A mind of its own while cutting

New Home Forum Mostly Printed CNC – MPCNC Troubleshooting – MPCNC A mind of its own while cutting

Tagged: 

This topic contains 35 replies, has 11 voices, and was last updated by  John 6 days, 8 hours ago.

Viewing 30 posts - 1 through 30 (of 36 total)
  • Author
    Posts
  • #45706

    John
    Participant

    Any idea why it would do this?  This was in the middle of the 3rd pass.  I left it unattended for 5 minutes, came back to find it had viad off the tool path. Completely off from where it was supposed to be.

    Can someone tell me what may have happened?  It’s not the gcode because I run this again on wood and it did not duplicate this path.

    I have shown with red arrows the direction of travel.  It was cutting 0.5mm at a time and it was 1.5mm deep at the time. Was using 3.175 single flute endmill with 25mm/s feedrate

    Thanks for any help.

    #45709

    John
    Participant

    Forgot to add, using ramps+arduino with DIV268N drivers, the recommended firmware and was running from an SD card.

    #45710

    Ryan
    Keymaster

    Weren’t you having some other issues as well? Could they be related?

    Lots of things I would do different here, I never go over 15mm/s and typically go much slower, slotting on aluminum is not preferred now that we have tricoidal, and you walked away. That cut doesn’t look good at all, how did your test cuts come out? New bit or old? That one is probably dead now.

    Looks like it choked and the bit caught and did the steering.

     

    #45718

    John
    Participant

    Yeah I had some other issues with Z plunging lower than programmed to do at random times during a cut.  I took the Z rod out and found a few turns were stripped, so I turned the Z rod over and tried again.

    I was using trocoidal.  So you think slower but deeper is better?

     

    In ragards to the path it decided to take.  I’m new to CNC and the likes but as I understand, the code tells the drivers the direction to cut, and as I said earlier, there’s no issue with the gcode because I did this with wood without an issue using the same file.  So what would make it via off like that?

    #45720

    John
    Participant

    The bit was brand new (WAS).  My test cuts were no better, and I yes I know, I shouldn’t have tried this before getting it right.  Also I was not using any lubricant.  It’s been suggested I can use WD40 but need to use something.

    #45722

    Ryan
    Keymaster

    The point of tricoidal is to cut deep. It looks like your cut was so shallow you welded the chips to the bottom.

    Wood and foam are different, most things don’t matter. With metal and plastic you have to have large enough chips for them to take the heat with them, and to not work harden the surface. You should be fine without coolant if you get actual chips, but you should also get perfect test cuts before you dig into something so large. I would even say you should get the cut perfect in plastic first. Aluminum has a very exact window in which you can cut for all parameters, wood is basically foolproof.

    Have you verified you machine works tight yet? Long cuts with lots of big Z moves?

    This sort of mistake will usually break something. We have increase the power of the machine greatly so mistake aren’t as forgiving as they used to be.

    #45726

    John
    Participant

    Yeah on wood it cuts really well, other than the Z plunge issue I had (waiting for lead-screw to install).  I’ll keep persisting.

    Since I mentioned lead-screw, I only just noticed I ordered a T8 type.  Will this work?  So I have an 8mm T8 coming.  I think I read somewhere it should be T2.  Your thoughts on this?

     

    Thanks heaps for all your help Ryan (and everyone else that’s helped), not just here but throughout the whole time I’ve been posting here.

     

    #45731

    Jeffeb3
    Participant

    You aren’t using endstops for anything are you? It almost looks like it got a max endstop and then decided it needed to go far to the left before continuing.

    #45732

    Dalton Hinkle
    Participant

    same problem here and im on a sd card too maybe something with that?

    #45734

    John
    Participant

    No Jeff, no endstops at all

    #45736

    Robert Kyle
    Participant

    Just a thought/suggestion to check – the LCD cables from the display to the ramps/arduino boards. When I was building my MPCNC I had really long cables over 4 feet, this resulted in garbled messages on the display when the SD card was inserted, sometimes the display would be OK then would not. I then shortened the cables to just under 4 feet, display looked fine however when I ran some tests on paper (pen with standard crown) I did get some odd random movements. I therefore installed the standard ribbon cables supplied with the boards and LCD and have had no issues since. Would be worth checking.

