Recalibrating Z on a Delta
-
Trying to do it manually, I am in this endless loop of changing end stop offsets and height settings using the paper technique at the center, X, Y, and Z towers. Center is nice, move to a tower, it is either too tight or too loose, make adjustment using M666, then center gets loose, adjust height, then towers are tight/loose again.
-
Are you using config-override.g at all?
-
Is there any chance that you can get a piece of glass to fit the build plate? Remove all of the tape and clean the residue off. Paint the bottom side of the glass black and tomorrow night, 22 1/2 hours from now, I'll get on Skype and help you. Let me know your user name.
-
Hi,
I am using glass.
Dont want to paint it. -
I swear there is a bug with the dive height on first run… The first time through on power up, is different than next runs!!! Can someone please confirm?
No bug that I know of, but I suspect that the H parameter in the M665 command in your config.g and/or config-override.g file is a long way from the true value.
-
So what is next?
-
If you don't want to paint the bottom side of the glass, do you have a piece of black paper to put under it? Need to get accurate readings from the sensor.
-
If I was to paint the glass, would I paint the top or the bottom? – Sorry I see you said bottom. Hmm.
Are we saying blue masking tape is not the right type of material to use for this probe?
How about white?
How about kapton tape?
How about some other material to use as a test for the probe?One other thing, my kossel is a mini, it uses 2020 extrusions for towers and 3 wheels for the sliders.
Perhaps I should look to replace those with 4 wheel versions? The three wheel has some play in it.But yeah, let's get good reading from the probe, it may help fix my incredibly high H parameter (dive height).
I still say there's a difference between the first run and future runs. -
It isn't the color of the tape, but just the application itself. If you could get the pieces exactly aligned with each other, with no gaps, no humps, it might work. Plain glass painted black on the bottom, or even a black piece of paper under the glass, is what DC42 has recommended. You don't want anything that might let the IR through to distort the reading.
If you don't have a config-override.g file that you are saving the results to, or don't copy the results to the config.g you will have weird results the first time compared to the next. If you get .300 before and .050 after on the first test, save the results to config-override.g with M500. If you don't save the results, I "believe" the results from the first test are still in memory and so the next test may end up like .060 before and .050 after. I am pretty sure they have to be kept somewhere because you can run a bed level test before a print without anything being saved.
So, what do you want to do? I'd like to help, but don't want to chase my own tail. We could even use Teamviewer to allow me to connect to your computer and control your Duet if you'd like.
-
Blue masking tape should work well with the IR sensor. What the IR sensor doesn't like is transparent surfaces with a light surface below, because then it sees two reflections, one from the top of the transparent surfaces and one from underneath.
-
/cry
-
I offered to help.
-
Update here:
I've worked a bit with Number40fan on a few things..
We found we were not getting a consistent reading from the IR probe when calibrating it and wanted to address this before going further. I replaced the IR probe with a newer version.
Both IR probes' LED will tend to flicker when it is at the trigger point if you let it sit there long enough; is this expected?
I've been working on updates to the effector with the creator, added M3 nut holes for rods to attach to, beveled the rod mounts for more range of motion, created IR probe mount in code rather than copy/paste job. Over the last week I've printed about 5 different effectors.
During my experimentation, I discovered that the IR probe will actually go back to a reading of 1 if the probe gets TOO close to the bed. I think my probe may have been too low compared to the location of the hotend, possibly causing false readings inbetween probe points. (is that even possible?) - Taking the leap of faith on this, I've moved the IR probe mount up 2mm from the previous location, it sits about 4mm above the bed now vs 2mm.
I discovered there are some dimensional inperfections on the printer I am using to print the effector on. That is, when adding the M3 nut holes to the rod mounts, the nuts were a very tight fit on the X axis but the Y and Z axis were just fine. I do not now how this would translate into offsets for an effector, not even sure if it is enough to be concerned with, which is why I mention it here, possibly reprint on another printer and see how it works. I can imagine this being a common issue?? Then I can imagine it not being one too..
Ok back to Reprap firmware…
Number40fan and I made changes to bed.g, config.g files, trying to keep track of everything I may have missed something and find that when I go to kick off a bed leveling job, coordinates that 'should' be reachable suddenly send the head flying off the bed and causing issues with crashing pullies, bad. Here is an example -
Homing the printer then sending the command G1 X0 Y75 Z5 moves the hotend toward the Z tower, ok, now, change this to a bed probe -
G30 P0 X0 Y75 Z-9999 (for example)
and the head will go flying off the print area. I have had to cut it down to Y55 to make it fit, only to find the same problem later on down, when Y was -75.. Y-75 .. poor printer.Tonight I plan to confirm all wires and connections are good and firm, no bad readings.
