Ellipsis instead of circle

This topic contains 35 replies, has 5 voices, and was last updated by Profile photo of Gianluca Magnani Gianluca Magnani 1 year ago.

Viewing 30 posts - 1 through 30 (of 36 total)
  • Author
    Posts
  • #6057
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Hi all,
    I finished building my mpcnc and I’m very excited about the result. It’s a great project!
    But tonight I gave my mpcnc the very first try, a simple circle on a piece of wood:

    ellipsis

    The result is an ellipsis instead of a perfect circle.
    I checked my frame and I found a perfect square.
    I checked my rollers and I found some play on the two axis. My ellipsis is due to a non perfect square on x and y directions, or so it seems to me.

    Is it normal to have some play on the axis? I hope so… To be honest I had to reprint the rollers scaling from the original model to 98% of the size. With the original size I had the bearing not touching the conduit tube.
    With my scaled prints all the bearing adhere to the tube and I thought everything was right.
    So my question is: is it normal to have some play on the rollers? I think I could start my mpcnc keeping the axis attached at the frame and send a M84 S0 to keep the rollers in place whit the steppers and then start my cut. Is it a solution?
    Anyway tomorrow I’ll try this method to see if it works for me…

    Anyone had issues like the this?

    Thank you in advance…
    Gianluca

    Attachments:
    #6063
    Profile photo of vicious1
    vicious1
    Keymaster

    Did you get everything from here? Did you modify anything?

    Scaling of just the roller F should be okay but you really should not have had to scale it down most people are scaling it up. What size conduit are you using?

    Both of mine cut great circles. If yours is sitting cockeyed before it starts I suggest loosening all the roller locks and all the center bolts and all the z axis stuff while it is assembled and try to tighten it up a little at a time seeing if anything is making it crooked. If you look at the parts they should really be trying to pull it in to a nice square. If you assemble the center real tight and put it together you could have actually assembled that crooked. That is why the directions say not to tighten anything until you are ready to run it.

    Let me know.

    People have done some other fixes by aligning the axis every time before they start a cut but really it’s easier just to put in the work once and never have to touch it again.

    #6065
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Thank you Ryan for your quick reply.
    I forgot to mention some important things…

    – I printed all the parts with my printer
    – I printed the IE version, my conduit measures 24,95 mm
    – I had to scale middle joiners, middle ends and rollers at 98%. I didn’t scale the better middle Z, beacause the first print made at 100% scale seemed to fit correcly with bearings and tube.
    – I used bearings from banggood: low price but not great quality I think… they are not very sliding… my bearings

    I assembled the center real tight, maybe I have to scale at 98% also the better middle Z to have congruency with the other middle parts? In fact it makes a lot of sense…

    From your last sentence I argue that the play in the x and y axes is a common issue… is it so?

    I made 3 short video to show you the beahavior of my x axis… pardon the darkness, I had to take them with hurry this mornig before going to work:

    x axis play part 1
    x axis play part 2
    x axis play part 3

    Anyway, as soon as possible, I will make a new try aligning the axes before starting. If I’ll get a real circle I’ll be quite satisfied…

    Let me know what you think about reprinting the middle z at 98% and if the beahavior shown in the videos is acceptable or not…

    Thank you very much!!!
    Gianluca

    #6066
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    Hi,

    I have printed too the IE version at 100% scale, my tube are stainless steel and have an perfect 25mm. All the 608 bearing of the roller are in contact with the tube. I now building the Z axis and some of the bearing at first do not touch the tube, to solve this i simply removed some pla that make the curve that not allow the bearing to touch the pla. removing some of the pla put the bearing a little more inside. Some bearing have an little play in the 8mm screw, when you screw the M8 you can make some pressure to the bearing to have the bearing more inside.

    i have checked your roller play video and this will not be an problem when the stepper motor have electricity.
    When the motor is running it not allow to pulley to move and remove all the play you have. I do not have checked the marlin firmware given my Vicious, but you need to configure the firmware to allays put electricity on all motor. by default when an motor is not used for some time, the electricity is cut and this can be an big problem.

