Circles are oval

[vyndalin]

I see problem with bad tighten belt as well (and switching motors leads me to this much more). I had the same problem, but my belts was tightened too much. The easiest way how to fix this, is:
1) switch off the printer.
2) tight the belt, but not too much.
3) Turn the printer (whole printer) to the side. Motor should freely slide (not fall) over the rail by its own weight. If we are talking about the other axis (I just can't remember which one is X and Y, I've no problem with Z ) just push the print bed and it should slide freely as well.

Crapy thing, but it could ruin the day.

[hoolito]

I recall that I was also messing around with the belt tension when the ovalicity problem became apparent...

Could this have something to do with low quality tension belts that stretch under too much tension or creep over time? If the stretching occurs differently per axis the G-code for a perfect circle, driven by firmware with identical X/Y setting of steps/mm would give an oval as a result...

[seaton]

What I would suggest is try printing a calibration piece at different locations on the X and Y axis, i.e. do 4 calibration prints at each corner of the hotbed (think a calibration at each quarter)

I think you might find its's your belt which might have stretched, which will show up as being different measurements at the different locations.

Doing a calibration this way you should be able to find out who axis it is.

I found this quite by accident when my calibrations were out it was where I print all the time, even where I did my calibrations. then by accident I printed elsewhere on the table and it was all correct. Ordered some belts and replaced then all was back to normal.

Regards,

Stephen

[hoolito]

Brilliant. Makes a lot of sense.
I find it difficult to believe that these belts would stretch to the point they cause visible distortion but it would explain this problem and this is a clever way to find out!