Skip to main content
June 22, 2018 ...

News

  • Registration is currently closed.
    If you wish to register please use the contact form.

Topic: Calibrating towers to the build plate (Read 5951 times) previous topic - next topic

Re: Calibrating towers to the build plate
Reply #15
Here is a link to a video that helps show what I'm talking about in this post about moving the print head to in front of each tower to adjust the height of the extruder

http://101user.com/index.php/topic,130.new.html

  • rfray
  • [*][*]
Re: Calibrating towers to the build plate
Reply #16
Disclaimer: I haven't been able to test this. This is based off of reading about calibration on other delta printers. Give it a go and let us know how it works for you. I'll be able to try it myself this weekend.

So it's important that each tower is calibrated to the build plate surface to assure proper and level printing.
Here is some gcode you can use to move the print head to 4 points (in front of each tower & center) to test if your towers are calbrated correctly.

Easiest way to test that the height is correct is to use a piece of paper and slide it under the nozzle when z=0.
You should be able to feel light drag from the nozzle but the paper should still be able to slide easily under.
If it pins the paper down or if it doesn't touch the paper, you'll need to adjust accordingly.

LocationG-Code (script)
Tower Ag28 g0 f800 x-43.30 y-25 z0
Tower Bg28 g0 f800 x43.30 y-25 z0
Tower Cg28 g0 f800 x0 y50 z0
Centerg28 g0 f800 x0 y0 z0

What does this gcode do? Well lets break it down...
  • g28 homes the printer carriage.
  • g0 is a move to an absolute location.
  • f500 controls the speed of the move. See note below.
  • x, y, and z are followed by the corresponding cartesian coordinates.  The x and y coordinates should be points directly in front of each of the three towers.

So calibrate Tower A until the paper slides between the nozzle and print surface with only light drag. Test a couple times and adjust as needed. Then move on to Tower B and repeat those same steps before finally doing Tower C. Once all three towers have been calibrated, do the same test at the center (0,0). You should be able to now test any of those four locations and have it pass the paper test.  8)

(NOTES: While the 101hero glass plate itself has a diameter of 150mm, My math is based off of using a setting of 50mm as my build radius.
Travel rate is set to 500mm/min, though theoretically it could be higher (upper limit should be: 14mm/s * 60 = 840mm/min). But for now, slower is safer.)

Something I think I discovered while trying this, is you should re-home the extruder (using the back home command) after an adjustment of a screw, before retrying the level command for each tower. I found out the hard way by making several adjustments on the same arm and it didn't seem to be helping.

Re: Calibrating towers to the build plate
Reply #17

Something I think I discovered while trying this, is you should re-home the extruder (using the back home command) after an adjustment of a screw, before retrying the level command for each tower. I found out the hard way by making several adjustments on the same arm and it didn't seem to be helping.

Of course, that's what the G28 at the beginning is for. The printer needs to touch the endstops, if you change the screws, otherwise it does not know about the changed calibration values :) I'ts just a screw, you know...

Re: Calibrating towers to the build plate
Reply #18
It seems that others have this issue too, i.e. when calibrating at the coordinates you provided (diameter 100mm) then you get too much lift-off in the center. Maybe we should calibrate with a diameter of 80mm and stick to smaller prints.....

i.e. something like:

Tower A
G28
G0 F800 X-34 Y-20 Z0

Tower B
G28
G0 F800 X34 Y-20 Z0

Tower C
G28
G0 F800 X0 Y40 Z0


Center
G28
G0 F800 X0 Y0 Z0

How to tell which is tower A, tower B and tower C...............

Re: Calibrating towers to the build plate
Reply #19
It is defined by the coordinates.
Just put the both g-codes (G28..., G0...) in a file "101hero" and run it on the printer.
The first comand does a homing and the second comand drives the hotend to the z-axis zero plane to the appropriate pylone resp. the center of the printer bed.
It is not important to know the pylon number. It is important to have nearly the same distance to the printer bed on each position (about 0.1mm, the thickness of a standard 80g/m^2 paper sheet).
If the hotend is in position, move a little piece of paper between hotend an glass plate with masking tape.
If you feel a little resistance it is perfect, if not turn the appropriate screw to bring it a little bit more down. Press reset and the the process starts again. Repeat this if nescessary.

Hope this helps...
  • Last Edit: 07 Feb, 2017, 17:15:23 PM by TheBartman

Re: Calibrating towers to the build plate
Reply #20
The lift off would be down to the cooling process more so then anything else. If there was a way to get directed air to the part that's just been printed so that the part was cooling is a more uniform way, it would probably take away a lot of the lifting issues. PLA contracts as it cools which is going to cause stress and bowing at the best of times.

I think most people are still using fans mounted at the edge of the printer. This means that there is possibly going to be areas of still air where the flows meet and compete.

My fluid mechanics is quite rusty (18 yrs of rust ::) ) but I'm nearly sure that there is a high probability of dead air in the print area.

