Strange mesh grid heightmap

[Nxt-1]

@alankilian said in Strange mesh grid heightmap:

Could you rearrange the points in the bed mapping so it always probes left-to-right or right-to-left and see if the issue goes away?

That would not solve the root problem ie. the backlash. You just would not notice it anymore during probing.

[alankilian]

@Nxt-1 OH! I didn't know you did some experiments that show it's caused by backlash.

Nevermind then.

[Nxt-1]

@alankilian said in Strange mesh grid heightmap:

@Nxt-1 OH! I didn't know you did some experiments that show it's caused by backlash.

Nevermind then.

I seem (mis)interpreted your reply and tought you suggested the rearrangement of probesequence as a permantent solution instead of an experiment. I am sorry to have stepped on your toes.

[BigMoff96]

@alankilian Perhaps it has something to do with the smart effector. Are you using the ball cup arms and injection molded carriages? Not sure how many parts are shared between a V2 and a V3

[alankilian]

No toes have been stepped on man. We're ALL here to get good prints from our printers.

I figured that if you rearranged the order of the points probed so that they always went left-to-right and if you then got smooth results that it would demonstrate that the issue is indeed related to the direction of movement. It seems that you have determined that already.

So, on to your questions:

I have plastic arms and plastic balls that came on the SE300.

I also have "Cheapskate" carriages. HERE are the assembly instructions which should answer your questions.

I guess you could try tightening (or loosening) your belts and see if that makes any difference.

[TLAS]

I’ve seen this effect pretty commonly on my setups.

I attribute it to hysteresis is the measurement system (stain gauges most commonly).

You could resolve this a number of ways, but no one has really thought it important enough to address.

Solutions in increasing goodness:
Option 1: Randomize points
Random movements will spread out the hysteresis effects to a more common natural error.
Option 2: Perform random measurements for the first items in a series of linear points and take a correction factor on all remaining points for the line.
Option 3: Randomize points with multiple measurements at each point.
Increases measurement confidence by mean value distribution reduction.

[alankilian]

@TLAS said in Strange mesh grid heightmap:

I attribute it to hysteresis is the measurement system

I know what hysteresis is, but I can't quite see how it applies to this situation.

Can you help me understand this? It would be wonderful to know more.

I like the idea of multiple samples at the same point. It seems like that should make it go away if it's related to backlash somehow.

Or maybe, when moving to the target point, overshoot, then back up and undershoot by just a little, then move forward and go to the target point.

Since I don't see it, it's hard for me to experiment.

[BigMoff96]

@alankilian sounds like we have the same setup in terms of arms and carriages. I did see something about the ball cups themselves causing binding so I may see about lubricating those and messing with the belt tension.

Is there a way to do multiple probes in reprap? I know how to do it in marlin. Also unsure about randomizing the points.

This is my first duet board and I don’t have much experience with them.

[fcwilt]

@BigMoff96 said in [Strange mesh grid heightmap]
Is there a way to do multiple probes in reprap?

The M558 command takes these parameters to control that:

Annn

Maximum number of times to probe each point, default 1. Maximum, as of 2.03, is 31. Setting M558 A parameter to anything >31 set it to 0 instead of to 31

Snnn

Tolerance when probing multiple times, default 0.03mm

For example, with A5 it will probe at most 5 times until it gets two consecutive readings within the tolerance specified by S.

If it does not get two consecutive readings after the 5 tries it will average the 5 readings and use that value.

---

If you specify S-1 it will force averaging of the A number of readings.

Frederick

[TLAS]

@alankilian
Any multi-axis measurement setup will have some form of ‘memory’ due to the loads last placed on it. When you have a repeatable load pattern (such as between x-axis points), you get fairly consistent results. When you have a different movement (move multiple axes back to the next line), you get different results.

The degree of measurement variability you get will depend on specific contributors of both the motion and measurement systems. Backlash can show as step variation, inertial loads that strain a force sensor can also show up. Creep, temperature drift, etc...

Without solving each of these issues independently, it’s easiest to randomize the motion (turning repeatable movement errors into nearly random errors) and perform more measurements. Might not work as well on the edges though due to less random motion on one side.

@dc42
IMO that’s the biggest flaw in the auto calibration at the moment. There are multiple ‘solutions’ from a calibration that tend to vary between calibrations. While each will produce decent prints, it should be obvious that things like endstop locations aren’t actually changing between runs.

Again, random motion and statistical modeling would likely produce the ‘real’ calibration. Not that it matters too much as the ‘fake’ one-time measurements are good enough to print.