4th axis feedrates


Page 1 of 2 12 LastLast
Results 1 to 20 of 30

Thread: 4th axis feedrates

  1. #1
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default 4th axis feedrates

    Hi Tom,

    I have not used my rotary axis much for true 4th axis cutting, mostly indexing. But trying to cut a slot in a cylinder here today and having a bit of trouble. Doesnt seem like feed rates are working properly for me. Im trying to cut at 60 IPM on a 1.7 dia cylinder, 1.44 dia at tip of tool. My cam software is outputting feedrate at around 4500 degrees/min which seems right, but seems like its going way faster then this. Seems to be going to the maximum I have specified in the trajectory planner.

    The slot has a 90 degree turn, so part of it is moving only in X axis, and part of it is only in A axis, and you can tell the feed per tooth on the surface finish is way higher in the rotary moves. Any ideas? Thanks.

    Mark

    Similar Threads:


  2. #2
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Did a little more testing, it was definitely rotating at max speed. I looked over the help file for the trajectory planner and I see you recommend inverse feedrate for rotary axis. I switched over to inverse feed and everything seems to be working fine. Code is a bit messy looking but it works. So everything is good as far as I know. Thanks again.

    Mark

    Attached Thumbnails Attached Thumbnails 4th axis feedrates-20190130_224730-jpg  


  3. #3
    Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    4043
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Mark,

    Did you have a non-zero Radius specified in the Trajectory Planner? If so then the F4500 would not be degrees/min but rather inches/min at the specified radius.

    Also the feedrate is only in degrees/min with a pure angular move (no XYZ motion).

    So I think the only way to do it in general properly is to let the CAD system, that knows all the geometry, specify the motion time explicitly using inverse feed times.

    Nice part.

    Regards
    TK http://dynomotion.com


  4. #4
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    It was non-zero at first, then i tried zero and got some jerky motion, and finally everything worked great with inverse time.

    Just for your info, the Mazak I used to run at my last job would handle feedrate in degrees/min on any line with an angular move. Im thinking this is probably the industry standard, probably would make more sense for it to work this way if it were an easy change in the future. But in the meantime the inverse time seems to be working fine so no big deal either way. Thanks again.



  5. #5
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Tom,

    Im still having some issues with 4th axis feedrates. Seems like G93 inverse time is working fairly well for basic stuff. But the issue im having is that cam software is still outputting G94 for linear moves. And when theres a combination small segments, some g94 and some g93, it slows to a standstill when switching back and forth. Trying to run a part here today and its not working at all. Taking almost 2.5 minutes to do a single rotation, should be a matter of seconds.

    Im going to check if i can configure G93 for all moves. Other then that do you have any other suggestions? Theres no way you could setup G94 to simply feed degrees per minute whenever there is a rotary move could you? Thats how Mazak does it and it works well. Thanks again.

    Mark



  6. #6
    Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    4043
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Mark,

    Theres no way you could setup G94 to simply feed degrees per minute whenever there is a rotary move could you? Thats how Mazak does it and it works well.
    That's how it should work for a pure angular move if the axis is configured as Degrees with zero radius.

    If you can post a small GCode fragment, your configuration, and describe what is occurring we can look into it.

    Regards
    TK http://dynomotion.com


  7. #7
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Tom,

    See attached. The rotary moves are all using G93, and there are some Z axis moves only using G94 and i guess theres a small delay when switching back and forth. Its switching so many times with these small segments that its slowing the machine down. Its basically just sitting there stalling with this code. I timed it and it took 45 seconds to rotate 120 degrees (one chuck jaw to the next). I can post a video if you need it.

    Not having any luck modifying post to use G93 only. Thanks.

    Mark

    Attached Files Attached Files


  8. #8
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Quote Originally Posted by TomKerekes View Post
    Hi Mark,

    That's how it should work for a pure angular move if the axis is configured as Degrees with zero radius.

    If you can post a small GCode fragment, your configuration, and describe what is occurring we can look into it.
    Any way to change it so that any angular move is degrees per minute? For example a Mazak would do this:

    G0 X0. Y0. A0.
    G94
    G1 X2.0 Y1.0 F60 -Feed 60 IPM
    G1 A180.0 F750 -Feed 750 degrees per min
    G1 X4.0 Y2.0 A360. F750 - Feed 750 degrees per min, while completing the X/Y moves in the same amount of time

    My cam software works the same way by default. Seems like this would be the way to go.



