delta auto calibration rams into roof of printer

dinlo

@Phaedrux i have used G30 a few times and it probes the bed properly

Phaedrux

@garyd9 said in delta auto calibration rams into roof of printer:

@Phaedrux said in delta auto calibration rams into roof of printer:

Try removing the second M558 H30 entirely and change the first H30 to H5.

er.. no. That's a safety thing to prevent the nozzle from crashing into the build plate when the H parameter of M665 isn't set properly. Instead of having the nozzle move down at the travel speed to z0 + 5mm (and potentially smashing into the build plate), it leaves a larger 30mm gap before running the G30's.

Thanks.

garyd9

If you know your printer homes properly, then the endstops are working.

After homing (G28), try the following commands to ensure all the movement stuff is going in the proper direction. What all this should do is home, set relative positioning, move the nozzle down 1cm, toward the back of the printer 1cm, toward the left 1cm

G28
G91
G1 Z-10
G1 Y10
G1 X-10

Then edit your bed.g and put a G28 as the second command (after the M561.)

I don't think any of this will solve your problem, but it may point to an issue. I'm still trying to figure out why G32 would cause your nozzle to move UP from a home position.

garyd9

Also, please go into the console for DWC and type the following two commands and copy the results here into the forum. This will tell us if there might be some initial config issue that would be messing up the initial calibration:

M665

M666

(Why use the console and copy the results instead of using values from config files? I have found that getting the "live" data from M665/M666 is more reliable than copying values from config.g and/or config-override.g. The console commands return the values that are actually being used instead of values that are configured to be used.)

dinlo

@garyd9 said in delta auto calibration rams into roof of printer:

M665
M666

M665
Diagonals 440.000:440.000:440.000, delta radius 227.000, homed height 437.000, bed radius 185.0, X 0.000°, Y 0.000°, Z 0.000°
(seems my z height from the previous step wasnt saved?)

M666
Endstop adjustments X0.00 Y0.00 Z0.00, tilt X0.00% Y0.00%
(i assume these values will change only after delta calibration?)

i apologise for the late reply, i have to go to my dads place physically now due to a forced windows update changing some settings i was relying on to remote in.

dinlo

Also, i tried the commands you suggested, it moved in the expected directions but i can't measure how accurate the movement was.

garyd9

@dinlo said in delta auto calibration rams into roof of printer:

@garyd9 said in delta auto calibration rams into roof of printer:

M665
M666

M665
Diagonals 440.000:440.000:440.000, delta radius 227.000, homed height 437.000, bed radius 185.0, X 0.000°, Y 0.000°, Z 0.000°
(seems my z height from the previous step wasnt saved?)

To get settings to save, typically you'd add "M501" to the last comment in your config.g. That tells the duet to load information from a file called "config-override.g". Then, certain settings (heater tuning, delta calibration, etc) are written to that config-override.g file when you send the command "M500."

So, once you get your printer calibration working, you'd probably do a "M500" command afterwards to save the settings (which are then loaded via "M501" every time the machine resets.)

(i assume these values will change only after delta calibration?)

Yes.

Can you confirm that your delta arms are really 440mm long? When you home the printer, is the nozzle really approx 437mm from the bed? (The homed height can be approximate - that will get cleaned up after a calibration.)

Did you make the one change I suggested to your bed.g? (adding G28 to the top of it, after the M561) I doubt it would have made a difference to your problem, but I have to ask: did it help?

dinlo

@garyd9 i measured the arms as best i could with what i have available, i don't have an exact tool to measure them (my longest calipers are 300mm) but believe that's accurate to at least 2 mm, i should be able to make a jig and reglue them if needed.

the 437 comes from me jogging it from homed height to within a millimeter of the bed.

regarding the G28, i made the change and it homed properly before hitting the roof again (now at a lower speed and with M913 on)

garyd9

@dinlo

Can you please describe "hitting the roof" in detail? You have a different machine than I do, so I'm having a hard time understanding concepts such as "going past the endstops" and "hitting the roof"

(On my delta machine, it's physically impossible for the carriages to move up beyond the endstops. Not only is there no more rail for the carriages to ride on, there are physical things in the way.)

