Seems to be doing something : ))

3mm · 10 May 2018, 06:03

Hi,

Well, not sure, but I think its starting to work, now that I finally got the delta-radius dialed-in and have replaced virtually EVERYTHING. (I wrote AnyPubic, asking them for the printer specs which was missing from the box, but was supplied with raw Marlin code, an Arduino controller, configured for some other vendor's delta, etc. I asked the headqtrs, their answer was, "for another $250 we have new model, fixes all the problems!! I kid you not.) And I have SRD-100 SSR SSR to run the bed heater, which I will install after I return from having gallstone surgery, next week. arrrgg.

So, finally I think I have successfully calibrated the SE (Smart-Effector) Z probe and have successfully run the auto-cal function. But what does this message mean?

"Calibrated 3 factors using 10 points, deviation before 0.108 after 0.107"

Thanks for all the help, tips and advice from (Mysta, Dino, DC42...et al).

3mm

dc42 · 10 May 2018, 10:55

The message means that root-mean-square average height error at the probe points was 0.108mm before that auto calibration, and is predicted to be 0.107mm afterwards. So that auto calibration run was not able to improve the calibration significantly.

That is still a significant calibration error. See https://duet3d.dozuki.com/Wiki/Calibrating_a_delta_printer#Section_Make_sure_your_build_is_accurate for some possible causes.

3mm · 12 May 2018, 04:17

@dc42

Significant? What's good number ratio?

I've checked the parallelism of the towers.
I replaced the plastic corners with aluminum corners
and checked the angles.
I've checked the lengths of 2020 profiles, top and bottom,
they are all the same within 0,1mm.
I've checked the spacing between the towers, --top, middle and bottom, they are all the same within 0,1mm (using a 69,96cm dial caliper).

I've checked the angle between the bed and the towers, 90°.
I've replaced the Heim joints with Smart-Effector/Hayden mag-joints.

I don't know what else to do? Maybe the delta-radius is still wrong or maybe I don't understand the calibration process yet? I'm starting to believe that this AnyPubic delta is snake-bit.

Thanks in advance

3mm

Dougal1957 · 12 May 2018, 06:32

@3mm can you post your bed.g file and the M665 and M666 lines from your config.g and also what is the length of your hand rods (as marked on the stickers on them)

3mm · 13 May 2018, 01:56

@Dougal1957

Hi, thanks.

Delta config:
Originally AnyPubic kossel linear plus, but virtually everything has been replaced, except steppers, belt, top-of-tower end-stop-sw's & 2020 profiles.

DUET3D-EtherNet
Smart-Effector + authentic E3D V6 HotEnd
Metal corners,
360W switching power-supply
SRD-100, 100Vdc @100A SSR for hot bed switching.
RepRapFirmware - Duet 2-E v1.21

I scotch taped the labels so they can't fall-off:
6 rods, grouped by pairs:

288.13
288.12

288.12
288.12

288.11
288.11

config.g:
M665 R139.7 L288.12 B111.76 H215.9 ; Set delta radius, diagonal rod length, printable radius & homed height
M666 X0 Y0 Z0 ; Put your endstop adjustments here, or let auto calibration find them

Thanks

3mm

![0_1526171708494_delta_DUET-E & PanelDUE.jpg](Uploading 100%)

Dougal1957 · 13 May 2018, 06:56

@3mm can you post your bed.g as well please I think I know what may be happening but without that file I can't confirm

3mm · 14 May 2018, 08:31

@dougal1957

Hi,

Perhaps I should also mention that when I run "Mesh Grid Compensation" web-control 'preset', that while it does generate a bed map, the graphic display depicts only the axis and grid overlay with no measurements??

bed.g (text information removed):

M561 ; clear any bed transform

; Probe the bed at 6 peripheral and 3 halfway points, and perform 3-factor auto compensation.

G30 P0 X0 Y109.9 H0 Z-99999
G30 P1 X95.18 Y54.95 H0 Z-99999
G30 P2 X95.18 Y-54.95 H0 Z-99999
G30 P3 X0 Y-109.9 H0 Z-99999
G30 P4 X-95.18 Y-54.95 H0 Z-99999
G30 P5 X-95.18 Y54.95 H0 Z-99999
G30 P6 X0 Y54.9 H0 Z-99999
G30 P7 X47.54 Y-27.45 H0 Z-99999
G30 P8 X-47.54 Y-27.45 H0 Z-99999
G30 P9 X0 Y0 H0 Z-99999 S3

thanks in advance

3mm

Dougal1957 · 14 May 2018, 09:26

yep change the S3 on the last G30 line to be either S6 or S8 and that should improve things no end

dc42 · 15 May 2018, 08:13

@3mm said in Seems to be doing something : )):

Perhaps I should also mention that when I run "Mesh Grid Compensation" web-control 'preset', that while it does generate a bed map, the graphic display depicts only the axis and grid overlay with no measurements??

Hover the mouse cursor over the grid points to see the measurements.

3mm · 16 May 2018, 08:15

@dc42

Hovering over the grid does nothing.

3mm

3mm · 16 May 2018, 08:21

@dougal1957

The numbers changed! Was .109/.107 and now is consistently reading things like:
"Calibrated 8 factors using 17 points, deviation before 0.039 after 0.031"

I presume that these are better numbers than before.