  9. #9
    Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    4043
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Mark,

    Interesting. I think the problem isn't really a feed rate issue. The problem is with the Z only moves. Because only Z moves the A axis is stopped. So for example in the code fragment below on lines 290 and 340 the A axis is stopped. Lines 300-330 move the A Axis at commanded speed of 1850 degrees/sec. But because of acceleration limits on the A axis the A axis only begins to get some speed and then must quickly decelerate to a stop.

    N290 G94 Z1.5492 F50.
    N300 G93 Z1.5491 A3.541 F4777.61
    N310 Z1.5493 A4.165 F2963.58
    N320 Z1.5491 A4.604 F4208.28
    N330 Z1.5492 A4.814 F8842.92
    N340 G94 Z1.5491 F50.

    The Z only moves are sort of silly. They are typically only 0.0001 inches long. Any idea why the CAD is putting them in?

    I can't think of a workaround. I'll try to look into it.

    Any way to change it so that any angular move is degrees per minute? For example a Mazak would do this:

    G0 X0. Y0. A0.
    G94
    G1 X2.0 Y1.0 F60 -Feed 60 IPM
    G1 A180.0 F750 -Feed 750 degrees per min
    G1 X4.0 Y2.0 A360. F750 - Feed 750 degrees per min, while completing the X/Y moves in the same amount of time

    My cam software works the same way by default. Seems like this would be the way to go.
    I've never heard of a system that works like that. Do you have a reference? Also I don't think it would matter in this case as all mixed linear and angular moves use G93 to specify the time anyway.

    Here is a reference.

    Regards
    TK http://dynomotion.com


  10. #10
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Tom,

    Yes looks like your right. The Z only moves are slowing it down. I also tried everything in G94 and just limiting my max rotation speed as a way of setting feed, and it did the exact same thing. Still barely moving. Once it gets to about X0.3, there are no Z only moves and it runs fine. It takes nearly 15 mins to get trough this problem section of code, basically just sitting there barely moving.

    I cant really find a reference to the Mazak 4th axis feedrate right now. They dont publish much online. That is how it worked though for sure. I programmed/operated a 510c-II in a job shop for nearly 10 years and did a lot of 4th axis stuff. While googling this topic I ran across a bunch of other people with various feedrate problems and everyone seems to suggest using G93, so it should be fine I guess. Works for everybody else.

    Im not sure why the CAM software is doing this. Occasionally when doing 3d stuff it spits out some weird segments. Would there be any way you could do something like your "colinear tolerance" setting in the TP for the 4th axis to combine or ignore small moves like this somehow? Thanks again.



  11. #11
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Tom,

    This part has come up again a few times and Im still struggling with this problem. I cant seem to fix the problem of small "z only moves" with CAM at all. I did manage to get the post processor to included A values on every line but that didnt help since there was still no actual A motion anyway.

    Is there any way I can smooth the G code to help with this problem? I've tried all the options under the transform option but not having any luck. Ive tried the running average, 3 point, and low pass and the output doesnt seem right. Takes 1000 lines and outputs only 2 lines. How exactly do I use 6 axes smoothing properly?

    Really wish the trajectory planner could fix this somehow. Such a tiny segment causing the problems, if only it could just be combined with the other moves or removed somehow. Any other ideas? Thanks again.

    Mark

    PS: used the dro/2 feature to find center of 4th axis today for this job. Worked great!



  12. #12
    Member Mecanix's Avatar
    Join Date
    Aug 2019
    Posts
    339
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    What CAM software are you using?
    The software end should allow you to set linear motion tolerance, where tolerance should be 0.1% greater than those unwanted Z moves/distance, ideally.
    Also, and depending on your prog skills, there are several ways you could buffer the pre-processed output and create a method to combine rotary moves and/or set tolerance to both linear and circular moves while postprocessing. All depending on the software/technology you are using; that's why I was asking what CAM are you using.



  13. #13
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    I got my cam package from a shop that closed down, and heard something after about the licence may not be transferable. So id rather not to get into too many details lol. The software im using is a little older, around 10 years old i think. But works great for more then 95% of the stuff I do. I dont have an option for separate linear/rotary tolerance. Just one single tolerance block. Played around with it and didnt seem to help much.

    Im pretty sure this code would run fine on the mazak 510c I used to run. Somehow it seemed to filter out small things like this.



