Inductive Z probe trigger point varies with XY position

nbGU

Hello,

This may be a long post about my bed levelling issues because I want to give all of the information. Ultimately I want to know what I have to change to be able to get perfect results from mesh grid bed levelling

My i3 style printer has:

A "mk42" heatbed (9 probe points for an inductive sensor like a Prusa mk2)
A 5v inductive level sensor (from amazon, can find part number if required)
A rigid steel frame with IGUS anodized rails, IGUS bearings, all structural plastic components printed in carbon filled PETG

The printer in general produces great quality prints.

The problem is that my inductive probe trigger point varies depending on XY coordinates, but is repeatable probing the same point.

When I move to my probe points, set the nozzle to hold a 0.1mm feeler gauge, send G92 Z0 and probe with G30 S-1, the trigger height varies by up to 0.19mm from the trigger height in the centre of the bed which is 0.307mm.

I believe this variation in trigger height is the reason that I've never had much luck with mesh bed levelling. Today I set up G32 bed levelling with H offsets and have achieved acceptable, but not perfect first layers across the full area of the print bed.

Additional information is:

My bed is not flat, but I have it as good as I can get it without buying various thicknesses of shim washers.

My X and Y axis linear rails may be slightly bowed, its not visible with the naked eye but I once crashed the nozzle into the bed with enough force to at least elastically deform them.

My inductive probe is 22mm away from my nozzle in the Y direction, 0mm in X direction.

I cant imagine my nozzle/exturder tilting at all as it moves across the bed, my bearing clearance is minimal and the spacing of bearings makes for an extremely solid feeling setup.

Am I missing something? From everything I have described does anything stand out as a reason for Z probe trigger height changing with varying XY coordinates?

I would like solve the trigger point issue to allow me to use mesh grid bed levelling.

nbGU

I think i might have figured it out. If I deformed my X axis rails in a collision and have since rebuilt the printer, they may no longer be bowed in an arc coincident with the Z plane. The top rail could be bowed in the positive Y direction while the lower rail is bowed in the negative Y direction.

This would mean that as my extruder travelled along the X axis, the nozzle would tilt back and forth, rather than side to side, which with my Y offset between the probe and nozzle would cause the variations in trigger height.

Phaedrux

That sounds like a logical conclusion. i3 style printers are actually quite difficult to keep everything square and true.

nbGU

@Phaedrux I'm going to redesign my extruder bracket to get the probe as close as possible to the nozzle, and perhaps accept that I may not ever be able to get a truly perfect first layer across the entire bed with this printer format

Phaedrux

If you want to try eliminating tilt as a test you can set the probe type to P0 for manual probing and then you will be prompted to manually lower the nozzle to touch the bed at each point.

nbGU

@Phaedrux is this to create a mesh grid map rather that using G32 to call bed.g for the old 5 point levelling?

Argo

I think the problem is the Y offset of your probe. I couldn't level my bed properly with (about) 20mm Y offset.
Now my probe has an offset an Y offsest of 0mm (and X 5mm) and the results are much much better then before.

nbGU

@Argo thats a really good point, I will be able to get the probe offset to Y0 but my X offset will then be quite large because of the shape of my extruder (Titan Aero).

As long as I can still get the probe to all of my points with a large X offset I'll give it a try

Phaedrux

@nbGU said in Inductive Z probe trigger point varies with XY position:

@Phaedrux is this to create a mesh grid map rather that using G32 to call bed.g for the old 5 point levelling?

Neither really, using probe type P0 just means that any time a probe is called for you get a manual prompt to job it down. This applies to homing with G30, or doing the mesh with G29, or G32 and bed.g if it contains any calls for the probe.

nbGU

@Phaedrux OK great, really helpful information, thank you for taking the time to help me. I'll do some more tinkering next weekend!

nbGU

Wanted to update this post because i have solved my issue.

I did design and print a new extruder bracket which brought the inductive probe and nozzle closer together, and brought everything closer to the X axis rails. It didn't make much difference.

I found that the probe points on my heatbed (Mk42 from Orballo printing) didn't actually seem to be dead centered in the marked positions on the PCB heatbed, and also suspected that my Amazon probe might not be all that good.

The solution?

Buy the IR probe by DC42
Design a neat bracket for the sensor that sits nicely alongside the heaterblock (tested positions until the heat no longer softened a PLA bracket, then printed it in XT-CF20)
Remove my PEI print surface clamp on a polypropylene build surface.
Quick and easy setup and calibration following the online guides

And the result is absolutely beautiful first layers across the entire print area!