effector tilting into bed on print
-
No, the calibration is way off, because the deviation before is always more than 2mm even though auto calibration thinks it has reduced it.
-
Ok,
run this morning.
G32
Calibrated 6 factors using 10 points, deviation before 0.888 after 0.625
G32
Calibrated 6 factors using 10 points, deviation before 1.866 after 0.294
G32
Calibrated 6 factors using 10 points, deviation before 1.121 after 0.343I have saved this into my settings. Something that does not seem right in that I have now issued a homeall and all the carriages go up at the same rate. Z carriage is leading ahead of X and Y. X and Y column limit switches trip around the same time shortly after Z
Once triggered, all three carriages drop by the 5mm as expected. Z bounces off the switch 4 or 5 times on the subsquent rise back to the switches since my effector is no longer centered.
-
I suggest you increase the amount by which the carriages drop after homing to a value higher than 5mm, so that you do not get the Z bounce as the effector is centred.
It still sounds to me that there is some geometrical error that is causing calibration to make excessive corrections. Can you mount a spirit level on the effector, to monitor effector tilt?
-
I can add to the amount the carriage drops. Will that reset the effector back to the center of the bed?
I have no way to do the spirit level, but I did film the effector while running auto configuration.
It 166Mb so I doubt I can attach it.
I can scroll through and select frames though.
-
In looking at the video, I can absolutely state that the effector does not stay level as it moves from point to point. The base of the effector tilts through every movement. Is it possible that 360.21mm magnetic arms are too long to be used?
-
@timvukman can you put the Video in a cloud location so we can see it alternatively can you photograph the effector and carriages and paste that here?
-
@timvukman the length itself is not a factor. The spacing of the rods (pairs must be parallel) and the difference in length of a given pair(lengths must be identical) are critical in avoiding effector tilt.
I agree with the other poster, your next step should be to get the video and clear photos of the effector, carriages and arms uploaded. YouTube is pretty straightforward to upload to, but there are dozens of similar services.
-
Ok:
I have a video uploading to YouTube. It is of the effector while it is executing a G32
Hmmm if this worked, then my video is at https://youtu.be/xvyjY2aBhl8.
Facinating
-
So, now that I have found out how easy it is to post videos on you tube, I will gladly entertain specific requests
for whatever may be helpful to isolate the issues.(personal vote goes toward operator malfunction. RTFM has already proved too hard to handle.)
-
@timvukman can you post a pic that shows both the effector and the carriages it looks like the spacing of the rods maybe different at both the effector and carriage ends and that does produce that tilt effect
-
I can promise you that the rods are the same distance apart at the carriage end as they are at the effector end. That distance is 2 1/4 inches and it is the same at the top and bottom of each pair of rods.
-
@timvukman in that case can you state your Pulley sizes (No of teeth) and post your Config.g
-
Looking at that video, I can see that the effector is tilting a lot. Also, at the first probe point the fan duct fouls the bed (because of the tilt) before the nozzle can contact it.
My first thought was that the spacing of the ball centres was not the same at the top and bottom of the pairs of rods. But if you are certain that it is the same, then I can think of another explanation. You have mounted the ball studs on the sides of the effector, not on the top as is usual. This may mean that the range of movement of the joints is insufficient. Those ball joints can go to nearly 90 degrees away from the aligned position, but I think more than 90 degrees is required if the studs on the effector are mounted horizontally. I suspect that the Delrin rod ends are fouling on the neck of the ball studs at some positions.
-
Good Catch David I never thought of that but it is more than Logical
-
I never considered the mounting of the balls. It certainly makes sense that it could be the problem. I had arrived at the conclusion that the rods must be the issue based on this machine printing very nicely before a couple of rods separated from their fish eye screw sockets.
I have no other option but a horizontal mount for the balls on the effector. To that end, I refurbished all of the rods / sockets and they are now secured with heat shrink tubing. I have replaced the magnetic rods with the originals and I am now working my way through the setup to reset the machine as a new install.
My latest tracking shows the Duet effector arriving on July 31. Until then, I will see what happens with the old rods. That should get us an environment that will confirm your suspicions.
Thanks so much for staying with me on this. -
I am content with the initial settings. The printer will cycle between home and z=0 realiably.
I have determined and saved the trigger height at 0.528. That does change a little on occasion, 0.515 seems to be the other number that pops up.
M500 has been issued.
I believe I am ready to run G32. Agreed?
-
G32 results
G32
Calibrated 6 factors using 10 points, deviation before 1.141 after 1.055
12:02:09
PMG32
Calibrated 6 factors using 10 points, deviation before 3.913 after 0.461Feedback? Possibility of printing?
I can watch the effector through the process and it is gliding like a flying saucer - base is flat.
-
You still have a large calibration deviation. What probing radius are you using? You could try reducing it until you get good calibration results, and then restrict your prints to that radius until you can sport out the effector.
-
@timvukman said in effector tilting into bed on print:
I never considered the mounting of the balls. It certainly makes sense that it could be the problem.
Just my "eyeball" from the vid, so take with a grain of salt: The rods look WAY too long for the other dimensions of the printer.
Curious, what are the rods (ball center to ball center), what are the horizontal extrusions (not counting the corners) and what are the vertical extrusions (again, not counting corners)?
And, if you know, what "brand" of corners?
-
This is looking much better. I took 10 out of the radius for now.
6:24:31 PMG32
Calibrated 6 factors using 10 points, deviation before 0.945 after 0.943
6:22:57 PMG32
Calibrated 6 factors using 10 points, deviation before 0.954 after 0.949
Look better?
