[not yet solved] Tevo Little Monster DuetWifi errors

giostark

Hi,
I learned other stuffs and tried to understand better some functionalities ...
Here what I have done that present the same issue on large print:

I have generated a new bed.g and calculated the H parameter for each point:

G30 P0 X0.00 Y140.00 Z-99999 H0.066 ; 1.316 - 1.303 - 1.291 - 1.310 >>> 1.310-1.244=0.066
G30 P1 X121.24 Y70.00 Z-99999 H0.291 ; 1.512 - 1.537 - 1.549 - 1.543 >>> 1.535-1.244=0.291
G30 P2 X121.24 Y-70.00 Z-99999 H0.178 ; 1.433 - 1.445 - 1.395 - 1.408 >>> 1.422-1.244=0.178
G30 P3 X0.00 Y-140.00 Z-99999 H-0.023 ; 1.206 - 1.206 - 1.250 - 1.225 >>> 1.221-1.244=-0.023
G30 P4 X-121.24 Y-70.00 Z-99999 H-0.244 ; 1.000 - 0.988 -1.006 - 1.006 >>> 1.000-1.244=-0.244
G30 P5 X-121.24 Y70.00 Z-99999 H-0.296 ; 0.938 - 0.950 - 0.944 - 0.963 >>> 0.948-1.244=-0.296
G30 P6 X0.00 Y70.00 Z-99999 H0.057 ; 1.306 - 1.293 - 1.281 - 1.324 >>> 1.301-1.244 =0.057
G30 P7 X60.62 Y-35.00 Z-99999 H0.097 ; 1.331 - 1.350 - 1.356 - 1.328 >>> 1.341-1.244=0.097
G30 P8 X-60.62 Y-35.00 Z-99999 H-0.143 ; 1.107 - 1.100 - 1.100 - 1.100 >>> 1.101-1.244=-0.143
G30 P9 X0 Y0 Z-99999 S7 ; 1.227 - 1.252 - 1.246 - 1.252 >>> 1.244
This is a 5 pass of G32 without a G29 (I have done just a one G29 after the new M665 data inserted in config.g otherwise the nozzle stop 5mm before the bed at any printing)
Calibrated 7 factors using 10 points, deviation before 0.031 after 0.029 (last of 5)
Calibrated 7 factors using 10 points, deviation before 0.032 after 0.030
Calibrated 7 factors using 10 points, deviation before 0.032 after 0.030
Calibrated 7 factors using 10 points, deviation before 0.031 after 0.030
Calibrated 7 factors using 10 points, deviation before 1.487 after 0.029 (first of 5)

Now, The M665 from Eric is different from what the firmware have calculated for my printer. Never mind , I used what the firmware say .
I tried to recheck the Zprobe and all the point with new M665 data and the Hs are more or less identical. I mean 0.005/0.01 of tolerance .

When I try to print large object the segment closed between those point are :
x -80 , y-50
and
x-80 , y+20
are too much pressed on the bed.
Instead the :
x120 , y20
and
x120 , y-50
is too lose from the bed.

So it seems that the printer let stay the print more high where it believe the bed is high and low where the bed is low. BUT the result is that where it believe that is high the print dont stick and where is low the print can barely extrude. So the bed is not slope as it believe.
Can be the S7 parameter in the bed.g the problem?
The H parameter should correct all this but they didnt seems to be adopted during the print ....

Suggestions?

dc42

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

I have generated a new bed.g and calculated the H parameter for each point:

Did you establish the correct H parameters by measuring the trigger height at each point?

If you need to use H parameters to compensate for effector tilt, then you cannot also use G29 mesh bed compensation; because G29 has no facility for making trigger height corrections. But judging from your height maps, you won't need to use mesh bed correction.

giostark

Hi dc42 ,
yep , as the G30s show up, then I have done the average and then subtracted the center probed.. For each point 4x (nozzle touching the glass bed) G92 Z0 + G1 Z5 + M280 P3 S10 I1 + G30 S-1

Azz... I didnt know that.
So Now should be fine launch a M561(clear the actual bed compensation) and then a G29 S2 (disable the bed compensation) ?
I should put this command at the start of the bed.g?

edit:
im trying to do that ..

dc42

Just avoid running G29 or G29 S1 so that you never load a height map. But if you do load one, running M561 or G92 S2 will disable it.

giostark

It seems that the printer ignoring the H value !
I restarted with 6 point for make the things simple . All close to the circumference and the last in the middle.
The end stop are "perfect". Belt well tied. (not too much).
The point are well calibrated directly from the home
G29 disabled and ...
Bed compensation in use: none
Bed probe heights: 0.083 -0.063 0.133 -0.056 0.079
If I run the G32 event 10 times in a raw the result is the same.

