Jagged Diagonals?

This topic contains 32 replies, has 4 voices, and was last updated by  alan 1 year, 11 months ago.

Viewing 30 posts - 1 through 30 (of 33 total)
  • Author
    Posts
  • #4912

    alan
    Participant

    So, laser is on and running. Just dialing in various laser settings and trying to get a handle on Repetier..

    On one hand, things are going smooth. I did a square at 25mm, and a quick ruler test shows very square, right on the money.

    However, on diagonals, I’m getting a jagged “stair steps” effect. Any ideas on things to look at, machine wise?

    Anyone with ideas on checking out the laser as the culprit, I’m all ears.

    #4918

    Ryan
    Keymaster

    What program did you use to generate your gcode?

    #4919

    alan
    Participant

    Inkscape, and the JTECH photonics plugin.

    #4920

    Ryan
    Keymaster

    Did you check it with a visualizer, are the lines straight or jagged? Trying to determine if it’s the g code or the machine. It can’t really be the machine unless you were using full steps on your drivers, but it could be loose pulleys.

    If not zip it up and upload it here and I will take a look.

    #4921

    alan
    Participant

    I didn’t see anything odd, but…I’m kinda new to this.

    Circle is a random circle drawn in Inkscape, and used the plugin.

    The roundness in this is very jagged.

    Reminds me of looking at 1000x magnification on a bitmap, no real curves just individual pixels close to each other to simulate a circle when it’s zoomed out to normal size.

    #4923

    Ryan
    Keymaster

    That circle looks fine, a bit small to really tell. It is using the arc commands.

    If you are not using the ramps board I am not sure what other boards can handle them. I don’t think the uno based cnc boards handle arcs but I am not positive.

    Did you get it the kit from me?

    #4925

    alan
    Participant

    Yeah, everything here.

    Not sure if it’s a missing steps issue, off square issue, or what really to check.

    I did a 25x25mm square, and without high end tools, it was pretty square and of the right size. So nothing skipped there.

    Can you send me some sort of small file with a circle or triangle with a diagonal that you’re 100% sure would be straight on a well-set-up machine? I’ll try and run it.

    #4939

    alan
    Participant

    So I’ve now noticed (with some fans off that weren’t off before) that I get a distinct “tick” sound every time it moves X (positive and negative).

    I can also feel that ever so slightly while placing a finger on the pulleys as it turns.

    Is this a pot setting on the Ramps board? Maybe too low?

    What I don’t see is this issue on the Y direction. Smooth and very quiet.

    Also having some laser trouble, but this probably isn’t quite the forum for that. πŸ˜›

    #4946

    Ryan
    Keymaster

    It could be a loose connection or driver set wrong. If one wire is loose you will get that bad stepping. If you want to check the pot, If you are using my board, my driver, my power supply, my steppers, the voltage for the x and y should read about .7V.

    I had some issues with one of my MPCNC’s last night and it turned out I dropped a spool on it and it yanked the wires and made a bad connection on the far Y axis motor. reconnected it and working great. I was getting layers randomly shifting on my printers and you could hear the motor kind of grinding randomly.

    Are all your bolts loose, this could also be everything over tight, putting too much drag on the steppers? It should be much more loose than most people expect. I have people trying to return printed parts because they literally crush them with the bolts.

    #4948

    alan
    Participant

    I can try loosening some of them. Are you recommending the ones that hold the belt idler bearings on the rollers? If I am testing trying to roll just a bearing by finger, should it spin a little/at all without moving the roller?

    My issue with loosening them somewhat is that they’re just naturally kinda tight. Like, I had to thread the bolts through them, so even without a nut on them…they wouldn’t come loose if it wasn’t for needing to hold on the belt idlers.

    I’m concerned about one of the bolts from the inner assembly that isn’t reachable by any sort of wrench just due to small spaces after it’s assembled and one of the cross bars being kinda in the way.

    Going to double check all the wiring again too. I bought the wiring kit which was a lifesaver as far as time goes. And will check pot voltage too once I get another multi tester tool.

    I may undo the belting and see if I hear the same noises with no load on the motors.

    #4953

    Ryan
    Keymaster

    Without seeing it it could be so many things.
    The belt could also be too tight. 2 days ago in the forum a user was having the same issue it was because his belts were too tight.
    I think I have it in the directions as don’t actually tighten any of the bolts until you run it a bit.

    I was also just looking through thingiverse and the 5th set of belt tightening parts just popped up using 8mm bolts. I can promise you right now those are a bad option. But if you read the comments it is from a person that hasn’t actually used his machine yet.

    So many “improved” parts are form people that have not even run there machine 1 time. Tons of builds are getting made with lots of extra parts and add ons it is getting hard to trouble shoot.

    Please post a picture of your machine so I know what we are working with here.

