[not yet solved] Tevo Little Monster DuetWifi errors

[dc42]

Yes, using S6 normally improves the delta calibration by calibrating inaccuracies in the tower positions too.

When you run G32 several times in succession, what deviation does it converge to?

[giostark]

@dc42 said in [help] Tevo Little Monster + DuetWiFi + DuetTouch7" errors:

converge

It look like always between max of 0.1x/0.1x with S6.

I tried also S9 and this is the result , of max of 0.1x/0.0x :

With S6 there were still trouble. The print with big base come really attached from one side and few attached to the other.
edit:
Still with S9 the problem remain the same (also if the mesh seems a little better) . On one side of the glass the print stay too high and the other low. Uff...so frustrating. The mesh seems loaded.

Suggestions?

ps: extruder arrived and mounted with 0.9 motor.

[dc42]

What type of Z probe does your machine have? If it's not a nozzle contact probe then the problem may be caused by effector tilt.

[giostark]

@dc42
I have the BL-Touch , original one mounted on the Tevo LM.
Is strange because with the original components there was not those difference. So I can image that is a software problem. If the calibration would be correct the 1srt layer should stick like expected in all its parts...instead it do not.

[dc42]

If it's a BLTouch then the probe must be offset some way for the nozzle, so I am 99% sure that the problem is that the effector tilt is varying with XY position. It's very hard to get the geometry of a delta printer sufficiently accurate to eliminate this completely.

You can test this by measuring the trigger height of the Z probe at different XY positions.

[giostark]

Thanks...I'll check it and I'll report the results...
In the same time I'll try to understand better other stuffs related to this..
https://duet3d.dozuki.com/Wiki/Using_mesh_bed_compensation

[giostark]

I tried make good homework ...but i miss still something.
As the file Eric posted I tried do the same. Now the bed.g look like this:

; 7 points, 6 factors, probing radius: 150, probe offset (0, 0)
G30 P0 X0.00 Y150.00 Z-99999 H-0.080 ; trigger height 1.640 > done
G30 P1 X129.90 Y75.00 Z-99999 H0.320 ; trigger height 1.880 > done
G30 P2 X129.90 Y-75.00 Z-99999 H0.240 ; trigger height 1.800 > done
G30 P3 X0.00 Y-150.00 Z-99999 H-0.020 ; trigger height 1.540 > done
G30 P4 X-129.90 Y-75.00 Z-99999 H-0.262 ; trigger height 1.298 > done
G30 P5 X-129.90 Y75.00 Z-99999 H-0.320 ; trigger height 1.240 > done
G30 P6 X0 Y0 Z-99999 S7 ; Bed center, main trigger height 1.560
; For G30: Use S-1 for measurements only, without calculations. Use S4 for endstop heights and Z-height only. Use S6 for full 6 factors
; If your Z probe has significantly different trigger heights depending on XY position, adjust the H parameters in the G30 commands accordingly.
; The value of each H parameter should be (trigger height at that XY position) - (trigger height at center of bed)

; Bed center, main trigger height 1.560 - sottrarre all'altezza centrale le altezze per ciascun punto e mettere la differenza positiva o negativa come altezza nei vari G30.

I taken 3 times all the 7 point. I commanded down the head close the bed and lowered by 0.05 steps against the glass. I noticed that if i use bigger step the head one time against the glass lift a little on xy , so I used the smallest steps.
And the result is this:

BUT the result in printing is identical.... one side barely stick , the other stick like hell.

I tried also see if the BLtouch was aware of changing bed high. I removed two spacer that sustain the bed (screwing a little) and this is the result:

I'm stuck.

Also if I manually level the bed with spacer to make the calibration look like perfect , I'm sure the printing will not go well. As mentioned the delta printer have more tolerances so the solution should be where you suggest but it didnt work

[giostark]

Eric add those line in the comment of the bed.g :
- entered in config-override.g G31
And this is the config-override.g :
; This is a system-generated file - do not edit
; Delta parameters
M665 L397.048 R156.896 H524.725 B175.0 X0.206 Y-0.111 Z0.000
M666 X-0.295 Y0.233 Z0.062 A0.00 B0.00
; Heater model parameters
M307 H0 A93.8 C266.0 D9.4 S1.00 B0
M307 H1 A372.4 C90.9 D4.4 S1.00 B0
M307 H2 A340.0 C140.0 D5.5 S1.00 B0
M307 H4 A340.0 C140.0 D5.5 S1.00 B0
M307 H5 A340.0 C140.0 D5.5 S1.00 B0
M307 H6 A340.0 C140.0 D5.5 S1.00 B0
M307 H7 A340.0 C140.0 D5.5 S1.00 B0

How he calculated the new parameters for the M665 ?!?!