I suppose that the H value is more or less ignored because if I revert the + and - of the bottom numbers nothing change.
Also if i leave all point with H0 stock the printed part look like the same.
Rise the Z offset increment the problem.

; bed.g file for RepRapFirmware, generated by Escher3D calculator
; 7 points, 6 factors, probing radius: 150, probe offset (0, 20)
G30 P0 X0.00 Y132.35 Z-99999 H-0.069 ; >>> 1.272-1.341=-0.069 >> ok
G30 P1 X121.08 Y69.90 Z-99999 H0.135 ; >>> 1.476-1.341=0.135 >> very heigh printing on the bed
G30 P2 X129.90 Y-75.00 Z-99999 H0.012 ; >>> 1.353-1.341=0.012 >> heigh
G30 P3 X0.00 Y-150.00 Z-99999 H-0.204 ; >>> 1.137-1.341=-0.204 >> ok
G30 P4 X-129.90 Y-75.00 Z-99999 H-0.341 ; >>> 0.900-1.341=-0.341 >> low
G30 P5 X-121.08 Y69.90 Z-99999 H-0.484 ; >>> 0.857-1.341=-0.484 >> very low printing on the bed
G30 P6 X0 Y0 Z-99999 S6 ; 1.341

The bottom side is too pressed (and gem on the right when rise a little when release the material compressed before) , the top is too high and dont stick...

Any ideas why are ignored that H values? Also if I give a +0.700 or -0.700 to the point that are more high and low , nothing change in the print ...
So I have done tons of time those passages for discover that are not recognized

dc42

It's a known issue that the H value is ignored when deployprobe.g and retractprobe.g files are used. Fixed in 2.02RC4 which I am trying to get finished today.

giostark

Ah
Ok so im not dumb
Now I mount the RC3. I'll be the first on the other side the Atlantic Ocean that will try the RC4
So my brain can stop fry for some hour

giostark

Damn found it... but just now that you told me. Any way thanks for the patience and the support...

@kuhnikuehnast said in Firmware 2.02 Release candidate 3 now available:

Interesting thing I found about G30:

If you use:
G30 X20 Y20 
to probe the bed at a specific point, you always have to calculate the offset of the probe, because the coordinates of the nozzle are used.

Whereas if you use 2:
G30 P0 X20 Y20 H0 Z-99999 
the offset of the probe is already calculated and the real probe coordinates are used but not the "actually" coordinates of the nozzle

The H- parameter is still not taken into account when probing. Therefore, I performed a little test:
M561                                                    ; delete any bed transformation
G1 X45.457 Y8.602 Z10 F15000				; go to probing point on floating bed
G30 X45.457 Y8.602 H0 S-1   				; define this point as Z=0

M561                                                    ; delete any bed transformation
G1 X45.457 Y8.602 Z10 F15000				; go to probing point on floating bed
G30 X45.457 Y8.602 H5 S-1   				; define this point as Z=0

M561                                                    ; delete any bed transformation
G30 P0 X20 Y20 H0 Z-99999 S-1			        ; probe point P0

M561                                                    ; delete any bed transformation
G30 P0 X20 Y20 H3 Z-99999 S-1			        ; probe point P0

M561                                                    ; delete any bed transformation
G30 P1 X20 Y20 H3 Z-99999 S-1			        ; probe point P1
This ended up in the results:
20:20:45G32 bed probe heights:, mean nan, deviation from mean nan
20:20:42G32 bed probe heights: 0.004, mean 0.004, deviation from mean 0.000
20:20:39G32 bed probe heights: 0.007, mean 0.007, deviation from mean nan
20:20:36Stopped at height 1.399 mm
20:20:31M98 P"0:/macros/Test H-Factor"
Stopped at height 1.399 mm
So as the H-factor differs every probe, the results are all the same...

greetings kuhni

giostark

Im still trying...lets see what we reach.
In the same time I bought the original Genuine E3D V6 Hot-End
https://www.amazon.it/gp/product/B07CN8JP47/ref=od_aui_detailpages00?ie=UTF8&psc=1
I was looking the the Pico 3D but on the site is sold out and they never answered to my mail.
I'll let you know how the work proceed.

giostark

So...
Some considerations...

1)There is not any difference if I select S6 or S7-S8-S9 in the bed.g, the print result is pretty the same , also if the parameter auto calculated differ a lot.

I produced right now several bed.g file and all of them have "H" parameter reconfigured. (12 external point and 3 inside or 6 outside and 6 inside)
BUT there is something strange in the procedure.
The trigger height for each point have a reasonable difference between each other. I calculated the difference and added into the bed.g BUT still the point stay too high or too low.
So the trigger eight is not the only parameter that affect the distance from the bed.
I choose to lock the S parameter in the bed.g at 6 because the automatic calculation fix the rod length at 411+ instead than 397+ as I measured with the ruler. (in the original firmware there is settled 397.1073 that confirm the distance).