Can you describe in detail everything that happens when you send the "G32" command to home and calibrate the printer? My expectation based on your config files (and the modification made to bed.g I suggested) would be:

The 3 carriages ride up to the top of the printer, hit the endstops, move down slightly, move back up again, and then again drop down a bit. (This is the "g28" command in action.)
Then the carriages move down to bring the nozzle somewhat near the bed.
Then the carriages move in such a way to move the nozzle towards the back of the printer, then the nozzle dips down to the bed, and comes back up
Repeat step 3 for each of the 16 points specified in your bed.g.
After all 16 points are probed, the printer stops.

Would it be possible to record a short video showing the printer in action when you send a G32 command and put a link to it here?

(I've never heard of what your describing on a delta printer unless the motors were set up in reverse, but you confirmed that they are moving in the correct direction in an earlier reply.)

Did you ever have the printer running properly with an earlier version of RRF? (You mentioned it came with 2.5.1. Did you get it working properly with that version? If so, do you still have a copy of the config files you used then?)

dinlo

G28 (as stated)
the axes goes up to the endstops (these are optical and are triggered by a shim on top of the carriages, hard to take a photo due to them being black)
they dont stop at the endstops or wait for any other axis (i.e. if x reaches it first, it hits the roof first)
i shut it down

2 included images showing (well, trying anyway) the shim and a carriage hitting the roof

dunno if you can even see it but that black shim on top of the carriage reaches up and hits an optical switch above the roof

when it reaches this point the steppers just grind away until i shut it down

as for earlier versions, no. I immediately upgraded to 3.2.2 since that's what the instructions said

i have the generic config files that came with the board but they were set up for a cartesian.

jay_s_uk

Can't see anything wrong with your config but I did only have a quick look.
My config for the predator can be found here https://github.com/jaysuk/Anycubic-Predator-Duet/tree/SmartEffectorRRF3
just bear in mind that mine had Haydn's delta arms

dinlo

@jay_s_uk oh hey, i must have missed that, i saw your blog just as my fourth hotend blobbed itself to death and it was actually what tipped me over the edge on a duet board (that and the possibility to use the effector+v6 on the pred while the duet goes into a future sk-go). I used your guide for most of the predator-specific wiring (like the mentioned endstops) and it all worked first try. (note: i read through the blog again and realised i somehow missed a lot of updates, having last read about when you used the stock effector)

I will try your config tomorrow and return with the results.

As for the Haydn arms, i am planning to buy some either when shipping isnt mad expensive or when i've saved enough to have $250 just laying around (which might take a while as i'm saving for a house rn)

Veti

check with m119 manually that each endstop is triggerd.
also check that they correspond to the motor assigned to it.
having the endstop on the wrong motor would cause that.

dinlo

@Veti i have done so previously but i will again, never hurts to be sure.

garyd9

I'm at a loss at this point. I've never heard of a duet-based delta properly homing, moving in the proper direction for G1 commands, but moving UP for a bed.g similar to yours.

Perhaps @dc42 can offer some help here. Here's very proficient with delta machines (and obviously he knows more about the Duet than probably anyone.)

Phaedrux

How bout a video of the behaviour?

dinlo

@Phaedrux sure, i'll do it today if i have time after trying out the files from Jay.

alankilian

I installed your homedelta.g and bed.g on my Rostock Max V3.2.

After making all the values in bed.g no larger than 95, I homed and ran the G32 button and everything went along fine. No crashing.

I can't use your config.g unedited because my motors go the opposite direction, but if you still have no progress, I'll try using it and doing the bare minimum of edits to change it to my configuration.

I'm running a Duet2/Ethernet with version 3.2

FIRMWARE_NAME: RepRapFirmware for Duet 2 WiFi/Ethernet FIRMWARE_VERSION: 3.2 ELECTRONICS: Duet Ethernet 1.02 or later FIRMWARE_DATE: 2021-01-05

dinlo

So, i think i've found the area where the issue is.

The machine still did the same thing with the config from Jay. I had only modified one thing to fit my machine: stepper direction. After changing this in the wiring (rather than in config.g) so that it's correct when config.g says they "go forwards", calibration worked and i managed to get off a few probe points before i had a structural failure with the 3D-printed adapter i used (it had some minor stress lines after the failed calibrations and one of these developed into a crack). I'll have to consider whether to get the Haydn arms now or order a print from a printfarm (which will probably cost as much as the arms from what i saw last i checked).

should i mark this thread solved now or wait until i've fully verified that calibration works? (will likely take some time)

edit: thanks @alankilian for mentioning direction just as i was wondering why.

oliof

or you can check if a forum member is willing to print for you ...