›

How to fix non-round circle problem

RickMcLeodRickMcLeod
in Repetier-Host
I am running Repetier 0.92M firmware on a XYZ da vinci Duo 2.0 printer.  I re-flashed this printer from the original XYZ firmware about a month ago.  While I have the printer up and running, I had hoped the change of the firmware would allow me to fix issues I had while trying to print out curved chin supports for COVID-19 face shields.  I saw similar issues with the other face shield parts but the "slot" in the chin support is completely unusable.  So I have spent the last month - off and on - trying to fix this.

Since updating the firmware, I have done the following things to try and resolve this problem:
 
1. Calibrated the X & Y steps/mm - now a 20mm cube prints out as 20mm on each side.
2. Checked to see that the extruder carriage butts up against the front belt pulleys - which has not made a difference
3. Found that the X-tensioning part was broken - at this point it is epoxied together, but have ordered bearings and made a replacement part.  Not sure if the new part will resolve the issue I had (and I had no problems printing out the new tensioning part).
4. Rebuilt and uploaded the firmware to enable the backlash compensation - loaded in a value of 0.39mm for the Y axis - but that does not seem to have made a difference.  Not at all sure what value I should use for this backlash, or how to figure out what the right value is.

I found a simple ring object online and I am using this as a test since other objects I've printed over the past several years were all very much linear - no obvious curves.  I have images of the resulting print but do not see a way to include these in my post.  So let me try and describe the result. 

It appears that the inner wall of the circle is actually round.  But when the printer goes to create the outer wall, at the top (furthest from the platform edge) and the bottom (closest to the platform edge) there is an obvious flat spot.  And it looks like the extruder overshoots where it should start following the circle again.  This results in an outside wall that somewhat resembles a lemon.  The peaks (one near the top and one near the bottom) are not in line with the Y axis but is angled 15 degrees or so counter clockwise.  I had hoped that the ability to adjust the backlash would solve this issue (based on some other posts I've seen).  When the printer fills in between the two walls there are gaps between both the outside wall and the inside wall where the outside wall bulges out.  These issues are apparent for each layer of the object (the object I printed has 12 layers).

For all I know, this problem has existed with this printer from day one (I've had it for 6 years or so), but since I have not really printed out many, if any, parts that have curves, I have not noticed this issue.  After going through the process of uploading the Repetier firmare and using the Repetier Host I am getting frustrated that I cannot seem to resolve this issue.


Comments

  • RepetierRepetier
    You can use dropbox/google or bindump or something like that to upload files. Can't really imagine how it looks without image.

    Note with backlash that if you have a line in x direction that should be a curve you have backlash in y direction!

    The typical reasons for circles not being round:
    - Backlash
    - Discretization is done with triangles, so not perfect round anyway
    - Drag of filament narrows circles in general. Depends also on layer height how big the effect is.
    - wear in nozzle might change preferred direction a bit.
  • RickMcLeodRickMcLeod
    No need to upload an image at this point since I have resolved the issue.  And for the sake of any other fool like me who's trying to resolve a similar issue on an XYZ da vinci Duo (or I suspect on the original da vinci as well), I will tell you what I found.

    Initially in talking with some local folks I was told "check that the belts are tight".  I did, or at least I thought I did.  I saw the two "Y" belts on either side of the extruder gantry and the "X" belt going across and attached to the gantry.  I paid close attention to these THREE belts.  I ended up finding that the X belt tensioner part had broken, so I found a replacement object online that used bearings.  So I printed out the object, fitted it with the pulley and the bearings and did a test print of the "ring" object I've been using.

    With the top open to watch the first couple of layers go down I was seeing the same odd behavior.  But then I noticed that there's a FOURTH belt on the da vinci duo.  It is driven by the stepper motor for the Y axis and connects the pulley on the stepper motor shaft with the pulley on the shaft that runs across the back of the printer.  I had never noticed this belt before because I had assumed (and don't remind me what that means...) that the shaft at the back of the printer was driven directly by the Y stepper motor.

    Watching this belt as the printer worked it was stunningly obvious that this belt was WAY LOOSE!  So I killed the print and loosened the Y stepper motor screws, moved the motor so the belt was tight and reran the print.  Lo and behold, the problem I described in my initial post had gone away.  When the object print was finished I measured it.  It turned out to not be exactly round (62mm in one axis and 64mm in the other) the quality of the print was 1000% better than what I had been seeing before.  I will spend some time now fine tuning the printer (maybe removing the backlash on the Y axis to see if I can get the circle to be perfectly round.

    Since I encountered this problem while trying to print out a chin support for a COVID-19 face shield, I then printed out this object and for the first time saw an acceptable print.

    So the moral of the story is to check ALL the belts - especially if you are using a da vinci printer!
Sign In or Register to comment.