RRF3.2Beta4.1 Heightmap vs. Spindle/Driver Mapping

Frederik

@3dML
the old mounting:
img_20201224_101915.jpg)

at the moment mounted with this original springs:

![IMG_20201224_101939.jpg]

The openbuild blocks:

the subframe assembly from the side:

on top is a Prima Creator Magnetic Flex Steel Plate
and the Keenovo Silicone Heatpad below + isolating padding

3dML

@Frederik
My guess is excessive force got applied to the center of the bed at some point and the nature of the L brackets with a single bolt allowed them to swivel inward creating the bend in the bed along the Y axis. I don't have any experience with that style of bed mount so not sure what would be the best recommendation for mounting.

Frederik

@3dML The Bed was shipped with DHL Express from Portugal to Germany. Perhaps it got damaged along the way?

What are the most thermal stable and least problematic solutions for the bed ?

i wanted to print a 3 or 4 point kinematic mount like the Hevort V2 without the need of a subframe . only 3x sfu1204 Spindles and a mgn12 or 15 rail behind to eliminate the ply of the floating mount (without FK/BK Blocks).

option 1: borosilikat plate 400x400mm without flex plate or similar
option 2. Aluminium Plate, 10-20mm thick with precision ground surface to +0.025 to -0.025 and flex plate or something like that

option 3 the same precision plate , only with a DIY coating with Ultem Pellets (PEI) solved in dichloromethane

Frederik

the corresponding .csv values:
xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
10.00,400.00,10.00,390.00,-1.00,43.30,42.20,10,10
0.238, -0.344, -0.824, -1.021, -1.114, -1.004, -0.833, -0.581, -0.169, 0
-0.003, -0.475, -0.694, -0.875, -0.938, -0.924, -0.785, -0.514, -0.146, 0
0.029, -0.307, -0.608, -0.732, -0.866, -0.804, -0.649, -0.469, -0.089, 0
0.132, -0.153, -0.518, -0.618, -0.739, -0.704, -0.600, -0.338, -0.053, 0
0.280, -0.028, -0.245, -0.518, -0.637, -0.599, -0.485, -0.265, 0.003, 0
0.421, 0.117, -0. 114, -0.345, -0.470, -0.528, -0.417, -0.231, 0.083, 0
0.631, 0.253, -0.006, -0.208, -0.333, -0.352, -0.258, -0.115, 0.202, 0
0.842, 0.446, 0.163, -0.082, -0.162, -0.266, -0.206, 0.006, 0.268, 0
1.037, 0.647, 0.309, 0.086, -0.064, -0.107, -0.045, 0.151, 0.456, 0
0, 0, 0, 0, 0, 0, 0, 0, 0,

Thats the last map after i tested a few little things, i thought could help to validate the measurements and improve the probing further:

firmware upgrade to 3.2RC2

Then only for safety
M997 B121
M997 B0

new macros with every BLTouch command including the EEProm Conversion request from open drain to 5v and vice versa
additional, i measured how deep the set crew is on top of the BLTouch and tryed if the ^ on the input pin made a difference in accuracy or error count (error on every probing try or only on every 10th...).

loosen the mounting screws to the subframe, so that the bed was only held by its own weight on top of the springs to release every little bit of mounting stress like you mentioned above (perhaps a force was applied in the center...).

next, i will rotate the bed 90° and 180 ° to see if the Heightmap in DWC is changing accordingly.

If it doesn't change, i try and change driver 0.3 and 0.4 (Hardware wiring) and check what happens with the compensation map

Set the Heater to 120°C, because I thought the thermal expansion could help a little bit, or it flexed while printing PLA with 50°C Bed Temp.
Trying to bring it down on the left lower, upper and right corners with the mounting screws and shim it on the right lower corner

Frederik

i made a few tests with different bed mounts.

when i probe 4 points on the outer corners i get these values:

RepRapFirmware height map file v2 generated at 2020-12-28 19:39
xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
10.00,380.00,10.00,380.00,-1.00,350.00,350.00,2,2
  0.083, -0.042
 -0.082,  0.077

and the corresponding graphic in DWC:

when probing more points i get this values:

