Arrow key movement of gantry

Home Forums TinyG TinyG Support Arrow key movement of gantry

Viewing 15 posts - 1 through 15 (of 15 total)
  • Author
    Posts
  • #9601
    hgriffin1
    Member

    Hello,

    I am using the TinyG with Chilipeppr on a Big Ox Cnc and have it working pretty well, however when I try and jog the gantry in the x or y axis using the arrow keys on my keyboard arrow keys it “stutters” instead of moving smoothly. I am using Chrome on a Windows 10 machine and have a TinyG V8 board with the latest firmware and use a TinyG configuration identical to: http://pastebin.com/VsUnxgHd
    Please advise.

    Thanks,
    HG

    #9603
    cmcgrath5035
    Moderator

    The minimum move tinyG can make in the X direction is the distance moved in a microstep.
    That is 60mm_per_rev/(200 steps/rev *8 microsteps/step) = 0.0375mm

    If your Jog increment is set to 0.1, it will take 3 pushes of the arrow key to yield a 0.1mm move, with some remainder.

    If my quick math is correct, if your setting is 0.1, it will take three arrow pushes to get a move, 3 more to get a second move, but only two more to get a third move.
    Welcome to the world of quantized movement

    What jog increment setting are you using?

    #9610
    hgriffin1
    Member

    I have tried each of the jog settings on the menu.

    I decided to try running a cut file beyond the “chilipeppr” sample file. I found an SVG of a Fender Telecaster Body. I simplified the file to use only the outline. Inkscape svg loaded into JSCUT, G Code to Chilipeppr to TinyG. Lenovo Laptop running Win 10, Chrome Browser. Here is a video showing the stuttering which becomes very pronounced when cutting curves:
    https://www.dropbox.com/sc/turs0xpi74i3k62/AACLdmaZm4uTwk-f8ewrCcAXa

    #9614
    cmcgrath5035
    Moderator

    I would suggest you set $_pm=2 in-cycle for all 4 axes.
    That will improve accuracy, but probably not affect what you call stuttering.

    Did the Chilipeppr logo run OK?
    There are other test files in the tinyG workspace – they stutter too?

    Sort of difficult to tell, is the cut pattern relatively (or very) accurate?
    Perhaps post your Gcode file for a look-see.

    Nice looking machine, by the way

    #9615
    hgriffin1
    Member

    The Chilipeppr logo cut just fine. It’s a much smaller file.. I will try another test file tomorrow. I will change pm as well.

    The cut pattern for the file I ran looked fine. The stuttering actually shakes the table that I have the CNC mounted to, but the CNC is very solid and seems to be holding to the cut pattern quite well. It’s just a bit disturbing to watch it stutter so much…

    I am going to reapply the firmware upgrade again, just in case there is an issue with that. Apparently there is an issue with processor lockbits being set that isn’t that won’t allow verification… so I am not certain that the firmware isn’t corrupted.

    Thanks for the assistance!

    #9617
    cmcgrath5035
    Moderator

    Table shaking, particularly with the relatively heavy spindle you have, is not unusual. It one reason I migrated to a 400W quiet cut DC spindle, my original dewalt was overkill for what I was doing.

    You experiment by reducing the values of $xjm and $yjm from 5000 to , say 500, a dramatic reduction in ‘aggressiveness’. Run the job in air(or on waste material), see if you can discern a difference in shaking between 500 and 5000.
    If you milling speeds (F speed) are relatively high, that would increase shaking as well,
    Run a separate experiment, using the Gcode widget to slow speed down by factor of 0.5

    • This reply was modified 8 years, 6 months ago by cmcgrath5035.
    #9624
    hgriffin1
    Member

    The spindle isn’t a contributing factor. I had the same issue before the spindle was mounted (just the empty mount attached).

    I dropped the Jerk to 500 and it did reduce the aggressiveness as expected put the stuttering persisted. The milling speed didn’t affect it either. Very perplexing..