  14. #14
    Hi1's Avatar
    Join Date
    Oct 2019
    Posts
    2
    Downloads
    0
    Uploads
    0

    Default

    Powermill 2017 postproocesser can give me ????



  15. #15
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Tom, im running these parts again tonight and I think part of my problem is Kflop isnt getting the code from the computer fast enough. If its running slow and i hit feed hold, wait a few seconds for buffer to catch up and resume it runs WAY faster for a few seconds and then slows down again. This is very consistent and repeatable.

    The other problem is still there, its still jerky because of these Z only moves. But still runs way faster for a few seconds each time after a feed hold.

    Does this make sense? Any other tests I can do to confirm? Anything I can do to improve transfer speed other then buying new computer? Come to think of it, I have noticed when doing helix entry that sometimes it would run fast for a couple seconds and then slow down in some cases. I guess thats the same problem?



  16. #16
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Heres a video. Runs better (still jerky) for a few seconds then its really slows down (starts and stops). And if I do a feed hold for a few seconds it will run faster again for a few seconds and then run slower.

    https://drive.google.com/file/d/1Mei...ew?usp=sharing



  17. #17
    Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    4043
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Mark,

    We're looking into using the Collinear tolerance to combine/filter/smooth out the microscopic Z motions. If they were in X and Z like this:



    Then they would be smoothed, but since they are in A and Z they are not. So we are looking at handling A and Z smoothing. But like anything with Trajectory Planning it is more complicated than expected So no promises.

    I wonder if a workaround would be to change the post processor to output only 3 decimals for Z. That might round those 0.0001 Z moves to zero. You should also be able to round them with the KMotionCNC Editor "Transform Selection" feature. Just do a scale of 1 and offset 0 but set Decimal Digits in Output to 3. But then simple rounding up can make things worse in certain cases. And of course errors of +/-0.0005 inches will be introduced.



    Tom, im running these parts again tonight and I think part of my problem is Kflop isnt getting the code from the computer fast enough. If its running slow and i hit feed hold, wait a few seconds for buffer to catch up and resume it runs WAY faster for a few seconds and then slows down again. This is very consistent and repeatable.
    It might be that the AZ GCode glitches are being "corner rounded". We are looking into that also. So for example the 90 degree zig-zags would turn each microscopic zig or zag into 90 segments to download if the Facet angle is set to 1 degree. So you might increase the Facet angle to see if that has an effect. What Trajectory Planner Settings are you using?

    Regards
    TK http://dynomotion.com


  18. #18
    Member TomKerekes's Avatar
    Join Date
    May 2006
    Location
    USA
    Posts
    4043
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Mark,

    Also what Version of KMotion are you using in case we are able to come up with a patch to try?

    Regards
    TK http://dynomotion.com


  19. #19
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Hi Tom,

    I played around with facet angle and you were right, that's what was slowing it down so much! My facet angle was set at 0.5 degrees and I increased it to 5 degrees and it ran much better. I ran another test with just 2 revolutions of the A axis (see attached test code). The backplot time was 26 seconds. Having the facet angle set at 0.5 degrees dragged that out to just over 2 minutes. Increasing facet angle to 5 degrees dropped the time to 30 seconds which seems right considering I still have some extra starting and stopped for the Z only moves. But this is way ahead of what I was getting before.

    Facet angle 0.5 degrees: https://drive.google.com/file/d/1nUj...ew?usp=sharing

    Facet angle 5 degrees: https://drive.google.com/file/d/1OrU...ew?usp=sharing

    I attached my TP settings. I just read through the TP help file again and it looks like the facet angle only applies to corner rounding right? So if im using a fairly small corner rounding value, and a larger facet angle, I probably wont even notice a difference in most cases? Thanks again for all the help!

    Mark

    Attached Thumbnails Attached Thumbnails 4th axis feedrates-tp-jpg  
    Attached Files Attached Files


  20. #20
    Member
    Join Date
    May 2012
    Location
    canada
    Posts
    537
    Downloads
    0
    Uploads
    0

    Default Re: 4th axis feedrates

    Also, im using V4.34. I can repeat this test if you guys come up with a patch to combine the motion somehow, it would be a nice feature. But no worries if you guys reach a dead end, im pretty happy with it the way it is running now.

    While we are on the subject of the Trajectory planner, are you guys still planning on adding the low pass filter to the trajectory planner someday? Still really hoping we get to see that on a future release sometime.



Page 1 of 2 12 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

4th axis feedrates

4th axis feedrates