RepRapFirmware height map file v2 generated at 2020-12-28 19:52
xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
10.00,380.00,10.00,380.00,-1.00,55.00,55.00,7,7
  0.067, -0.716, -1.026, -1.035, -0.892, -0.648, -0.194
 -0.543, -0.832, -0.956, -1.031, -0.845, -0.609, -0.239
 -0.586, -0.778, -0.884, -0.869, -0.774, -0.577, -0.268
 -0.531, -0.686, -0.745, -0.789, -0.716, -0.558, -0.302
 -0.446, -0.559, -0.581, -0.624, -0.567, -0.474, -0.275
 -0.328, -0.423, -0.451, -0.486, -0.405, -0.334, -0.193
 -0.150, -0.201, -0.192, -0.164, -0.126, -0.115, -0.083

Frederik

and this a second map generated right after the first two:

the values from heightmap.csv:

RepRapFirmware height map file v2 generated at 2020-12-28 20:15
xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
10.00,380.00,10.00,360.00,-1.00,37.00,35.00,11,11
  0.654,  0.049, -0.369, -0.588, -0.709, -0.703, -0.634, -0.514, -0.327, -0.081,      0
  0.105, -0.311, -0.518, -0.634, -0.707, -0.710, -0.635, -0.561, -0.365, -0.162,      0
 -0.046, -0.310, -0.495, -0.571, -0.688, -0.729, -0.654, -0.570, -0.404, -0.206,      0
 -0.118, -0.321, -0.484, -0.575, -0.659, -0.693, -0.621, -0.582, -0.434, -0.243,      0
 -0.121, -0.301, -0.431, -0.529, -0.636, -0.618, -0.610, -0.552, -0.451, -0.276,      0
 -0.113, -0.293, -0.383, -0.503, -0.560, -0.549, -0.601, -0.555, -0.456, -0.266,      0
 -0.110, -0.225, -0.334, -0.405, -0.512, -0.547, -0.524, -0.474, -0.436, -0.339,      0
 -0.051, -0.171, -0.239, -0.339, -0.444, -0.478, -0.474, -0.456, -0.409, -0.342,      0
  0.002, -0.133, -0.213, -0.276, -0.329, -0.369, -0.348, -0.359, -0.339, -0.285,      0
  0.086, -0.059, -0.100, -0.133, -0.233, -0.228, -0.225, -0.246, -0.260, -0.226,      0
  0.116,  0.061,  0.006, -0.048, -0.033, -0.045, -0.047, -0.141, -0.179, -0.155,      0

Phaedrux

The heightmap shows a bowl shape. The 4 point map shows that the bed is basically level.

Now the question is whether the mesh compensation is enough to compensate for the shape of the bed surface.

bedlevel_nozzle_0.4_200x200-0.3-0.8.stl

Print a test file once with compensation active and again with it disabled and see if the first layer is better with it active.

Frederik

@Phaedrux thanks for the file. i will test it and report back whats happening .

the last time i tried a 350x350 X as test, i had to abort the print after i saw that the printed line was getting squished more and more...and before the nozzle starts digging in the bed i stopped the test. that was perhaps 10 cm from the center

parts up to perhaps 8x8cm from the center where i home Z are no problem, but as soon as i try bigger parts with 15x 8 i already can see the height changes in the actual layer

You can see it goog here in the picture below. The actual part is 6.5cm wide

Phaedrux

That amount of brim is going to show some heavy over extrusion no matter what.

Frederik

@Phaedrux the picture was only to visualize the problem. The right side is where the height starts to decrease, and another 3-5 cm to the right the nozzle would hit the bed. Note: I was not at the printer and had no other Pic showing the differences:)

Frederik

@Phaedrux there are a lot of things that I had to improve. at the moment I think a combination of pressure advance 0.45 and FW Retract 6.5mm with 45mm/s solved my extrusion problem quite well.

for information:

the Bowden tube between my Bondtech and the start of the Heat block in the mosquito is 1065mm.

when I am running a g32 and start a print with ~ 75x75mm in the center...everything looks good.

the next thing I will test is your test file with and without compensation activated and with varying probing points around the bed.