1. Got a question or need help troubleshooting? Post to the troubleshooting forum or Search the forums!

Differential IR height sensor, Attemptin' Clone....ain't sure of success...now stage 10

Discussion in 'Mods and Upgrades' started by jim3Dbot, Aug 25, 2015.

  1. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
    -1.85? Why not a positive number?
     
  2. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    Positive brings me too far above the bed. Negative brings me closer. I'm not sure why it's -1.85. That just seems to be the number that puts me near where I need to be.
     
  3. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
    interesting. have you tried measuring the nozzle tip height when you home the Z compared to where it is when it starts to lay plastic? Not sure if you have a feeler gauge or not. I just want to make sure it is actually doing what it should be and going below the home position. Also, can you confirm by saving the model gcode to a file and search for what the M565 command line says? Very odd to me that it is negative, but maybe my setup is special for some reason :)

    HINT: When doing the measurement of the tip to bed once it is actually printing the first layer, I actually had no filament in the machine so then I could measure. Made it a bit easier.
     
    #263 danzca6, Aug 18, 2016
    Last edited: Aug 18, 2016
  4. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    I haven't tried that but I will. It just failed a print and when I restarted, the z-values for the autolevelling sequence were very different to when it began the print it failed previously. This did appear to make a big difference to the next first layer. This was the issue I was facing with the endstop autolevel. Different z-values every time by up to 2mm. It meant I had to tune the z-offset every print and hope that what I set matched what the z-values would be for that autolevel. It took around 35 to 40 minutes to do each time I wanted to print.

    Hopefully I can get this setup giving me consistent values so I don't have to do that anymore.

    I don't have a feeler gauge but I'll give getting the measurements a go anyway!

    I searched a gcode file that I just saved for the M565 command and got this:
    startingGcode,G28 ; home all axes,G1 Z5 F5000 ; lift Z by 5mm,M565 Z1 ; set the offset for auto-leveling mechanism,G29 ; run auto-level,

    and a few lines later: M565 Z1 ; set the offset for auto-leveling mechanism

    I changed it from a negative 1 to a positive 1 last night since you were asking me about it when I posted my start gcode
     
  5. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    Aaaand my PSU has stopped working. Damn. Looks like no more testing for me until I can figure out what happened.
     
  6. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
  7. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    Yeah I know. Probably just bad timing. Believe it or not I've spent almost 1100 euros getting this printer to a point where it almost works to an acceptable level some of half of the time. Could have got an UM2 for that price plus the original cost of the printer! Oh well. I wasn't to know and now I'm too stubborn to give up. I'll get back to calibrating the height sensor in a week or so when I get a new PSU or whenever I can figure out what's going on with the one I have.
     
  8. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,673
    Likes Received:
    7,318
    Sorry to hear that.
    Upgrading the hotend and adding autolevel (I have betas) was enough to get mine to where they work really reliably and consistently.
    They are not as fine a detail as our DLP, but they are fine for what they are.
    They can get you reliable decent prints with the proper tuning.
     
  9. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    Probably just bad luck, Mark. I've had mine for two years now. Any sane person would have given up long ago. Legitimately pumped 500+ hours and all that money in to it. Nothing original except the chasis. It's my job to sell these machines to different companies so it's nice to know the ins and outs of them. I've got quite a few other printers but gosh darn does this one cause me headaches!

    No matter though. People like Jim, danzca and others on this community make my life a little bit easier! I'm sure when I get the sensor dialed in it'll take a major time sink out of the equation.
     
    Geof and mark tomlinson like this.
  10. jim3Dbot

    jim3Dbot Active Member

    Joined:
    Jun 1, 2015
    Messages:
    246
    Likes Received:
    124
    Mr. O, M565 will be a positive offset number, Firmware will be a negative offset number............My offset number is 'firmware'...-0.15mm.

    Shoot for mini ir to trigger, LED solid ON, around 2 to 3mm from bed.....it should already be calibrated out of the box.......if it does not trigger that height, let me know.........I calibrated IR's on the black area of the boro glass, not sure how they trigger on buildtek.
     
  11. jim3Dbot

    jim3Dbot Active Member

    Joined:
    Jun 1, 2015
    Messages:
    246
    Likes Received:
    124
    Hey O, I was typing while you were, so sorry to hear 'bout the PSU, they are fairly cheap on Amazon, I hear.........
     
  12. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    That's ok Jim. My M565 value is 1mm, but I chose that at random to be honest. I thought it controlled how high the z-axis raises between it's two bounces during each autolevel point.

    I'm a little confused about all the offset numbers. So we have four sources of offset numbers?

    1). M565
    2). Firmware offset number
    3). Slicer z-offset.
    4). Z-values generated by the sensor that it then compensates for.