    About your circle problem, this can be related with missing steps.
    you need to check if you steppers have enough voltage to not skip any steeps, what voltage you have on each steppers?

    Other thing is important, for your first tests i think is more easy to find the cause of this problem if you make circle using an pen. The pen will not have any lateral force and pretty sure it will not miss any steps. If the circle drawing is perfect, then is very likely is an voltage problem.
    Check too if you have some play on the z piece.
    Are you sure your cnc has all corner with an perfect 90 degree? an easy way is to measure both diagonals, if both diagonal have the same dimension, then you have an perfect square or rectangle. But i pretty sure this is not your problem, to have the circle posted in the photo your cnc should be really out and this will be easy to see.

    #6077
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Hi Filipe,
    could you post a picture of the point where you “removed some pla that make the curve that not allow the bearing to touch the pla”?

    My problem cannot be solved pressing “more inside” the bearings, my problem was the opposite, I had the bearings not touching the tube. So I had to scale down the size of some pieces.

    It will be no problem to configure the firmware to give electricity to all motors continuously. That was may guess for a solution.

    I’m also quite sure I don’t have any missing step, first of all I checked very carefully the vref on my drivers, then I’m quite shure in case of missing step I should not obtai an ellipsis but an open curve… similarly at what happens on a 3d printer that produces slided layers.

    I’m sure that all of my corners are 90 degree, the diagonals measures exactly the same. My axes, as I wrote, may not always be perpendicular due tu the play in the rollers.

    #6080
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    The M8 screw im using are not an perfect fit inside the 608 bearing, they have a little gap of halt mm or less. this is not much space, but in some case it can make the difference between all 608 touching or not the tube. but In your case the main problem is not having an tube with the correct 25mm… if the gap is to big, i guess the only solution is to reduce more the size and print again!

    “It will be no problem to configure the firmware to give electricity to all motors continuously. That was may guess for a solution.”
    This is not the reason of your circle problem. The electricity is only cut when an motor is not used for several seconds. When making an circle, all the x and y motors are always in use. This cut of electricity is mostly an problem exclusive of the z axis.

    “I’m also quite sure I don’t have any missing step, first of all I checked very carefully the vref on my drivers, then I’m quite shure in case of missing step I should not obtai an ellipsis but an open curve… similarly at what happens on a 3d printer that produces slided layers.”
    Most 3d printer use only one motor for the x and y axis. In this cnc they are 2 motors for each axis, what will happens if only one of the x axis motor miss steps?

    #6086
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Ok thankyou..
    About electricity cut, you suggested it…

    Your last question is interesting, I didn’t think at it.
    If only one motor misses one or more steps I will get 2 axed not in square and then theorically an ellipsis instead of a circle, but, missing the steps in the middle of the cnc program I would see a different path for the subsequent layer passed by the end mill. My ellipsis instead is identical on the three layers passed by the end mill. And I’m not sure I should get an ellipsis instead of an open curve.

    Anyway, let’s return to my original question: is the play on the axis shown in the videos normal?

    Thank you!!!

    #6088
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    I was only to try to help you and in the process it was made some confusions… first, i never say the electricity cut could be the reason for the circle problem.. I speak about this only to explain that the play you are showing in the 3 videos only are happening because the motor are not energized. When all motor have electricity, the play will disappear. To check what i talking, you only need to power on you cnc and make an homing, after this check if you still have any play?

    is the play on the axis shown in the videos normal? yes.. i using stainless steel tube, was result i think i have less play, but still have some… And roller that runs using 608 bearings and tubes i think is expected this type of play.

    And again, about your circle problem. If you say you do not have any play on the Z axis and your cnc was all corner 90 degree, i put my money that your problem is missed steps. I could be wrong, but i think it will not be a lose of time if you check this by printing an circle using an pen attached to your Z tube. This is an easy and quick test to be sure if is an geometry problem of your cnc or missing steps.