    #4957

    alan
    Participant

    Circle

    There’s a good straight line, that’s a Z axis cut, then the diagonals coming out of it (and the circle) are where all the mess is.

    I’ve done a “box” that is all straight lines down the rails, and those cut great. It’s only when trying to do a diagonal cut.

    Everything is stock MPCNC, bought from here, except the laser tool and mount.

    Belts come off tomorrow. Thankfully they’re pretty easy to redo thanks to the zip tie idea. Will see if they still tick and stutter without having to move.

    #4958

    Ryan
    Keymaster

    no picture

    #4961

    alan
    Participant

    It’s there now. Typed faster than it could upload…heh

    #4962

    Ryan
    Keymaster

    To get a sense of scale how big is that circle?
    Are you always using the same spot or is that a different area of the build? maybe you have a rough rail?

    At this point I am shooting in the dark and I feel making my machine look bad. I’ll let you give all the other suggestions a try, and wait for a full machine picture.

    #4963

    alan
    Participant

    The straight line is like 75mm so circle diameter is close. I have tried a few spots. Will try different ones with my tests tomorrow.

    Not trying to make it look bad. Just trying to get it dialed in before trying some big cuts.

    What kind of machine pic(s) would help?

    #4964

    Ryan
    Keymaster

    I’d say just check out all the earlier suggestions and see if that fixes anything. Looks like your machine is really close, so it might be hard to find. Hopefully one of the drivers is just set a little low.

    #4982

    alan
    Participant

    Today’s progress…

    I took off the belts and loosened everything on the 4 outside rollers. Things slide pretty well without belts on, so I’m going to leave it like that for a bit.

    I put the belts back on a little looser, and tried some jogging from Repetier, and slowly tightened up to where I think things are ok from a belt perspective.

    However, I’m able to visibly watch the center assembly shudder while at slow lasing speeds (Under ~150mm/min).

    I just took the whole center assembly off and did some temp wiring to see if I can get the rollers to run ok without the weight of the inner conduit and Z assembly. If things go smooth, I’m going to put just the rails on, and see how it moves from there.

    I need to get a better multi test meter before I can check any of the pots values at this point, so I’m just doing basic mechanicals.

    Will see how it goes.

    #4987

    SteveC
    Participant

    Alan,
    Are you absolutely sure that all 4 four of the XY steppers are connected? When the belts were off did you make sure all four pulleys were moving in the correct direction and none were “stuttering”? This sounds like it could be a symptom of one open wire to one of the steppers. An open wire will cause one side to randomly step back and forth with the other side dragging it along in the right direction.

    #4988

    alan
    Participant

    Short answer…no. πŸ™‚

    Long answer…I think you may be right. One of my stepper plugs is kinda fubar. It looks like the underside panel fell off or something, and one of the little connectors in the plug SEEMS to be slid down a little. I was hoping the fingers on the wiring kit from here were long enough to make contact.

    I think I’m going to rewire via the solderless connection on all motors to see if that changes anything.

    #4989

    SteveC
    Participant

    My steppers have built-in plugs and are not very secure so I tacked both sides of them all with hot melt glue. Also I have a lot of 0.1″ connectors hidden in the conduit and I secured them all with electrical tape.

    Another thing to watch out for when testing is that an intermittent open circuit like this can be very sensitive to the position of the wire and may open or close at different X/Y positions. So I suggest putting a piece of tape or a sticky note or a sharpie dot on all the pulleys and remove the belts. Then manually tell the steppers to move a large distance and move the axes by hand all over repeatedly. Look for the back-and-forth stuttering that I mentioned on one of the steppers.

    #5039

    alan
    Participant

    Have to do some wiring stuff tonight maybe.

    After playing around with the conduit rails off, the motors still “tick” on my fingers…but I think that’s just me feeling the steps with some sensitivity. Looking at the pulleys, they don’t stutter or anything visibly, under load or otherwise. Everything looked very smooth without the middle rails and Z axis on.

    So, either this is the wiring issue that @stevec mentioned, something to do with my inner conduit rail, or something to do with the whole middles assembly.

    What I am mainly seeing in the “stutter” now, appears to be the X side of the middle assembly slagging behind and then catching up. Maybe it’s not as square as I think it is? Or the middle conduit possibly bent slightly but not enough for me to see. A jacked up bearing possibly? Though I like to run ~200mm back and forth on X direction and try to keep my finger on the each bearing for a full rotation or two just to see if there’s any tiny nicks or wobbles or anything “odd”. But they seem to roll really well.

    One thing I asked previously, either in this thread or another one. Realistically, how tight should the bearings be to the conduit? IE: Should I be able to spin them ever so lightly without it forcing the whole axis to move? Or when I try to roll them with my fingers, it grips enough to start pushing the whole thing?

