Print Bed / Y-Rod Alignment - How Accurate is It?
In preparation for my upcoming series on adjusting rods for higher accuracy alignment, I investigated a related issue. How much is the print bed physically skewed from the Y-rod axis? I'm not talking about XY skew as measured by the firmware. This is about how well the Y-carriage bearing mounts are aligned with the print bed's calibration points. The bed can be rotated relative to relative to the Y-rods, but still have zero XY skew. The firmware (and we) assume the Y-rods and the print bed's Y axis are aligned with each other, but are they?
To check the print bed / Y-rod alignment, I used a digital microscope and measured how much x-shift is present between print bed calibration points as the bed moves.
As a pre-requisite, XY-Skew for the printer was low enough for the firmware to call it perpendicular. That should avoid the printer doing corrections to x as y is moved.
The microscope was positioned to capture the left PC board circular traces. I was not moved as the Y-axis was moved.
Y-axis motion was via Octoprint because touching the the control panel knob would move the printer.
I imaged the left front and left rear calibration pads before and after the bed moved 191 mm Y from one pad to another. The top image shows the left edge of left rear pad. Middle is left front pad. Bottom shows size of 1 mm.
The two calibration pads were only 0.144 mm different in x-position as the bed moved 191 mm. That is how much my y-rods deviate from the y-axis of my print bed. It's not zero, but pretty small.
No, you can't correct this by shifting the y-rods. That also shifts the print bed along with the rods. The deviation would remain. Only by correcting the bearing mount points or rotating the print bed relative to the y-carriage can this be changed.
What does this deviation mean? Even when the system thinks the print bed is aligned (or firmware skew compensated) with the x-axis of the printer, moving the bed along y-axis shifts the print bed slightly left/right depending on how far forward or backward the printing is occurring. Resultant x/y squareness will be slightly deviated in prints. This is the ultimate limit of x/y squareness accuracy for my prints.
I probably should also check agains the right side calibration points, but the deviation is pretty small.
Why is a small deviation here so important? This alignment will be an assumed sufficiently accurate relationship when fine tuning y-bushing seating.
Re: Print Bed / Y-Rod Alignment - How Accurate is It?
This variation is probably due to the bend in the rods and the offset of the bearing contact to the bed frame. Just how accurate is the bearing shipped with the printer? Roundness and general machining is specified on fair bearings and amounts to 0.00/-0.018 mm grind accuracy on OD. So there's 0.04 mm without any rod straightness tolerance added. Cheap rods are spec'd 0.05 mm per 100 mm length. So over 210 mm, a bit more than 0.1 mm is possible. Together, that's 0.14 mm of variation just in the spec sheets.