    #6091
    Profile photo of vicious1
    vicious1
    Keymaster

    We all might be over thinking this. Check to make sure your pulleys can not be pulled off by hand. I had this happen when one of my pulleys was loose and moved a few mm in each direction on the flat spot. exact same problem

    Everything looks like it is working right, what ever this is is something so small you keep over looking it.

    #6115
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Filipe I think you are right when you say that my issue could be the play on the z axis. I didn’t understand what you were meaning at the first read of your post.
    My middle definitely keeps out of square the x and y axes tending to make an angle lower than 90 degree. Pulling by hand the two axes I can easily keep them square, taking off the hands the middle flexes the two axes of a couple (more or less) of degree.

    I think the cause of this problem could be the “better z middle” printed at 100% scale instead of the 98% scale that I used to print the other middle components. What do you think about this?

    But, then again, Filipe, maybe the solution you mention for the play in the z axis could be right also for me, but I didn’t understand the exact point where you removed some pla and which bearing you put more inside. Don’t you have a picture explaining that?

    Thank you very much!!!

    #6120
    Profile photo of vicious1
    vicious1
    Keymaster

    Parts should not have to be scaled, they are actually sized small, that part is making me wonder if something else is going on. Does your printer pass a calibration cube test, inside and outside dimensions?

    If you did scale one part, every single part should have been printed at the same scale, most important being the center pieces. Center pieces set the conduit distance so then all the other pieces also need to be scaled.

    #6121
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    I scaled everything except the better z middle and the corner blocks…
    I’ ll try to print the better z middle scaled too!!!

    #6122
    Profile photo of vicious1
    vicious1
    Keymaster

    Corner blocks also control the distance between the conduit, might as well just do them all.

    #6123
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Ahhhhhh yeah, I didn’t focus that! Now it’s clear: also the corner blocks, if not scaled at the same factor are a problem..

    #6125
    Profile photo of Asho777
    Asho777
    Participant

    My take on what I have read!

    The movement you have shown in all three videos, I would expect if the steppers were not energized!
    So that side of it all looks good to me.

    The oval shape that you seem to be achieving, I suspect it is caused by the fact that your “better Z middle” is slightly longer then it should be.
    By longer, I mean that your X and Y rail will be pushed out of square, in the quadrant of the Z assembly, I’m guessing by 1 to 2 degrees!

    To check this I would move the tool over the same position where you cut that circle/oval. From both ends of the X rail, and on the same side of the
    X rail drop a plumb bob down to your work surface/table. Mark the two points on this surface and draw a straight line between these two points.
    Now do the exact same thing on the Y rail. Take a square and check where the lines intersect, are they square?

    Cheers
    Greg.

    #6152
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    >>The movement you have shown in all three videos, I would expect if the steppers were not energized!

    Yes the steppers were not energized. I confirm that when the steppers are energized I have no play at all.

    >>The oval shape that you seem to be achieving, I suspect it is caused by the fact that your “better Z middle” is slightly longer then it should be.
    >>By longer, I mean that your X and Y rail will be pushed out of square, in the quadrant of the Z assembly, I’m guessing by 1 to 2 degrees!

    I fully agrre with you. I’ll print a scaled “better Z middle” and see what happens.

    >>To check this I would move the tool over the same position where you cut that circle/oval. From both ends of the X rail, and on the same side of the
    >>X rail drop a plumb bob down to your work surface/table. Mark the two points on this surface and draw a straight line between these two points.
    >>Now do the exact same thing on the Y rail. Take a square and check where the lines intersect, are they square?

    I didn’t do it but the axes out of sqaure are really evident looking with a square on the tubes.

    Thank you!

    #6155
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    Hi,

    About the removing of the pla is very simple, only remove a little of pla (sending) on the spots at red in the image. But you will only gain at the most haft of millimeter, no more..
    so if you have 1 or 2 mm of gap, this will not correct your problem. I used this trick only to have a better contact of the 608 with the tubes, but i already have contact before doing this.

    It appears you have printed several pla piece in different sizes, i think this was an bad move. This will introduce possible geometry problem that are very difficult to solve. In my opinion the best you have to do is to print again to have all at the same scale.