[T3P3Tony]

@giostark

M665 and M666 values are obtained though using delta calibration. They are written to config-override.g through the command M500, and then loaded when the printer starts up as long as M501 is at the end of your config.g (which it is from your earlier post). You can see what values are currently set from M666 and M665 by typing the commands into the console with nothing after them:

Are you running delta calibration before trying to get the mesh bed compensation working? It looks to me like you are showing pictures of the mes bed compensation but not the results of delta calibration. you may need to run delta calibration a few times to get it to converge.

Use the command G32 in the gcode console a few times and post the results.

[giostark]

Hi Tony,
thanks for your time too...
5 tries with cold bed :
First G32 cold bed
Calibrated 7 factors using 7 points, deviation before 1.534 after 0.000
37 points probed, mean error 0.003, deviation 0.068
Height map saved to file heightmap.csv

Second G32 cold bed
Calibrated 7 factors using 7 points, deviation before 0.033 after 0.000
37 points probed, mean error -0.000, deviation 0.064
Height map saved to file heightmap.csv

Third G32 cold bed
Calibrated 7 factors using 7 points, deviation before 0.052 after 0.000
37 points probed, mean error 0.047, deviation 0.063
Height map saved to file heightmap.csv

Fourth G32 cold bed
Calibrated 7 factors using 7 points, deviation before 0.034 after 0.000
37 points probed, mean error -0.001, deviation 0.059
Height map saved to file heightmap.csv

Fifth G32 cold bed
Calibrated 7 factors using 7 points, deviation before 0.018 after 0.000
37 points probed, mean error -0.002, deviation 0.062
Height map saved to file heightmap.csv

Now 5 tries with hot bed and hot nozzle :
First G32 hot bed
Calibrated 7 factors using 7 points, deviation before 0.130 after 0.000
37 points probed, mean error -0.021, deviation 0.056
Height map saved to file heightmap.csv

Second G32 hot bed
Calibrated 7 factors using 7 points, deviation before 0.066 after 0.000
37 points probed, mean error 0.030, deviation 0.063
Height map saved to file heightmap.csv

Third G32 hot bed
Calibrated 7 factors using 7 points, deviation before 0.055 after 0.000
37 points probed, mean error 0.037, deviation 0.064
Height map saved to file heightmap.csv

Fourth G32 hot bed
Calibrated 7 factors using 7 points, deviation before 0.049 after 0.000
37 points probed, mean error 0.000, deviation 0.057
Height map saved to file heightmap.csv

Fifth G32 hotbed :
Calibrated 7 factors using 7 points, deviation before 0.065 after 0.000
37 points probed, mean error 0.038, deviation 0.062
Height map saved to file heightmap.csv

The M665
Diagonal 407.831, delta radius 159.010, homed height 523.247, bed radius 175.0, X 0.508°, Y 0.646°, Z 0.000°

[giostark]

I have added the M665 to the config.g then I restarted the printer and run twice the G32 with hotted bed , but this doasnt change a thing... the problem still remain the same:

I settled the 0.1 offset. If I set 0.05 i go thorough mouse bite even with the new Bondtech extruder.

Still something is not well settled ...

[T3P3Tony]

@giostark you should run G32 a few times, until it gets as close as possible to converged. Leave the mesh bed levelling out for now.

Once you have it converged as close as possible then you can do mesh bed levelling if you need to. They are not the same thing.

As you are using a BL touch off to one side it is worth manually checking the probe trigger height at each point you are using for G32 in bed.g. Start by confirming that probe correctly sets Z0 in the bed center. Then move to each of the other point and check the Z trigger height. If you see any difference (cause by effector tilt) the capture that in the H offset in the G30 commands in bed.g

Cheers

Tony

[giostark]

Sorry , but is not what I have done already some post above?
After I change the M665 with the above content the new grid look like this . It seems more regular than before:

Sorry again , what you mean as "converged"?
I have done :
1)z-probe for the center point.
2)for all the 7 points taken the trigger height and subtracted it at the center trigger height and added the numbers int the bed.g as above posted.
3)changes the M665 in the config.g with new data calculated by the firmware.
4)run the G32 several times (auto delta calibration) and also runt the auto grid calibration.
5)tried to print with the above results, even with the map of this post.

I have not understood your last sentence....

[giostark]

Is arriving also the BLTouch-Smart V2... lets see if I can solve those problems...

[T3P3Tony]

@giostark I think you are making progress!

