What could cause this heightmap pattern
-
Ok, repeated the height map with the glass turned 90 degrees and the pattern features of the map did not rotate with the glass. I am left assuming that maybe there is a slight abnormality in the gantry producing this pattern.
Of more importance to all of this is the tremendous amount of change in the height pattern that is introduced by simple bed clamps. These aren't even heavy bulldog clamps but light picture frame clamps! This is something completely unexpected and I must admit to dealing rather haphazardly with clamps in the past in general. I have been moving them here or there without reason. Sometimes I added a bulldog clamp - random stuff really.
What this test shows is that leaving absolutely EVERYTHING identical between a calibration run and any subsequent prints is CRITICAL !!! Even the lowly glass plate to bed plate can throw the whole works off.This has been most enlightening indeed !
-
I'd be inclined to do some light disassembly of the Y axis and see how it slides.
-
One of the reasons I went to indepandant z motors was that the gantry would slide down way too easy when there was no holding current so I doubt there is a stiff spot. In the great picture of things, the aberrations are immaterial but it was important to me to know why I see them. My curiosity has been satisfied and I will leave 'good enough' alone.
It has been a tremendous learning experience no matter what! -
Are you probing with bed heat on? If so, do you have the B1 parameter in your M558 command, to turn the heaters off during probing? Some bed heaters generate enough magnetic field to affect a BLTouch.
Duet WiFi hardware designer and firmware engineer
Please do not ask me for Duet support via PM or email, use the forum
http://www.escher3d.com, https://miscsolutions.wordpress.com -
If you are going to use clamps you should only use them in 3 places. 3 points determine a plane and 4 clamps over constrains a plane.
Is there anything in your wire loom that is binding the carriage? -
@dc42 , B1 is turned on
-
@4lathe , I am aware of the 3 point thing but the bed is supported at three points so I decided to keep that going to the glass plate.
No on the binding wire loom. -
Where you have the ridges and valleys in the Y direction towards the right hand end of that height map, I would be interested to see what it looks like if you probe that part of the bed with a smaller X spacing.
-
I just started a 9 hr print but I will run a mesh probe tomorrow as per request.
-
Interesting pattern ....! Just not sure what it tells me.
-
Interesting indeed! Did you try rotating the bed 90 degrees?
-
It is a CR10S-5 .... whoa .... the original post vanished
Anyway, aluminum v rails, aluminum bed, glass plate on bed.
Ha .. a new question appears
Yes, tried rotating, no difference. -
I still think your Z axis is stuck due to different movements of left and right. I would put 0.5 kg additional weight on the hotend temporarily and measure again. This will show you the problem, because it will make it worse or better.
-
I will see if there is a way of adding some weight to explore that possibility.
-
To joergs5’s point, can you probe the bed from a different direction in any way than you have been? As in always from the right or in the opposite direction in each row from what you have been been doing. That
might answer the question of whether loading from one side or the other is creating the ridges and troughs. -
@4lathe , the printer uses the standard Duet probing method starting at the front left, going to the front right, doing one step over (depending on probing density), starting on the right and probing to the left.
In other words the probing happens from both sides and no difference is seen in the bed pattern.
The pattern is seen in the Y direction and short of coding a completely manual scan, I have no idea how to tell Duet to scan 'backwards'. -
@jens55 said in What could cause this heightmap pattern:
I have no idea how to tell Duet to scan 'backwards'.
I guess you could swap your coordinate system to make the rear right corner 0,0
-
Great idea Phaedrux. When the probes happen from the opposite direction, if there is indeed a loading or stacking issue, shouldn't the valleys and ridges on the map be reversed?
-
@JoergS5 said in What could cause this heightmap pattern:
I still think your Z axis is stuck due to different movements of left and right. I would put 0.5 kg additional weight on the hotend temporarily and measure again. This will show you the problem, because it will make it worse or better.
Well it's not happening. I added a 600 gr chunk of lead and it promptly got loose . I only lost my BLTouch mount (thank God) in the process. I am not going to risk the glass plate.
-
@jens55 Sorry to hear that! I hope, it's not too much damaged.
The test was intended for:
a) if it gets worse, this may be due to stuck axis because of the difference between the two vertical aluminium extensions are too narrow or too wide. Can you test this manually?
b) if it gets better, the reason may be that you have backlash effects. This means, if Z changes, the movement changes the position at your left and right spindle with play. Alternatively, the connector between stepper and spindle is flexible too, it has backlash also. I don't know a different possibility than to make the hotend heavier to find it out without disassambling and replacing the connector.c) is possible also, as David suggested, that measuting by BLtouch has a problem
In your last screenshot the left side was lower than the right. You can move the left stepper isolated so that the distance of left and right are smaller (lift left side for about 0.25 mm) like described in https://duet3d.dozuki.com/Wiki/Bed_levelling_using_multiple_independent_Z_motors
