SOLVED Calibrating E steps

Discussion in 'Calibration, Help, and Troubleshooting' started by Alex9779, Apr 18, 2016.

  1. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    Hey,
    I am playing with the new extruder design and Titans and this is something I actually had since the beginning with the stock firmware RC3, and all of my following versions.

    I am somehow unable to get a reliable workflow to calibrate E steps.
    I tried several methods but it is irritating...

    Method 1: Using E3D's scripts from the SD and the LCD
    This is annoying because the box always homes and the moves back again and again each execution.
    But the issue is there too.
    What I do is heat up the nozzle. When the nozzle is at temp I mark the position on the filament and start the script.
    I calculate the new E steps and change them with the LCD.
    I execute the script again, the extrusion is still off.
    I execute the script again, the extrusion is different, maybe off, maybe matching.
    I execute the script again, the extrusion is the same as the try before.
    I calculate the new E steps and change the with the LCD.
    I execute the script again, the extrusion is still off.
    I execute the script again, the extrusion is different, maybe off, maybe matching.
    Rinse and repeat until I have it...
    You see my problem? I have to do this at least two time after changing the E steps to get the real result somehow...

    Method 2: Using OctoPrint and LCD
    Same phenomena like in Method 1.

    Method 3 Using only OctoPrint and M92 command
    Still the same but I was not able to get a foreseeable result, the extrusion is always off whatever I do calculate and change.

    I change the script of E3D to only extruder 50 or 10mm. But then I have much more strange results. For example I use the 50mm distance and extrude more times between changing the E steps then I can get to a point where the extrusion seems ok. Then I just use the 10mm script and the extrusion is wrong. I switch back to 50mm and the extrusion is wrong too...

    I do not know what to do to get reliable results...

    Did anyone have a comparable experience?
     
  2. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    The e3d method as you describe works for me.

    Did you remember to "Store Memory" after changing e-steps on the LCD?
     
  3. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    Yes...
    I also used scripts that just extruded and did nothing else... No homing, no setting of M83.
    I tried absolute extrusion with G92 E0 but always the same behaviour.
    The first extrusion after changing the E steps seems to do whatever it likes.
    And switching extrusion lengths sometime also does strange things...
    Strange strange strange...
     
  4. Springfield-Jack

    Springfield-Jack Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    58
    Likes Received:
    4
    I was getting some very inconsistent results until I changed the drivers.It's still not 100% consistent but much better.
     
  5. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    I did not so extensive test before I switched to Titan but now I seem to be unable to get good results.
    The prints I have done are bad.
    I tried for an hour now. I calculated the new E steps for the roll I am using by the E steps I had before and used that as a base.

    Then I extruded 100mm. (462.5 steps/mm)
    The result were 92mm.
    I Adjusted E steps. (502.7)
    The result were 99mm.
    I adjusted E steps. (507.8)
    The result were 92mm.
    I adjusted E steps. (552)
    The result were 92mm.
    WTF??? I knew that could not be but I continued to push the steps up but the result stayed the same...
    I don't know what to do. I tried a print with the E step for the 99mm, it was horrible. Much too much filament...

    Then I tried to do the extrusion in steps. I did 10 times 10mm and the result was different than when I did 1 time 100mm. WTF???
    Can someone tell me what I am doing wrong???

    I tried my hacked firmware and the RC4 of Greg, all the same, no change in behaviour... :confused:
     
    #5 Alex9779, Apr 18, 2016
    Last edited: Apr 18, 2016
  6. Miasmictruth

    Miasmictruth Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    758
    Likes Received:
    112
    I'm sure this isn't it, but how are you measuring the filament?
     
  7. Miasmictruth

    Miasmictruth Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    758
    Likes Received:
    112
    Another stupid question but have you triple checked your not adjusting esteps on the wrong extruder or axis?
     
  8. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    I am using a ruler on top of the extruder where you put in the filament. I put the ruler press on the idler and mark a position on the filament. Then I extrude. With the Titan I did guide the filament a bit so it is centered in the hobb gear. Last times I even pulled the Bowden back over the filament into the idler. When down extruding I measure again the distance of the mark how far it moved.