Step 4 above: run the G32 several times (auto delta calibration) . That is where you should see the value with deviation before and after reduce after each step once it has reduced enough that it's no longer changing "much" between each step then it's converged. The "much" depends on your system and how repeatable it is.

Please show the output of G32 after running it a few times.

Cheers

Tony

[giostark]

Ok Tony , understood I'll do that soon as possible....
this is what I have done yesterday night.
I removed the B175 from the bed.g accidentally and those was the result:

Then I checked the belt just to be sure they were not too lose and inserted back the B175 and restored the original M665 with standard data (for TLM).
I exchanged the bed point with those :
G30 P0 X0 Y132.28 H0 Z-99999
G30 P1 X121 Y69.86 H0 Z-99999
G30 P2 X129.9 Y-75 H0 Z-99999
G30 P3 X0 Y-150 H0 Z-99999
G30 P4 X-129.9 Y-75 H0 Z-99999
G30 P5 X-121 Y69.86 H0 Z-99999
G30 P6 X0 Y59.15 H0 Z-99999
G30 P7 X56.09 Y32.38 H0 Z-99999
G30 P8 X64.95 Y-37.5 H0 Z-99999
G30 P9 X0 Y-75 H0 Z-99999
G30 P10 X-64.95 Y-37.5 H0 Z-99999
G30 P11 X-56.09 Y32.38 H0 Z-99999
G30 P12 X0 Y0 H0 Z-99999 S9

And the result was this one:
1rst pass

second pass:

Third pass:

Fourth pass:

It seems clear that the system is not accurate. Hope is the BL sensor that make this discrepancy.
The result in printing is the same. One side stick one not.

I'll try with the new BL-touch to do what you suggest. It that should arrive Friday , with the hope that I can come out all this.

[giostark]

@giostark said in [help] Tevo Little Monster + DuetWiFi + DuetTouch7" errors:

Hi ,

Can you tell me if is right this procedure?
I have followed the procedure for calibrate the bed.g
new sensor BLTouch smart V2 original:
For each point I have done those passage
1)G31 Z0 ; clear the previous Z
2)Lift the nozzle close the bed . Then Z-5 to touch the bed, then rise of 5 and start lowering 0.5 to touch the bed again, then rise of 0.5 and start lowering again of 0.005 to touch the bed again. Those steps for 3 time for each point. I had to do this because on delta printer I noticed that the nozzle tend to shift if pressed to much with big steps of lowering. So tried to be the more accurate I could be.
2)G92 Z0 ; set the zero Z
3)G1 Z5 ; rise of 5mm
4)M280 P3 S10 I1 ; release the trigger
5)G30 S-1
6)copy the result for 3 times and take the medium.
7)take the center Z height and subtract each number for each point.

I selected the right number with the ">" , the one of the 3 pass done that is the medium
; 7 points, 6 factors, probing radius: 150, probe offset (0, 0)
G30 P0 X0.00 Y150.00 Z-99999 H-0.124 ; trigger height 1.194 - 1.238 - >1.194> done
G30 P1 X129.90 Y75.00 Z-99999 H-0.430 ; trigger height 1.501 - >1.507 - 1.513 > done
G30 P2 X129.90 Y-75.00 Z-99999 H-0.265 ; trigger height 1.355 - 1.392 - >1.355 > done
G30 P3 X0.00 Y-150.00 Z-99999 H0.023 ; trigger height 1.047 - >1.047 - 1.053 > done
G30 P4 X-129.90 Y-75.00 Z-99999 H0.316 ; trigger height 0.754 - >0.754 - 0.773 > done
G30 P5 X-129.90 Y75.00 Z-99999 H0.370 ; trigger height 0.718 - 0.693 - >0.700 > done
G30 P6 X0 Y0 Z-99999 S7 ; trigger height 1.077 - 1.064 - >1.070 done

first pass:

Calibrated 7 factors using 7 points, deviation before 1.596 after 0.000

second pass :

Calibrated 7 factors using 7 points, deviation before 0.146 after 0.000

[giostark]

After I changed the M665 with new data:
M665
Diagonal 408.220, delta radius 159.047, homed height 524.265, bed radius 175.0, X 0.640°, Y 0.720°, Z 0.000°

First pass:
Calibrated 7 factors using 7 points, deviation before 1.499 after 0.000

So I would ask why the deviation is so different each time?

[giostark]

The F******* problem is still the same.... Im going mad....

[T3P3Tony]

@giostark you do not appear to be running the G32 multiple times. For now ignore the bed mesh compensation (G29) and run G32 a number of time (let's say 5) and report back those results.