45 degree worm M2.5

[philbur]

But not if the cutter were off center, I assumed you were talking about the cutter off center on the y axis, what other off center is there, or am I missing something. Did you mean not zeroed on the z axis?

Phil :)

[philbur]

Art, if I understand correctly the less the horizontal offset for the roughing the more closely the cutter would follow the involute curve. If so ..... is it a possible solution to get the roughing to follow exactly on the curve of an offset, dummy involute, being offset either by linear or angular displacement from or about the rotational axis. Or something along those lines.

Phil :)

[ArtF]

Phil:

  No, I was speaking offcenter as in raising the bit. As you raise the bit the offset to the profile gets larger only on the
tight portions created by having too few teeth. What you describe for an offset is exact what it does now, the crux of the issue is that the offset works fine except in your case where the profile lies flat on the X plane..only happens on less than 8 teeth or so..
        Again, look at the sample toolpath, image it up just a small amount, see how the distance from the profile
grows larger as you raise it? And the effect is equal on both flanks.
 
Art

[philbur]

Hi Art, I'm not sure what's going on when rooting. With a step of 0.5 things look as expected. If I have a step of 0.25mm it steps correctly to a depth of 2.0mm then it jumps to the full depth of 5.625mm?

[philbur]

Success, very well done Art.

Phil :)

PS: Again the tool marks look worse than they are, although I thing they may be significant enough to result in excessive wear. Does the algorithm put the flank of the cutter at the tangent point or the cutter tip at the tangent point?

[philbur]

Regarding the tool marks on the tooth flanks. I checked for lost motion on my machine and I have only 0.02mm on the y axis and 0.003mm at a 10mm radius on the A axis. In any case looking at the code the y axis movements are all in one direction and the only cut where the end corner of the end mill is at the tangent point is the first cut. So the most probable cause of the tool marks may be left-overs from the rooting stage, although the number of marks doesn't seem to match the depth per step for the rooting. I may have to play with the rooting clearance some more. I'm not sure using the same tool clearance number as used at the end/beginning of each pass is a good long term solution.

Phil :)

[ArtF]

Phil:

  Good looking result, Im very pleased.

The marks are more from that plane thing I was talking about, as the plane of the profile goes flatter and flatter to one Z plane, the tangent functions approach infinity. ( Tan(90) = Infinity) and it causes such things. Raising the rooting path can help.

    Im unsure why the rooting was so deep, there is a bug that causes it to be weird if the stepdown is too large, the stepdown has to be small enough for at least 1 pass otherwise it fails to make any passes. Im looking into it.

That having been said, its a good result considering the near impossible angles involved. Not maky tangental shaving algorithms will handle a 4 tooth object in addiotn to sprockets and non involute shapes.. Im pretty happy, but Ill continue to work on making it better.

Art

[philbur]

Hi Art, I was beginning to think you might be away somewhere nice for the weekend.

The marks are more from that plane thing I was talking about, as the plane of the profile goes flatter and flatter to one Z plane, the tangent functions approach infinity. ( Tan(90) = Infinity) and it causes such things. Raising the rooting path can help.

Yes but the first tool marks are way down close to the base diameter so the tangent shouldn't be to tough down there, or should it? The tool marks are the same magnitude all the way up the tooth flank?

If I raise the rooting path will this increase the root diameter?

Phil :)

[ArtF]

Phil:

  Your right... How about just increasing the clearance setting. It affects only the rooting, so upping it a bit will not affect the end size, just how much the rooting takes off..

Art

[philbur]

Sorry Art but I don't think we are there yet. See attached photo: Left tooth space is set to 25% clearance and the right tooth space is set to 50%. The tooth profile includes the toothing cut in both cases. I'm not sure I still have the file but I think yesterdays run must have been set at 10% clearance.