    #45739

    John
    Participant

    Thanks.  I’m using standard cables supplied with LCD

    #45745

    Jason
    Participant

    I’ve had some issues when using “G92 X0 Y0 Z0″ to zero my machine instead of resetting the RAMPS.  I don’t like resetting to zero since that lets the steppers power down and risks allowing the machine to shift out of square (and my particular machine with the original Z parts and a slightly wonky roller on a 36″x 48” frame really likes to shift when the steppers power down.)  When I do that it seems to mess up Marlin’s internal coordinates somehow and it will start acting like it hit an end stop…then go totally crazy:

    20170716_102300

    That rectangle on the left should have been almost the full height of the sheet, basically everything got truncated at that point in Y….until it got a bit further and it started doing the crazy triangles thing.  Same gcode run after resetting the RAMPS and not doing a G28 works just fine.

    However – I’m also running an outdated version of Marlin because I’m too lazy to update.  (I’m still on a modified version of RC6.)  One of these days I really need to update it and see if it still does this.

     

    #45748

    Ryan
    Keymaster

    That is odd, RC6 is right around when the “?’s” showed up in the LCD. I use that command with the lowrider on every cut so I think it isn’t an issue with the current firmware but I could be wrong.

    #45776

    It doesn’t look like the bit got stuck or the CNC missed steps, otherwise it wouldn’t have gone in this direction. There is no mechanical explanation for this kind of path, or at least I don’t see any, especially considering the fact that it was trochoidal milling.

    The software limits issue is the one who seems to make the most sense from my point of view. I never use the Gcode for reseting the axis, I simply switch off and back on the machine, and never had this issue so far.

    #45777

    John
    Participant

    This is what I do, switch it off, align the machine where I want it to start, manually set the height then turn power back on

    #45778

    Well, thinking back again at this issue, the soft limits don’t make sense either, because they would have been triggered also during the first passes…

    So, maybe some electromagnetic noise.

    You mentionned that you left the CNC alone for a few minutes. Is it possible that you used some electrical appliances or turned on a switch somewhere else in your home, which may have affected the machine?

    I had some similar weird patterns while using the plasma cutter, the gantry moved in a straight random line at some point and continued the cut from another place. This was due to very strong electromagnetic interferences who messed up with the Arduino. Could be something like that.

    #45779

    thesfreader
    Participant
    Well, thinking back again at this issue, the soft limits don’t make sense either, because they would have been triggered also during the first passes… So, maybe some electromagnetic noise. 

    That would be my guess too.  I have had similar troubles when I switched from a dremel to a chinese spindle that I attribute to EMI.  It produced randomly vertical/horizontal lines and went back to the drawing.

    I’ve yet to test with shielding.

    #45780

    Heh that strange 2 days ago I was making some v carving and in middle of code suddenly it’s make to stroke one in x one in y direction and continue with carving. gcode generated with Estlcam.

    i also have spindle 500w but I don’t think so it’s related becose yesterday I was routing some parts for 4 hours and nothing happen code generated with cambam. But I will check power supply both on spindle and on Arduino if it have some pick voltage.

    #45781

    Dalton Hinkle
    Participant

    I’m having a similr problem. There’s not a problem for a few cycles of the code them I put another piece on to run it again and it adds an extra row of stars. I’m using it to mill stars into flags

    #45786

    Ryan
    Keymaster

    Dalton I think you should start a separate thread, The chances of these problems being the same are very slim. Post your code so we can see it there. Or you can go back through the replies and try all the trouble shooting we have done so far.

    John – Can we get a picture of your setup as it is, controls wires and such. Like a overall pictures and something a little closer of your controller and LCD. Have you ran my crown test code a few times? I f you can run the crown many times back to back with no issues this is a gcode thing, we can sort that out, if you have crown issues then it is an electrical thing and we can sort that out but right now we do not know where to start so it is best to stick to some proven gcode for a while. You need to run it back to back to see if it is a driver heat issues as well. once is not enough.

    This is a very non-standard problem.