    Kinda same question goes for the roller large washers. Should they be so tight they don’t spin by hand? Or should you be able to spin them a little/a lot?

    I don’t think I have enough pipe to cut another X middle rail to see if a new rail fixes it, and since my machine isn’t a square design (roughly 24×36-ish), I can’t just swap the middle rails to see if it starts happening on the other axis.

    Ooh…final question…does it matter which direction the Z axis is pointed in relation to X0 Y0?

    The thingiverse picture kinda has the machine at a diagonal, so I’m not entirely sure which is X and which is Y.

    If the left hand (not pictured) corner block would be 0,0 – then that’s how mine is.
    If the lowest corner is 0,0, then the tool is pointed in and towards 0,0 – then maybe I need to turn my assembly around.

    #5040

    Ryan
    Keymaster

    You found the problem. You should not be able to feel the stepper tick. I ticks 3200 per revolution, should feel buttery smooth.

    So now as said before, its either the driver voltage, or loose wire.

    X is left to right Y is front to back, typically.

    #5041

    alan
    Participant

    Yeah, I know that. πŸ™‚

    What I meant was in that particular picture, I couldn’t tell which was X and which was Y.

    In this pic:

    The tool/mount is on the closer side of X0, Y0…presuming that’s the lower left corner of the picture.

    Depending on which way your picture is oriented, it could mean the tool is pointed towards X-max, Y-max. Was wondering if it made a difference in direction?

    I’ve got a meter and going to do some motor testing.

    #5042

    Ryan
    Keymaster

    Makes no difference.

    #5043

    Ryan
    Keymaster

    if you switch the x and y plugs you can narrow down the issue. if the same motor is bad it the wires, if the other motor is bad it’s the driver. take you 2 minutes. Make sure the power is off.

    #5046

    SteveC
    Participant

    Alan,
    I like to have the rear left (in your photo) to be origin/X0Y0 and the front left to be XmaxY0. As Allted says, it does not matter but I am right handed and that orientation makes more sense to me when I am locking down the stock to the spoil board. Also the Y axis rail is not in the way. Also it make sense to rotate your table 90 degrees counterclockwise.

    #5057

    alan
    Participant

    All 4 motors give a little “pulse” when putting a single finger on the pulley/shaft when going at slow steps (75mm/m). So either I’ve got 4 bum motors, or it’s more normal than we think.

    I resquared everything, Re-belted, and tightened down things that were loose (motor mounts etc).

    The only major thing I did was on the trouble axis I just switched which plug was connected to each end which reversed my X. My pot on the “trouble” axis reads .68/.69, so I think I’m right at good there. I can’t seem to find .7 exactly but I don’t know how much 1/100 of a volt will hurt.

    Don’t know what seemed to do it, but things are getting there. A test diagonal today was MUCH better than the other day.
    I need to get a “real” table. Right now my table isn’t perfect and wiggles a bit during rapids when the motors whirl up and then come to a stop. That doesn’t affect the feed rate @ 75mm/m nearly as much as running some 4000mm/m or more during rapids.

    But another thing I just need to tighten up and make 100% perfect.

    I think another issue is going to be smoke and probably doing an air-assist or at least getting the laser mount it’s own fan. I don’t know why I didn’t include that when I bought, but will end up finding a small fan and maybe printing an air assist mount down the line. Having even a small fan blowing onto the workpiece seemed to help a little.

    Going to do some further test objects as I get things dialed in to precisely how I need it.

    #5067

    alan
    Participant


    So, I’m thinking a LOT of this might be a combination of the JTech Plugin and/or Repetier-Host.

    First picture is something I cut last night. It missed some lines, elsewhere on there is a straight line that isn’t in my image (but is in the gcode previewer in Repetier).

    I set it in inkscape to do 42mm by 15 mm…and it came out somewhere near 21 inches by 8 inches. Can post the gcode later if anyone wants to take a look. Some I attribute to lack of smoke clearing. When I have a fan turned on blowing onto the laser, it does much better. But my fan doesn’t provide 21 inches of coverage so some spots the smoke lingered and the laser got caught in it. Better fan connected to the laser mount absolutely needed!

    #5068

    alan
    Participant

    If anyone has a laser, can you send me some Repetier-ready gcode to draw a simple object or two (circle, triange, square) that are “known” to be pretty dialed in and perfect on the gcode end?

    Heck, I can add the laser part myself. Right now my only other gcode generator is Easel, which is set up more for the XCarve and GRBL shields.

    I just need to find out how well I can run something that is known to be perfect, then will work on how to get some better gcode outside of the JTech plugin (which is also kinda intended for their own GRBL firmware they developed.

Viewing 30 posts - 1 through 30 (of 33 total)

You must be logged in to reply to this topic.