Also the tool marks are from the toothing cut not left over from the rooting cut. I managed to get close enough to see what was happening. Z may stay constant but because of the blank rotation the tool tip effectively rises up the tooth flank for each rotation step. If the tool flank shaved at the tangent point rather than the tool tip at the tangent point then the tool marks would be avoided. Probably much easier said than done. Possibly all that is needed is to subtract a small % of the tool diameter form the z height. for each toothing pass.

Phil :)

PS: If I try to load a none allowed file and then correct the mistake it tells me I have already loaded that post!!!!!

PPS: I just checked the gcode from yesterdays good run, it was only 10% clearance.

PPPS: Why not use the tangent shaving algorithm to do the roughing above the base circle. and keep the rooting for the root.

[ArtF]

Phil:

  Shows every sign of a clearnace thats backwards... Im not sure how that can happen..

Did the toolpath show it getting wider as you increased the clearance or narrower ( Should be narrower)..
  Ill check it. As to the tip..I know in GM I made it keep the tip down to the base circle.. Im pretty sure I still do. Problem is that sharp angle again.. By the time you get tangent with that damn flat plane,
the base circle is probably up near the tangent point.. (Ill look into this.. ) In essense I think we're seeing limitations about the small tooth count again. (But that clearance looks backwards.. It isnt here so I need to check that..

Art

[ArtF]

Phil:

  Still not sure why your clearance appears to increase the amount taken. I did have time to check the tangents though. They work the way I remembered. The tool is not being raised for each shave, its being kept at what it figures is the base circle. BUT your rotating so much that that level is climbing the profile quite sharply. Again, I think most of that is the limits of the tooth count. Really your not supposed to shave a gear with undercut and a 4 tooth has a lot of it. On the other hand perhaps I can allow a drop to the root level if its an involute..that would probably relieve those tooth marks on the profile. Im still considering why the clearance error....maybe its the tool is too tight for clearance.... If the tool just fits in the root, and is then told to back away from the edge by .5mm, it will cut into the other side. Ill bet thats what happened. I failed to check for that so Ill add that tomorrow. Ill also consider if its possible to drop the tool to root level instead
of base curve level..that should clear up a lot as well if its possible..

  Thx for the info, Ill try to get a version out tomorrow with some corrections...

Art


 

[philbur]

I'm still curious as to why you can't use the tangent shaving routine to also do the roughing above the base circle. That would solve the original problem and you wouldn't need to add clearance for the rooting phase?

Phil :)

[ArtF]

Phil:

>>I'm still curious as to why you can't use the tangent shaving routine to also do the roughing above the base circle

  I used to do that in early GM. Users found though that it made the flute take too much off. It cause the flute to flex overly so that in the end rooting was extended to take up as much as possible. There are differences though from GM. I did find tonight that while GM keeps the depth at the base radius, GT was recalculating that value on each tangental step. Ive stopped that for tomorrows version. Now the Z will not move while doing the tangent. It will stay at Base radius. It can be argued that the depth could be a bit deeper so the tangent point is always above the tip, but as I recall we tried that in GM and it didnt work well for many gears as it causes overshoot where it shouldnt occur.... Ill look at a way to automate that better.

  Some of this is because the routines are meant to do sprockets and such as well, unlike GM which could do ony involute gears on 4th axis...so the profile follower is trying to be intelligent about it.. Ive told it to stop that for involute gears only. We'll see how that smooths things out.. I may allow for an option to not root above the baseline..though that can leave a high load on small bits..

Art

[philbur]

Users found though that it made the flute take too much off.

Can't you step the shaving across and down like you currently do the rooting.

1. Run the end mill straight across the middle with a user defined step down per pass.
2. Run the tangential shaving above the base circle with a user defined step across and down per pass.
3. Run rooting below the base line with a user defined step across and down per pass.
4. Then add a finishing cut if chosen.

That brings me to the layout of the CNC screen. It?s very confusing as to which parameters work with which machining strategy. A separate box for parameters for each strategy (and sub strategy) would make it much clearer and also help greatly with the users understanding of how each strategy works.

Phil :)