There seems to have been some type of mismatch somewhere, I think it was between the config.g file and the config.json file, as I was seeing via the console, an error msg, when ever I did a calibration, being a G32 'bad command' error. I reran the RepRap wizard and the error msg disapeared. So what does the config.json file do that the config.g file doesn't? Which one is read by the firmware?

Thanks for your help.

3mm

dc42 · 16 May 2018, 09:05

@3mm said in Seems to be doing something : )):

Hovering over the grid does nothing.

3mm

In that image it isn't displaying a height map at all. Did that auto calibration fail?

3mm · 16 May 2018, 21:42

@dc42 said in Seems to be doing something : )):

@3mm said in Seems to be doing something : )):

Hovering over the grid does nothing.

3mm

In that image it isn't displaying a height map at all. Did that auto calibration fail?

@dc42

No, that's what I've been saying. I've not seen anything on that grid in all of the runs that I've done. Its seems like its working ok?? and I'm not seeing any error messages anywhere. Both 'AutoCalibration' and 'Mesh Grid Compensation' macros run the entire process. It creates a height-map .csv file, but nothing is displayed? And hovering over it does nothing, and before I mentioned this matter...well...I tried zooming in, and out, rotating, panning. The grid is empty.

Here's the content of the most recent HeightMap.csv file that I generated today.

HeightMap.csv:
RepRapFirmware height map file v2, mean error -0.001, deviation 0.044
xmin,xmax,ymin,ymax,radius,xspacing,yspacing,xnum,ynum
-100.00,100.10,-100.00,100.10,111.80,20.00,20.00,11,11
0, 0, 0, -0.057, -0.049, 0.052, 0.040, 0.104, 0, 0, 0
0, 0, 0.017, -0.001, -0.008, 0.027, 0.007, 0.023, 0.027, 0, 0
0, 0.053, -0.019, -0.011, 0.020, 0.063, 0.026, 0.067, 0.075, 0.098, 0
0.085, 0.064, 0.037, 0.018, 0.013, 0.017, 0.019, 0.003, -0.096, -0.100, 0.012
0.051, 0.040, 0.012, -0.002, 0.002, -0.001, -0.010, 0.001, -0.013, 0.003, -0.006
0.066, 0.040, 0.017, -0.010, -0.034, -0.050, -0.049, -0.061, -0.058, -0.071, -0.073
0.014, 0.003, 0.003, -0.001, -0.034, 0.015, -0.048, -0.060, -0.043, -0.050, -0.068
-0.013, 0.010, 0.008, 0.041, 0.014, -0.063, -0.074, -0.071, -0.060, 0.025, -0.046
0, 0.041, 0.048, 0.050, -0.026, -0.047, 0.031, 0.032, 0.039, 0.051, 0
0, 0, -0.036, 0.020, -0.055, -0.060, 0.004, 0.004, -0.009, 0, 0
0, 0, 0, -0.032, 0.023, -0.037, -0.044, -0.023, 0, 0, 0

On an unrelated subject...I find it peculiar that the web 'G-Console' displays different messages than does the PanelDUE, in fact, there is an obvious disparity between the two readouts, eg, the web 'G-Console' and the PanelDUE 'Console'. It would be convenient to watch the PanelDUE 'Console' display for messages while the web-interface remains in Machine-Control or another mode.

Thanks...3mm

dc42 · 17 May 2018, 08:47

What browser are you using? Also, please confirm what firmware and DWC versions you are using.

3mm · 17 May 2018, 21:50

@dc42

Hi...

Web-Browser: Mozilla - FireFox 'Quantum' v57.0 (64-bit)

PanelDUE: v1.20(15b1)
DUET-E HW v1.02
Smart-Effector v?

RepRap FW for Duet 2 WiFi/Ethernet
HW Electronics: Duet Ethernet 1.02 or later
FW Ver: 1.21 (2018-03-21)
Web Interface Ver: 1.15a

thanks...3mm
ps...screen grab of 'Machine-Control' page:

dc42 · 18 May 2018, 05:22

You are running a very old version of DWC. Please upgrade it to 1.21.

3mm · 18 May 2018, 05:33

@dc42

I'm confused, I thought I had the latest Firmware...which of these two firmwares are you refering?

FW Ver: 1.21 (2018-03-21)
Web Interface Ver: 1.15a

The web-interface, presume is an internal module of the DUET-E firmware v1.21, do I need to manually update that firmware? eg, 1.15a?

thanks...3mm

3mm · 18 May 2018, 06:53

@dc42

Also, which specific (DUET-E v1.02 HW, SAM4E8E) firmware do you recommend? I'd appreciate the URL location, so as not to make any mistakes.

thanks...3mm

Dougal1957 · 18 May 2018, 06:58

This is the link to the file you need to Download then drag the .zip file (do not unzip it) to the Upload button on the settings page of the web interface

DWC V 1.21

3mm · 18 May 2018, 09:24

@dougal1957 & @dc42

Thanks guys, that fixed the graph anomally!!

What about the latest FW releases? I notice some releases at URL:
https://github.com/dc42/RepRapFirmware/tags
depict a 'verified' tag, while other's don't? What's the difference, are the un-'verified' beta or not-necessarially-stable test releases?

Thanks in advance

3mm