(Please note the word rusty, it really has been a long time  ;) )

I'm not talking about the print lifting off the plate, I'm talking about the nozzle not keeping the same distance over the plate, although it's calibrated to be exact near the 3 pylons, it has a bigger gap over the center. It moves in a slight arc with the high-point above 0/0/0.

I never found a solution to the center lifting further off the plate.  Has anyone figured how to overcome the printer's arc?  Mine can squeeze the plate hard enough on the outer edges that no filament will extrude, however lift far enough off the plate in the middle to leave blobs in the air that don't quite adhere.

Re: Calibrating towers to the build plate
Reply #21
have you tried printing with a raft and a skirt? with those two it should compensate enough get okay adhesion and reduce warping quite a bit. have you also tried flipping the glass over? the glass itsself might be warped and causing the weird lifting issues.

  • Admin
  • [*][*][*][*][*]
  • Administrator
Re: Calibrating towers to the build plate
Reply #22
I print with a raft and use glue stick on the tape to keep the print tight on the bed. It has worked for me so far

Re: Calibrating towers to the build plate
Reply #23
Thanks to both of you for the suggestion of the raft.  I had been printing only with a brim.  The raft does indeed help with adherence, despite the uneven travel of the extruder.

Re: Calibrating towers to the build plate
Reply #24
My printer on command is down to z 0,0, reaching a level of about 50 mm from the table. And moves to next command. I can not make it down till glass

  • fedplast
  • [*][*]
Re: Calibrating towers to the build plate
Reply #25
sorry if these are silly questions,
If I want to feed that gcode through my SD card, how do I do it? can I type up a notepad file and rename it? should I basically be doing 4 different files, or is there a pause command where I can give myself 60 seconds to calibrate? (I do not have the dev version, if that matters)
thanks

Re: Calibrating towers to the build plate
Reply #26
I use the code:
Tower A   g28 g0 f800 x-43.30 y-25 z0
Tower B   g28 g0 f800 x43.30 y-25 z0
Tower C   g28 g0 f800 x0 y50 z0
Center   g28 g0 f800 x0 y0 z0
 to calibrate printer, but when I start the center - the head of the printer does not reach the table a few millimeters

  • BobC
  • [*][*]
Re: Calibrating towers to the build plate
Reply #27
It seems there may be problems with some of the configuration constants set in the 101Hero's Marlin firmware.

They apparently entered the arm length from their CAD drawings, and didn't measure it on an actual printer, so the values entered aren't quite right.  If the arms are physically shorter than the firmware is configured for, you can have perfect height at the edges yet be too high in the middle.  Inverse of that when the nozzle has good clearance at the center, but not at the edges.

I believe the most likely cause is that the arms shrunk a bit after coming out of the mold.  It is very common for plastics to shrink by different amounts based not only on the kind of plastic used, but also on the batch of plastic, the mold release used, the mold temperature profile, and even the weather that day.

Arm length error may be visible by very closely watching the print carriage to see if it tilts when away from the center.  If so, the arm lengths likely are wrong.  (Or your pylons are really crooked...)  A good test would be to write some G-Code to move in a 75mm radius circle about 10mm above the build plate.

Arm length error is a common problem that is straightforward to fix on all other Delta printers.  Unfortunately, the 101Hero folks are in violation of the GPL by not releasing everything needed to rebuild the firmware, so we can't fix this error on our own.

One alternative is to replace the arms with ones that are adjustable, then set them to match whatever the firmware expects.

I'm told there are slicers that can account for Delta arm length errors, but I haven't found one yet.

Re: Calibrating towers to the build plate
Reply #28
What arms length needed for correct this problem?

Re: Calibrating towers to the build plate
Reply #29
The delta arm length might indeed be too long. I just spent the whole weekend calibrating my 101hero after a few upgrades (NEMA14 steppers, bowden extruder, frame stabilizations) and finally made it print pretty well.

I'm using Marlin 1.1.5 on the original 101hero controller board, the configuration is available on github:

https://github.com/drdelaney/101Hero-Marlin-Config

I ended up setting DELTA_RADIUS to 78 and the print head was moving in a plane parallel to the print bed. However, printed parts came out a little too wide, so instead of incrementing the radius, the arm length has to be shortened. Setting the rod length to 142 will probably fix my issue. I'm not sure if that was supported in the Marlin build on the stock firmware, but you could try adding

M665 L142

to the gcode.start section of your slicer. The "delta radius" can also be configured using M665 Rxx. Also, the tower height can be set using M665 Hyy, so if you get a vertical offset to the plate by changing the rod length/delta radius, you can also move the whole setup up or down a little without having to screw with the end stop screws.

All parameters can also be set at once with one command, e.g.

M665 L142 H104.8 R78

As a side note, a good reference to measure distance between the nozzle and the print bed is a single row lego brick on the side. Lego bricks are produced very precisely, a single row brick is exactly 7.85mm wide. So if you do something like

G0 F800 X0 Y50 Z7.95

And make the brick fit exactly between pad and nozzle, you'd end up with a 0.1mm gap between nozzle and pad.