     

    #45794
    Dalton I think you should start a separate thread, The chances of these problems being the same are very slim. Post your code so we can see it there. Or you can go back through the replies and try all the trouble shooting we have done so far. John – Can we get a picture of your setup as it is, controls wires and such. Like a overall pictures and something a little closer of your controller and LCD. Have you ran my crown test code a few times? I f you can run the crown many times back to back with no issues this is a gcode thing, we can sort that out, if you have crown issues then it is an electrical thing and we can sort that out but right now we do not know where to start so it is best to stick to some proven gcode for a while. You need to run it back to back to see if it is a driver heat issues as well. once is not enough. This is a very non-standard problem.

    Yes, pictures of the machine might indeed help.

    I don’t think it is a gcode file issue, since he said that running the same file again on wood showed no problem (provided it was running at the same speeds and same parameters).

    It’s a very weird problem, I’m not even sure an EMI issue could be the cause since the CNC continued on doing trochoidal passes. My guess would be that it would go nuts in a straight line in case of EMI issue, like it happened to me a few times.

    Very, very weird issue.

    #45826

    John
    Participant

    Sorry for late reply guys, been really busy at work.

    As far as EMI goes, I run an air compressor from the same switchboard to blow the chips/dust away.  When I left the machine going for that 5 minutes or so, I went inside the house to check on my 3D printer and see how the print is going, made a coffee then went back to see what I have shown you guys.

    I will take some closeup pictures either later tonight or tomorrow morning and post back.

    Ryan, I have only run the crown with a pen when I first got it going.  I have some new 10 degree V bits that I want to try anyway, so I will run the crown through a few times over back to back as suggested and also post results.

    Thanks

    #45827

    John
    Participant

    Oops I had forgotten that I have dismantled it si I can change the rod.   I will put it back together tomorrow and di those tests

     

    #45939

    John
    Participant

    So I run the code for the crown twice and both times it turned out normal without a problem.   The left was the first one.  The second one carved a little deeper but that was my fault as I had the bit lower to start with.  I took a video of both cuts but am unable to upload here.   Guys let me know if you want me to upload the vids on fb page or somewhere and give link if you want to check it out

    #45941

    Ryan
    Keymaster

    So it seems like this is just a gcode issue? Do you agree?

    #45955

    John
    Participant

    I’d like to agree with you Ryan since I’m only a beginner here.  I would say unlikely since I ran the same code on MDF again right after this happened and it cut more than 3 passes without an issue, I ended up stopping it on the 4th pass as the issue happened on the 3rd pass with aluminium.

    It makes me think this happened due to a force put on the machine.   Maybe the belts slipped enough on both X and Y to create this weird direction cut.

    You know maybe some things (like aliens) cannot be explained or solved, lol  I think maybe we move on and if this happens again then will definately re visit the issue.  Trouble is there are so many ifs at the moment, and it doesn’t help that I wasn’t there at the time it happened so I can’t even say I saw anything weird on the machine while it viad off to the left.

     

    I’m going to try alu again today, new 1/8 bit, maybe 1mm per pass, 8mm/s troc at 50% step length, see if I can produce some real chips rather than fine granuals.

    #45959

    Ryan
    Keymaster

    I would stick to wood, and maybe plastic for awhile, put in some time with the standard, I can’t really help troubleshoot aluminum, anything could go wrong with aluminum.

    #45964

    Moose
    Participant

    I am trying to plane the table to be level with the machine.  When I run the program it will run a little deeper or higher in the Z axis on a few lines of the code and then it will go back to normal.  I am using fusion 360 to produce the G code and Repetier host on my computer to control the machine.  What do i need to do to fix the Z axis from doing its own thin?

    #45965

    Ryan
    Keymaster

    There are a ton of variables in fusion, I am not sure where to start. Make sure your model is perfect? I know this isn’t the answer you want but you should really spend a lot of time in estlcam before you switch to fusion, so much easier to find issues if you have any.

    Attach your gcode or your setup file if you want someone to have a deeper look.

Viewing 30 posts - 1 through 30 (of 36 total)

You must be logged in to reply to this topic.