    I think I am absolutely sure. I heat it. It rotates. It feeds the filament. The other is not moving. Cannot because it is not heated. Marlin prevents cold extrusion if not explicitly activated with M302...
     
  9. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    Oh and yeah I am sure I am using the millimeters scale! ;)

    I did a print now with the the filament I used before the switch to Titan.
    I reduced the E steps now from 462.5 (this is 309 from before Titan) reduced by 10%!!! The results are good no under extrusion.

    My concern is how I should calibrate new spoils if I am unable to measure?
     
  10. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Could there be a Marlin bug that causes it to mishandle numbers above eg 400 such that it was not apparent until Titan?
     
  11. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    I don't know I have to do further testing. Will post some photos. Maybe a video of how I am doing this and what the results are. Maybe someone find my error or is able to reproduce...
     
  12. PsyVision

    PsyVision Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    798
    Likes Received:
    248
    @Alex9779 What motor are you using out of interest? I presume not the one that was originally on the bigbox?

    My e-steps should be ~920 based on information but I'm finding 729 is more reliable in terms of measurement although from last nights print I need to put it nearer 800 for good print results.

    My process is almost the same but I do it via Octo. Heat up, measure 120 on the filament using metal ruler ontop of the extruder. Mark with bluetac, extrude 100, calculate the difference and apply a new e-steps using the LCD. Rinse and repeat. I have had times when this works _really_ well and times when it keeps going way off, as you have seen.

    I wouldnt be surprised if there was some weird Marlin bug. Don't forget also that you are on RC5, I don't know if this has introduced any changes that my have an effect. I don't know the size of the user base for RC5.
     
  13. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    I am using the 1.8° motors that came with the Titan.
    So @PsyVision you are lower than calculated and about 10% too...

    @PsyVision when you set the E steps using the LCD do you save the settings with M500 or using the LCD before the next extrusion? Don't think this should have an influence though, what I know you set things that are used for the current "session", only if you want to preserve them over a reboot then you have to save...

    Can you post your script?
    Here is mine:
    Code:
    G21 ; set to millimeters
    G90 ; use absolute positioning
    T0 ; select extruder 0
    G92 E0 ; zero the extrusion distance
    G1 E100 F400 ; extrude 100mm
    G92 E0 ; zero the extrusion distance
    I am not doing every time the lengthy 100mm extrusion but if I didn't totally fail in maths this should not matter, does it?
    I have no ruler starting at 0 on the edge but this also should not matter.
    Normally I just put one edge of the ruler on the extruder, then I mark at the 100mm mark, extrude 100mm and then the mark on the filament should be at 0mm on the ruler.
    My E steps I calculate with [original E steps] * [target extruded length] / [real extruded length].

    Mayby I am totally wrong but if I extrude 50mm it should be 50mm, if I extrude 100mm it should be 100mm, if I extrude 10mm it should be 10mm. We only use 100 for easier calculations, do we?

    For further troubleshooting I am going to use the firmware Greg posted without any changes. If I do not move XYZ and try to home it should be fine... I don't wanna run through the whole motion system commissioning process to get that bed leveled with autoleveling...

    EDIT: I tried script with "M83" to use relative extrusion too with the same results...
     
  14. PsyVision

    PsyVision Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    798
    Likes Received:
    248
    I don't save settings and I don't use a script, just extrude an amount using the buttons on the Octopi portal. The e3d filament is the only stuff I've really been using so not needed to change e-steps much.

    You should just be able to flash it and then extrude without any homing. I read last night that even after flashing you can have problems so a flush of the eeprom should also be done.
     
  15. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    Did that but I thought script are easier. Tire that too again now.

    Works, for my last tests to figure that out when using a different firmware I always loaded the failsafe defaults and saved the eeprom, so this should be ok. Then I started the tests.