>>> Concrete example: I'm gone for this point during a test with (6 out points + 6 inner points)
-Removed any old Z offset.
-G1 X112.41 Y64.90 (nozzle down at G1 Z3 from home position, moved to the XY coordinate and with incremental steps dug to the bed)
Then M280 P3 S160 I1 - G92 Z0 - G1 Z5 - M280 P3 S10 I1 - G30 S-1.
The trigger height is 1.320 . Measured several time . Is consistent, all the attempts are more or less the same.
In the center (x0 y0) the trigger height is 1.100
So 1.320-1.100=0.220 The H parameter is H0.220.
BUT if after the calibration (G32x3) I reach that point I can see that there is still 0.250mm from the bed!!! (used a thickness gauge - and if nozzle down 5 times of 0.05 i can touch the bed so is a real 0.25)
The other points do almost the same despite they are too low or high.
So...

WHY?

I'm missing something?
(the endstop are all leveled with digital caliper - The motors are all the same and well wired. I can print a more or less amazing test cube.

Thanks for any suggestion

I understood that the delta printer have those tolerances ..BUT !
If is this the case the trigger procedure as explained dont solve the problem.
There is not solution that go by attempts all the times for each point?

dc42

There was a bug in the firmware whereby H parameters on G30 commands were ignored if deployprobe.g and retractprobe.g files were used, for example for BLTouch. This has been fixed in firmware 2.02RC5.

giostark

I ordered the extension volcano for the original D3D hotend. Both should arrive in this week .
I'll restart the calibration process after I assembled the new components. Don't want do the work twice. I bought also some copper nozzle. Ole!

giostark

Dear absoluT...
I got it....
I made some further change at the setup I'll explain in an other post but some summing here:
-D3D original hotend + volcano extension as the original TLM. + copper nozzle (nickel coated)
-Changed belts with ones iron cored.
-Cartridge modified for a far better precision (with Teflon thickness + bi-adessive scotch) .
-Borosilicate custom made glass (x2) 41,5cmX5mm + 3 auto-build support for block the glass.
-The damn white glue really work! Moisture of 2/5 glue-water , I used just a sponge for apply it. Than the bed at 50 degree. The big piece doesn't move from the glass!

Thanks to any one for the support...for sure I'll ask other stuffs...but right now I'll give back my contribution opening a new thread with my accomplishments.
I'll mark as solved the 3d.

RafB

Hello,

I also have a Tevo Little Monster and I would like some help to understand what I badly parameterized to have this form of card in comparison to that of giostark.
I enclose you bed.g and config.g!
Thanks

0_1549405026893_Capture d’écran 2019-02-05 à 23.13.29.png

0_1549405189432_Capture d’écran 2019-02-05 à 23.19.30.png

bed.g
0_1549405330644_bed.g

config.g
0_1549405365128_config.g

dc42

That height map looks very good to me. Is there a problem?

RafB

Yes, it's the same printer as Giostark so why the map is not more circular as can be seen on the Giostark map above.

I am french and use googletrad and the tracduction of technical words to link to the impression are poorly translated. I still do not understand why my 1st layer is still so bad.
I can tinker with some modify G31 P25 X0 Y20 Z0.84 but it's perfect. I think that on the mechanical side the printer is well tuned but I must not understand something in the calibration procedure.

I do not know what information to bring for help with my problems.

dc42

My guess is that you have a Z probe that is offset from the nozzle and you have a geometrical error that causes the effector tilt to vary with XY position. This causes the apparent trigger height of the Z probe to vary with XY position. It's possible to correct for this using H parameters on the G30 commands in bed.g, but tedious.

RafB

I will try to control all these points but it is not obvious to be precise. By the way, what is the order of magnitude XX.XXXmm?
0_1549987197649_Capture d’écran 2019-02-12 à 16.55.57.png

When I do my calibrated to find the value G31 ZX and that I launch my impression I always 0.20mm too high. So I change the value of G31 Z0.687 + 0.20 to have a 1st layer that hangs and rotates between 0.25 and 0.35mm.

I controlled the distances between the top of the printer and a point of the carts that had different values. I now have about 4cm everywhere I will try to do this with the caliper.

But it did not really change the quality of my first layer (Nozzle 0.8mm)

L'imprimante

RafB

dc42
what should be changed as parts to solve this problem of geometry?

boldnuts

Try this guide and join this TLM group, https://www.facebook.com/groups/TEVO.Monster.Delta.Owners/

0_1550049851528_TLM Ultimate Calibration Tutorial (1).pdf