Critical Flaw in the MK4S → Core One Upgrade Kit: Don't Reuse Your POM Nuts
Hi everyone,
After four months of using the MK4S-to-Core One upgrade, and after hours of chasing “Uneven Bed” errors and unexplained layer inconsistencies, I’ve identified a significant design/policy oversight in Prusa’s upgrade path that I think users should be aware of.
The issue: transferred mechanical wear
The official upgrade kit instructs you to reuse the two POM trapezoidal nuts from the donor MK4S, while providing only one new nut for the third Z motor. After relatively limited use, I’ve measured 0.2–0.3 mm of axial play in the two reused nuts, while the new one remains rock solid.
Why this is an engineering oversight:
• Architecture change
In a CoreXY machine like the Core One, the Z axis is under constant axial load (supporting the full bed weight) and is subjected to frequent and aggressive Z-hops. This is a significantly more demanding scenario than the MK4S Cartesian setup.
• POM wear and “memory”
POM is an excellent low-friction material, but it is still a consumable. Reusing nuts that have already settled into a specific wear pattern on an MK4S and placing them into a more demanding three-point Z system is a recipe for Z-banding and long-term instability.
• Loadcell interaction
That seemingly small 0.2 mm of axial play is enough to confuse the loadcell during mesh bed leveling. The resulting errors often look like extrusion or calibration issues, when in reality the bed is physically shifting under nozzle pressure.
My advice:
If you are performing this upgrade, do not reuse the Z nuts or the LM8UU bearings. For the cost of roughly $10 in parts, you avoid compromising the precision of a high-end CoreXY machine by recycling components with hundreds or thousands of hours of prior wear.
I’ve sent a formal technical note to Josef Prusa about this. In my professional opinion, saving a few dollars on critical wear components in an official upgrade kit is a mistake that negatively affects long-term reliability.
Check your Z-axis play before you lose your mind troubleshooting the loadcell.
RE: Critical Flaw in the MK4S → Core One Upgrade Kit: Don't Reuse Your POM Nuts
Hmm -- how would an axial play in the POM nuts actually have an effect on the bed position? I would expect that the weight of the bed is always sufficient to let it rest at the bottom end of the tolerance range, i.e. take any play out of the nuts.
My Core One (built from a complete kit, no reused parts) has a similar play in one of the Z nuts. I can hear and feel it when I lift the bed mount in that corner, but I can also hear and feel it drop back to the bottom as soon as I release the mount.
RE: Critical Flaw in the MK4S → Core One Upgrade Kit: Don't Reuse Your POM Nuts
The mechanism is simpler than it might seem:
During Z homing / bed alignment, the bed is driven downward until it hits the mechanical stop. The firmware then continues to drive the Z motors slightly further to make sure the system is fully seated.
At that moment, the trapezoidal nuts are forced against the upper end of their axial clearance.Once homing is finished and the bed is raised again, gravity takes over. Any nut with axial play will settle back to the lower end of its clearance. If one nut has 0.2–0.3 mm of axial play, that corner of the bed will drop by that amount.In a three-point Z system, this directly translates into a loss of bed level, even though everything appeared correctly aligned during homing.
RE: Critical Flaw in the MK4S → Core One Upgrade Kit: Don't Reuse Your POM Nuts
But you weren't talking about bed leveling in your original post, were you?
I agree with the mechanism you describe regarding the leveling. But the frame geometry of the Core One won't be precise to within 0.2 mm anyway. If you really want to get the bed leveled so accurately, you need to install adjustable end stops or shims, and adjust them based on profiling the bed with the nozzle and loadcell. And if you do that, you will correct the effect of the nut's play automatically.