    BTW, I was set up with the $_pm=2 on all axis.. I tried $_pm=3 as well with no change.. Back to $_pm=2

    I am suspecting it has to do with the y-axis (and uneducated guess) so I removed the belts from the y-axis steppers to see if it was a friction or drive issue.. No change…

    I see several of the same setup Big Ox machines running smoothly. I am going to try and contact the people with the similar machines to see if they can share their config file..

    #9625
    cmcgrath5035
    Moderator

    Just watching the video, I might guess that CP is not keeping the buffer full, tinyG is stopping, waiting for more instructions.

    Describe your CP setup (OS, SPJS host version and OS, etc)

    #9628
    hgriffin1
    Member

    Latest version of Win 10 home version 1511, OS build 10586.218, Intel Core i7-6500U cpu @ 2.50 Ghz, 12GB ram, 64-biy operating system running on x64 based processor. There should be plenty of firepower but it isn’t a dedicated machine…

    Serial Port Json Server 1.92_windows_amd64

    TinyG V8 running tinyg-master-440.20 firmware (reloaded again today just to be sure)

    I have a couple laptops and tried another one today to see if it made any difference. It was a Lenovo running Win 10 as well.. No difference.

    #9631
    cmcgrath5035
    Moderator

    As your Gcode is running (just run in Air), do you see strange variations in the fubber status from the CP display, particularly during an area where Gcode stutters

    Can you post your Gcode, or a piece of it long enough to produce stutter, so we can try it?
    Copy to a CLoud Drive and post URL.

    #9632
    hgriffin1
    Member

    Not familiar with the term Fubber Status.. Could you elaborate?

    #9634
    cmcgrath5035
    Moderator

    Sorry, buffer status (actually, the number of buffers left available for use).
    It is a maximum of 32, if SPJS is doing it’s job keeping tinyG filled the number of available buffers will be low.
    If SPJS stops streaming commands, the available buffers may spike up, a pause (stutter) occurs then streaming of data continues.

    #9638
    hgriffin1
    Member

    I tried to run the ShapeOko Calibration file that you provided the link for. The gantry did a rapid move diagonally from home, stopped, and the spDir led started flashing. I was using Json 1.88 as you suggested.
    After resetting, I got the same result again. I didn’t have enough time to run another file to check the buffer status. I will try and get to that tomorrow.

    I hope to be doing an axis rewire and replacement of the limit micro switches with Hall Effect switches late in the week when the materials show up. It may not be related to the issue but I think shielded cable can’t hurt and the wire gauge is heavier so currently delivery will improve.

    Thanks for the suggestions.

    #9639
    cmcgrath5035
    Moderator

    Wow, it has been a long time since I ran the file.

    Turns out it has no real setup code, such as a G21 to say it is mm code.

    So if you were in G22, inch mode at the time, it took off for 6″ when it was authored for 6mm. Good thing it hit a limit – It would have headed for 1 inch deep next.
    Sorry about that.
    The code was actually developed to run on the original, rather small, ShapeoOko1.

    It did not have an M03 to turn on the spindle either, sorry about that too.

    I edited the file to add a G21 and Mo3 at front end, M05 at end.

    Try your Gcode file first with SPJS 1.88

    Redownload the ShapeOko file, run in air.

    By the way, things always run better with tinyG in mm mode at the moment.
    Consider regenerating you Guitar file in mm.

    #9640
    cmcgrath5035
    Moderator

    Wow, it has been a long time since I ran the file.

    Turns out it has no real setup code, such as a G21 to say it is mm code.

    So if you were in G22, inch mode at the time, it took off for 6″ when it was authored for 6mm. Good thing it hit a limit – It would have headed for 1 inch deep next.
    Sorry about that.
    The code was actually developed to run on the original, rather small, ShapeoOko1.

    It did not have an M03 to turn on the spindle either, sorry about that too.

    I edited the file to add a G21 and Mo3 at front end, M05 at end.

    Try your Gcode file first with SPJS 1.88

    Redownload the ShapeOko file, run in air.

    By the way, things always run better with tinyG in mm mode at the moment.
    Consider regenerating you Guitar file in mm.

Viewing 15 posts - 1 through 15 (of 15 total)
  • You must be logged in to reply to this topic.