BL(3D) touch woes... Any ideas?



  • So I'm now running the printer and doing tests but I'm getting a few oddities I want to sort out before I move onto the next stages of putting my printer together...

    Now I am using a triangle labs clone, as I can't seem to find a reasonably priced legit BL Touch in the UK. (ie compared to eu and us prices)

    Also I'm typing on my mobile so apologies for any typos or dodgy auto corrects.

    Now first issue with the probe... 1 out of 5 or 10 initial z homings after startup, the probe won't react to reaching the bed, and the bed would crash. However if it does react the probe will seem to run fine and react to all 64odd double provinces of a mesh check...

    I have looked for shorts and can't seem to find any, and my crimps I think are good.

    The second issue is when probing, sometimes the pin would redeploy early and bounce off the bed and retract while the z is still moving away for the next move... It doesn't seem to effect the mesh readings but I'm sure it's not right. It doesn't happen every probe point just maybe a quarter...

    Are my issues down to having a clone? Or a setting? Or software? Is this a known bug?

    The other issue, which might not be the probe, but is puzzling is my bed mesh....

    alt text

    Speaking with kraegar from railcore (which my printer is based off) , hes said he's seen this hump happen with a few people... Is this a bug? My bed that I'm probing is 6mm machine polished cast aluminium. And when I run my steel ruler over what's meant to be a humped zone on its side, I can't see the deviation at all... The thing is flat....
    But according to the mesh there's nearly half a mm difference... If there was, I should be able to see that with the ruler laying against it'll it... This one is puzzling, and I'm not convinced it's the probe giving error ed readings because if it was the mesh points would be noisy, and not produce a regular pattern. Oh and repeatable pattern... Maybe it is warped and I just can't see it 😛

    I also don't think it's my gantry sagging, it's all on twin 10mm rods with quad lm10uu bearings spaced about 6cm apart, and the hump would be far more centered, this hump is more to the back right...

    I'm using mode 9 and on one of the latest release candidates. Not sure what one off the top of my head... Maybe latest.

    Any ideas?




  • administrators

    A hump near the middle of the X axis can be caused by the X rail sagging slightly due to the weight of the carriage on it.



  • i bought and tested one of those clones and found them to be far to inaccurate. (+/- 0.05mm)

    what i did in the end is build a BFPTouch from thingiverse. It uses an optical endstop and is far more accurate. Marlins M48 gave me a SD of 0.002. It is not susceptible to heating being on compared to the bltouch.



  • Thanks for the trouble shoot... But I think I have all points covered. It did give me th link to buy genuine a BL touch, so ordered one...

    As for the hump... Yeah maybe the gantry is sagging 😕 not sure what to do about it. I honestly thought 2 10mm hardened steel with spaced out bearings would stop any sag like that.... Any ideas there?


  • administrators

    @mangy_dog well, that's also what the mesh bed compensation is there to some extent. Once you have the BL touch working repeatability then try the printer with the compensation.



  • Have you tried printing a first layer calibration print to see if the bulge is actually there or not?

    0_1542133661353_bedlevel_nozzle_0.4_200x200-0.3-0.8.stl

    If you print that scaled to your bed size first without compensation active, and then again with it on you should be able to tell if the heightmap is correct or not.



  • Approximately how much does the gantry weigh? Including the bearings, the carriage, hotend, etc. excluding the rails?



  • @bot I don't know, but It has 2 mini turbo fans, 4030 I think... A chimera with volcano blocks. A small but not pancake stepper for direct drive extrusion, a titan like setup (custom design) and the BL touch. 4 ln10uu ball bearings and a slightly wide design thanks to the fans layout. All printed in cf petg.

    Honestly wouldn't have thought that be enough to make the gantry sag.

    alt text

    The span is 500mm..... Maybe that is just enough to make this sag.

    Oh and the sag is real. Did a paper grip test and the hot end pinched the paper very tightly in that jumped area



  • @mangy_dog Sometimes the sag can also be a result of the rods getting bent during a head crash.



  • @mangy_dog I did some quick and dirty FEA in fusion 360, and I was able to elicit 100 microns of displacement with 2 kg on one pair of rails. Steel rails with ABS plastic mounts, over a ~500 mm span

    0_1542215889798_FEA10mmrails.jpg

    Stack two of these on top of each other, and I could understand the displacement you're seeing in reality.



  • Have you done a test print with mesh compensation active? It should have no problem taking care of the distortion.



  • From the mesh image it looks like it's bowing UP not down.
    What do you have on top of the cast aluminum bed?
    My guess is your bed sheet is flexed either from the binder clips or from natural flex when it's heated and being held at the corners.

    Truth be told though I wouldn't worry about it. The bed mesh will compensate for that and your mean error is < 0.1mm. I wish mine was that close. 🙂

    I am also having the same issue that the probe doesn't always deploy when doing a G30 S-1. Mine does complete a bed mesh with no issues though.



  • @kenny66 said in BL(3D) touch woes... Any ideas?:

    From the mesh image it looks like it's bowing UP not down.

    Exactly. When the X axis bows down, the probe will trigger earlier which results in an elevated point
    I would have expected a more "cylindrical" shape of the probing result, though.

    Edit: As it seems to be a CoreXY, there would be bowing on the X and on the Y axis, which explains the result.



  • Could always take the bed sheet off and probe the aluminum directly and see what you get.



  • Hi all sorry for not replying sooner, the probing is direct on the 6mm thick cast aluminium. The machine is still in early testing and as such the beds not even been heated up... I've not even fitted the heat matt yet 😛

    Yeah its core xy, and the rods are alined vertically, with a 6cm space and 6cm separation of 4 bearings, I has thought that the spread load would negate downward stress.

    Anyway, I'm now designing a rail xy that I will fit, replacing the rods... The y will be a 12mm rail bolted to 2020 aluminium. If that ends up sagging I'll cry.... A lot...

    Using mesh compensation might correct the hump to some degree, but I would still know it's off. And it would never be perfect...

    My print carriage I'm certain is under 2kg, heck I think it's under 1... It didn't feel as heavy as a bag of sugar...


 

Looks like your connection to Duet3D was lost, please wait while we try to reconnect.