    So here what I did now:
    1. I flashed the firmware Greg posted, which seems to be RC4, loaded the default and saved eeprom.
    2. Default E steps are at 417.5.
    3. I heated the extruder to 245° C, have XT loaded that's why.
    4. I marked 50mm on the filament at the 50mm mark on my ruler.
    5. I extruded 50mm with my script, mark moved to 9mm on the ruler.
    6. New E steps 417.5 * 50 / 41 = 509.1, set with the LCD, saved settings.
    7. I marked 50mm on the filament at the 50mm mark on my ruler.
    8. I extruded 50mm with my script, mark moved too far for a reading.
    9. I marked 70mm on the filament at the 70mm mark on my ruler.
    10. I extruded 50mm with my script, mark moved to 0.9mm on the ruler.
    11. New E steps 509.1 * 50 / 61 = 417.3 (!!!)
    So I was there were I began... I did what I did, I checked I again, same result...
    I tried with different values, if I extrude 100mm the values are always short, my E steps will go to infinite.
    If I extrude 10mm on the stock settings 417.5 I get 10mm extruded!

    This sucks. What's up with that?

    Then I tried only with the LCD, E stepts to 417.5.
    I marked 100mm on my ruler. Moved 50mm with "Move axis", I got 50mm extruded, I moved again 30mm, I got 30mm extruded.
    Then I extruded again with OctoPrint, the extruded value was not correct.

    And here comes the strange thing:

    1. I reset the whole machine.
    2. I heated the extruder.
    3. I set the E steps to something too high, about 500.
    4. Marked the filament and extruded 10mm by moving the axis with the LCD.
    5. The filament moved too far, about 2mm.
    6. Then I changed the E steps back to 417.5.
    7. I went to move axix again, the value was still at 10mm, I moved to 11mm and the extruder retracted! The distance the mark was then was 11mm!!!
    There is something strange going on...
     
    #15 Alex9779, Apr 19, 2016
    Last edited: Apr 19, 2016
  16. Alex9779

    Alex9779 Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    2,405
    Likes Received:
    731
    I am going to open an issue at the Marlin repo to ask what I might be doing wrong. Maybe I didn't understand something.
     
  17. PsyVision

    PsyVision Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    798
    Likes Received:
    248
    I had a weird issue at the weekend where I was asking it to extrude 100mm at a time bringing in filament to the bowden side (as it's hard to reach round the machine and release the idler and feed the filament all the way up the tube) and it randomly retracted 100mm.

    I don't think you're doing anything wrong. To be honest I will usually get a rough figure in place and then just use the extrusion multipliers (on the fly or in slicer) to get it how I want.
     
  18. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    @Alex9779 I think it may be a Maths thing.

    Coming back to your reasoning:

    1. I flashed the firmware Greg posted, which seems to be RC4, loaded the default and saved eeprom.
    2. Default E steps are at 417.5.
    3. I heated the extruder to 245° C, have XT loaded that's why.
    4. I marked 50mm on the filament at the 50mm mark on my ruler.
    5. I extruded 50mm with my script, mark moved to 0.9mm on the ruler.
    6. New E steps 417.5 * 50 / 41 = 509.1, set with the LCD, saved settings.
    7. I marked 50mm on the filament at the 50mm mark on my ruler.
    8. I extruded 50mm with my script, mark moved too far for a reading.
    9. I marked 70mm on the filament at the 70mm mark on my ruler.
    10. I extruded 50mm with my script, mark moved to 0.9mm on the ruler.
    11. New E steps 509.1 * 50 / 61 = 417.3 (!!!)
    Step 6 should probably read:

    417.6 * 50 / 49.1 = 425

    If it moved to 0.9mm on the ruler that means that only 49.1mm were extruded and not 50mm

    So your "e-steps per 50mm" need to increase a bit.

    It's all infra red to me!

    ;-)

    ps as per the manual instructions, it's a good idea to mark a point a bit further than the script, in case it goes too far: so if the script is for 50mm then mark off 60mm with the ruler. At which point the maths becomes 417.6 e-steps per mm theoretical * 50mm extruded in script / 49.1mm extruded in reality. The 49.1mm you will get this time because the mark will have moved to the 10.9mm mark on your ruler, since it started further away.
     
    #18 R Design, Apr 19, 2016
    Last edited: Apr 19, 2016
  19. PsyVision

    PsyVision Moderator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    798
    Likes Received:
    248
    I think he meant 9mm, 0.9mm is quite hard to measure using a ruler.
     
  20. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    Let's see....

    I just spent 2hrs trying to fix something that actually wasn't broken.

    If only somebody would have put me out of my misery!
     

Share This Page