    I still finishing my own z axis, i hope not have any problem.

    I printing the rigid version, man… that an great piece of engineering!! Really appreciate the work done on this. I ready the cnc won the thingiverse competition. I receive an email to put my printer on the same competition, but the printer is nothing compared with the work done with this cnc.
    Well done Vicious, you have skills.. You need to create an 3d printer.

    #6157
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    I didn’t print several pla pieces in different sizes… I started printing at 100% scale and when I saw the it was not right form my tubes I reprinted the pieces at 98%. The bad move has been not reprinting the better middle z because it seemed to fit well… Now I have to reprint the better middle z and the corner blocks…

    #6167
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    ok

    The re-size of all the pieces will work, but you need to nail the correct value of the scale… and this is an risk.
    Your tube was 24.95mm, It will not be more cheap (counting the pla and time spend printing) and easy to buy an better tube with an perfect 25mm diameter?

    #6170
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    It’s not a risk… 98% on my tube fits perfectly…
    but the 0.05 mm difference doesn’t justify a scale factor of 2% so i think my 3D printer is not well calibrated. I’m a newbye in 3D printing…
    I thought at the option to search a perfect 25mm tube, but now that I know that at 98% all the sizes are correct I prefer reprinting the missing pieces.

    #6174
    Profile photo of vicious1
    vicious1
    Keymaster

    I am working on a printer! What contest?!?!?!?!?!

    I don’t think removing material from that spot will do anything. that will only move the bearings along the conduit tangent, not normal to it. I agree all the same scale pieces, preferably no scaling and just get the right size tube.

    #6175
    Profile photo of vicious1
    vicious1
    Keymaster

    24.95mm would fit the IE version no problem. 0.05mm is not the issue, pretty sure +/- 0.25mm would work just fine.
    Something else is fishy here. plenty of complaints about how tight the parts are, never once has anyone said too loose. 8mmbolts and 22mmOD bearings?

    #6186
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    About the 24.95mm… you are right, 0.05mm is way to small to produce this problem.
    If you calculate the scale in relation to the 25mm you get an 99.8% ((24.95 * 100) / 25), the value is mostly negligible.

    can you give more info about the 3d printer you are working? what type of printer, is an delta? corexy? prusa?
    About the context, some months ago i receive an email to submit the printer in some type of contest, but i never submitted anything because i was sure i will not win. At the time i only uploaded the project on thingiverse, so i was pretty sure it was some type of contest promoted by them. I see on the thingivers page of the cnc you have win an contest and assumed it was the same one.

    #6187
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    my printer is a delta, a geeetech g2 pro.

    #6189
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Yes 22mm od bearings and 8mm bolts.
    I think my printer needs a calibration… but for now reprinting in scale at 98% I think I should get everithing fine.

    #6193
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    An stupid question, your are sure that your printer prints objects with the correct dimensions?

    #6195
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    In fact, assuming that the IE version stl files are correct, and it’s surely so, I think my printer doesn’t print in correct dimensions, but scaling at 98% I find a perfect fit… and so I’ll try to reprint the missing pieces in this way. And then I’ll see.

    I should do a calibration, but if I calibrate now, I will not know if print at 100% or at whatever scale factor.
    Anyway, if reprinting with the actual settings will do the job, ok, if not, I’ll do a calibration and restart printing.

    #6197
    Profile photo of Filipe Campos
    Filipe Campos
    Participant

    you can simply print an 20mm cube calibration to check.. is an print made in less of 20min
    http://www.thingiverse.com/thing:34553

    #6201
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    I’ll do it as soon as possible and I will share the results! Thank you.

    #6236
    Profile photo of Gianluca Magnani
    Gianluca Magnani
    Participant

    Well, I printed the 20mm cube for calibration… It measures 20mm on all the sides… See the pics…
    I don’t know why I need to scale all by the 2%… But the new “better middle z” fits perfectly and now I have the axes in square… I didn’t finish mounting the CNC so I didn’t try a cut yet… Tomorrow I’ll see what happens …

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

You must be logged in to reply to this topic.