeroing just .1mm above the actual radius called for.

  The math says you should then do well. So welcome back, good luck, and show me the result. :)

Art

[Stephen Fornelius]

First test wasn't too successful.  The toolpath looked fine, but moving at 5 degrees/minute is REAL boring.  Let it run and then when the part rotated the first time to a new tooth flank, it drove at g0 speed into the stock and then tried to shave off the material.  That doesn't work at all.  The plunge speed seems to not work - I used 1 inch/min and it still went at probably 30 IPM.

Tried with an 1/8 inch bit and it actually moved the column because it was trying to cut too deep on that plunge and then it wanted to move along x. 

I'm going to do a search and replace on the G0 and change it to G01 so it feeds slowly on the plunge. 

This is using the latest version which I downloaded last night.

Steve

[philbur]

I had the same problem with my 4 toothed helical when I didn't relieve the stock either side if the blank. It plunges in to the stock at G0, I have a broken cutter to prove it. The plunge feed rate would normally be set independent of the cutting feed rate and in my experience would normally be significantly less than the cutting feed rate. In the mean time I guess it's possible to hand edit the code if necessary.

Phil :)

[ArtF]

Hi Guys:

  Sorry about that, the output was planned for a blank of the correct width so it assumed there was nothing to plunge into. I will look at changing that so that it can properly plunge down on the side of the blank it needs to change.....though Im unsure while I think about it if that will cause other trouble on various types. Its why the requiement to zero is to zero on the front edge, the system then assumes from there outward is empty space, and the backside is the same dimensions.
  The plunge I think is only used in 2d work as the 4th axis cutting was never supposed
to actually cut anywhere except from side to side at the correct depth..

Art


 

[ArtF]

BUT.. one more note on that...if you find that changing G0 to G1 works for you..let me know and Ill make it so the G0 is changed to a plunge rate..that way a person could just set the plunge rate high if using a blank as envisioned by the original code...

Art