Any words of support or thoughts are greatly appreciated.
-
It's usual for the IR sensor LED to flicker if you position it right at the trigger height. It's also usual for the LED to go off again if you move the sensor is very close to the bed. Neither of these causes a problem when doing normal bed probing, because each probing move starts at the trigger height plus dive height.
What surface are you using with the IR sensor now: the glass or the blue masking tape?
-
Hi,
I am using blue tape, sanded down.
We replaced the tape, no difference.
We used a piece of dark paper instead, found that we were getting different readings when calirating the probe and stopped there.
The plan is to get some black spray paint and paint the back side of the glass, reprint a new effector, move the IR probe to the Z axis. I did this but the head goes flying off the print bed when calibrating now, think we discovered the issue with this after exchanging some emails – it is the IR probe offsets, I moved the probe to the Z axis and did not re-comission (change probe offsets).
I also replaced the probe with a newer version on this round.At the moment I am more concerned about my printing printing things differently on one axis right now. I am printing on a Cartesian printer, there is no reason for this to happen; will reprint on another printer and see how it goes.
More later...
-
Printed new effector on different printer, everything fits well, moving on with reconnecting nozzles, tubes, fans, heater, temp sensor, ir probe.
Recommission IR probe as described here https://miscsolutions.wordpress.com/mini-height-sensor-board/ ..
Moving head 5mm up, then G30 S-1, several times I get the same result.
Moving head 10mm up, then G30 S-1 several time I get the same result.If I set max height of 355.00, calibrate the IR probe, get a result of 1.757, am I supposed to subtract the 1.757 from the max height of 355.00 and modify the M665 H parameter with the result or leave it at 355 and do a calibration every time? What if I decide to modify the H parameter and it turns out to be higher than what the probe dectects in some areas of my print bed - The hotend would stop, at the max length, eliminatin a flat surface?
At exactly a height of 15mm above print bed, the IR probe reading on screen goes from 1 to 466.
If I leave it there and put a piece of paper on the print bed, the reading starts to give random readings, 210, 293, 466, 350, 341, 420, 455, 466 …Using the webui, If I move to 16mm, step down Z-0.1, AJAX error on the 6th click. Am I clicking too fast?
Anyway, reconnect and the probe display is producing random readings, the height is now 15.30mm above print bed.
Ajax errors keep occurring now, must reset.What's going on?
Done for tonight..
-
Let auto calibration find the correct M665 H parameter for you, then either run M500 to save the M665 and M666 parameters in config-override.g, or run M665 and M666 without parameters and copy those values into config.g. See https://duet3d.com/wiki/Calibrating_a_delta_printer.
-
Ok, pretty sure I know what the wonky IR readings were.. I have a wifi camera aimed at the printer, that camera has a bunch of IR LEDs around it so you can see at night… So I am guessing it is just like shining an IR flashlight on the printbed. LOL
Eliminated.
The G31 X and Y offsets, if the Z probe is on the Z axis, and it is about 20mm away from the hotend, that would the parameters be positive or negative? G31 X-20 Y20 .... or G31 Y20 X20 ...
-
The X and Y parameters in G31 are the offsets of the probe measured from the head reference point (which in a single-nozzle machine is the position of the nozzle). So if the probe is on the +X side of the nozzle, the X parameter will be positive.
-
Generally, calibration issues with the IR probe come down to two issues:
Effector tilt- You have the probe offset from the nozzle by some distance, however due to the geometric shortcomings of your machine, the effector may be tilting very slightly at different x y positions (ie rods not space properly at both ends etc), because the probe is offset and the effector is tilting, it may be closer to the bed at one xy position than another.IR interference - reflective surfaces etc. Sounds like this isn't the issue. I'm sure the experienced gang here will help you sort out the gcode/config confusion you're having, but if that doesn't work out, use a spirit/target level on the effector to see if you're dealing with effector tilt. You can use the H parameter in your bed.g to attempt to calibrate out effector tilt, but I've had mixed results doing so personally, best to address why your effector is tilting if that turns out to be what's happening.