    All of these impact the leveling somehow? I know I'm probably mixing something up.
     
  13. jim3Dbot

    jim3Dbot Active Member

    Joined:
    Jun 1, 2015
    Messages:
    246
    Likes Received:
    124
    I'm thinking just two. M565 & firmware.............Maybe some one can better explain, for example the IR is an inch above the bed......as it drops it will trigger at some distance above the bed..........that distance can be measured..........now if you place that negative number in the firmware then assuming repeat-ability, the probe will touch the glass each & every time, however you don't want to print right on the glass.....so you need to tweak a little less negative for the filament not to get squashed......

    Added info......That is if you are probing over the same point......but since you are probing in other places, that is where ABL comes into play.......
     
    #273 jim3Dbot, Aug 18, 2016
    Last edited: Aug 18, 2016
  14. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    Strange....went downstairs to have a go at finding out why the PSU stopped working only to find the printer powers on fine again. Must be magic.

    Back to cautiously calibrating the sensor. Don't want the PSU to mess me up again.

    And Jim, I sort of get it. I'm not 100% sure how both of them relate to each other in terms of how they work though.
     
  15. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    Does changing the height of the sensor mean I'll have to modify my z-values? As of right now I'm having the same issues I was having with the endstops. Different z-values every autolevel meaning I have to guess at an offset until I get it right every single print.

    If I move the sensor down a little bit, will it throw off anything I've done so far?
     
  16. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
    The sensor should be no less that 1mm and no more than 2mm above the nozzle. I have mine somewhere in the middle of that.
     
  17. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
    So I am puzzled as to why your readings are off each time. Like Jim mentioned before, I wonder if it has anything to do with the reflective properties of the build plate you added vs the stock glass.
     
  18. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
    also...when it homes and sets zero point for the Z during the G28 command is really what matters when it comes to z offset. The G29 command that probes the bed for auto leveling is really just calculating any tilt of the bed across the 9 points. That calculation tells the software to raise raise or lower the z axis to compensate as it prints across the bed. That is supposed to aid in keeping from printing in one corner being to thin and another two thick so that the layers stay nice and even across the bed. So I am not sure why you would need to have to adjust the z offset for every print. Not saying that wasn't helping you with whatever issue you were having, but that shouldn't be the case. Like Jim mentioned, some just set the offset on the firmware and no longer issue the M565 command in their starting gcode. That way if you sliced something and put it on and sdcard as plain gcode you can print it today or a month from now and it wouldn't need to be sliced again. If there was a need to adjust the z offset between those prints, you would just do it in the firmware and all is good.
     
  19. danzca6

    danzca6 Well-Known Member

    Joined:
    Jul 27, 2015
    Messages:
    2,161
    Likes Received:
    1,077
    So, back to why you are getting variances...maybe we should put that in another thread, but have you gone through and made sure every screw and nut is tightened down? Have you lubed your z axis with a light machine oil like 3-n-1 wiping down the smooth rods and oiling them and dropping a few drops on the leadscrews and running the whole z axis up and down a few times to work it in? Just thinking is it something loose or sticking that is putting the two sides off each time. I have maybe .01 mm variance between G29 runs. Very consistent and repeatable. I'd also be curious on what you get with the buildtek off, but not sure how easy it is to remove and if it can be replaced.
     
  20. Oisin

    Oisin Member

    Joined:
    Apr 14, 2015
    Messages:
    384
    Likes Received:
    23
    It was like that when using the stock endstop method of autolevelling too. Different readings every time. I thought it was caused by bed flex but this proves it isn't.

    So the autolevel only compensates tilt? Not bulge or if the bed it tilted in one corner more than the other three etc... Just a straight tilt?

    The reason I have to adjust z-offset for every print is that for every print the printer thinks that the bed has moved up or down (the different z-values for each autolevel). I've never known why it does this.

    For example: I mentioned earlier that the offset that worked was -1.85mm. It has been changed 5 times for 5 test prints since then. If I used -1.85 right now it would not lay the filament on the bed. It's -2.25mm now.

    Since my firmware is the same as yours, danzca, would it be fair to assume this is a hardware issue?

    Come to think of it, the Y axis stepper is very rough to move. Could that be affecting the bed level? It doesn't move smoothly at all by hand.

    .01mm variance between G29 runs rather than 0.7mm or so would be a dream!

    I haven't tried it with the buildtek off as I could never get the PLA to stick to the glass. My heated bed is also very temperamental. The buildtek is easy to remove.

    But yes you're right about creating a different thread. I apologise to Jim for going off topic!
     
    #280 Oisin, Aug 18, 2016
    Last edited: Aug 18, 2016

Share This Page