Prototype BMG based 'tall' direct drive extruder tool

Discussion in 'Tool heads & ToolChanger' started by Jai Stanley, Dec 12, 2019.

  1. Paul Meyer

    Paul Meyer Well-Known Member

    Joined:
    Dec 18, 2019
    Messages:
    49
    Likes Received:
    8
    Following closely, ready to build another extruder or two. Thanks for the work!
     
  2. garethky

    garethky Well-Known Member

    Joined:
    Oct 8, 2019
    Messages:
    49
    Likes Received:
    28
    I started in on the cad for a mosquito/BMG variant and its going well. There are just fewer tough issues to solve with this configuration. Bonus: a 5015 size part cooling fan just fits off to one side.
     
  3. Paul Meyer

    Paul Meyer Well-Known Member

    Joined:
    Dec 18, 2019
    Messages:
    49
    Likes Received:
    8
    Thanks, Garethky. I want to try a mosquito. Even though v6 works great, I’m tired of them! Been doing them for years, but the pain of getting the heat-break out of the heat-sink without damaging it, or swapping nozzles on a mounted extruder: I’m ready to try something new.

    Keep us posted!
     
    garethky likes this.
  4. garethky

    garethky Well-Known Member

    Joined:
    Oct 8, 2019
    Messages:
    49
    Likes Received:
    28
    Thanks Paul. I'm printing the first prototype now. I decided that I my main goal was solving my cooling issues so I don't want to post files until it passes assembly and turns out a trouble free Benchy. I decided to not design anything new that I could avoid. I modified the E3D parts for both ends of the umbilical. For part cooling I modified the Prusa MK3S nozzle which we know works well. Here are some preview shots from CAD:
    Screen Shot 2020-01-26 at 12.18.49 PM.png
    Screen Shot 2020-01-26 at 12.18.21 PM.png

    Unfortunately, I need to get some 24V 5015 fans before I cant test it printing. It will take at least a week.
     
    #24 garethky, Jan 27, 2020
    Last edited: Jan 27, 2020
  5. Paul Meyer

    Paul Meyer Well-Known Member

    Joined:
    Dec 18, 2019
    Messages:
    49
    Likes Received:
    8
    Two mosquitos (+24v fans) and two BMGs on order. Any obvious bits and pieces needed that aren’t included (other than cooling fan)?
     
  6. garethky

    garethky Well-Known Member

    Joined:
    Oct 8, 2019
    Messages:
    49
    Likes Received:
    28
  7. Paul Meyer

    Paul Meyer Well-Known Member

    Joined:
    Dec 18, 2019
    Messages:
    49
    Likes Received:
    8
    Is that a cost saving measure, or are there parts of the BMG that should be reprinted for weight/area? Since I haven’t used a BMG before, I’m fine with the complete kit for my first stab, but it’ll be interesting if any of the parts can be improved.
     
  8. garethky

    garethky Well-Known Member

    Joined:
    Oct 8, 2019
    Messages:
    49
    Likes Received:
    28
    Yes, it saves cost. I'm not re-using the plastic printed parts from the BMG. But I bought one just to have a look at it.

    My design needs some tweaks but it prints nice Benchys:

    IMG_4978.jpeg IMG_4977.jpeg IMG_0470.JPG IMG_1587.JPG

    This Benchy is better than my "horizontal" tool Benchy in 2 ways. First there is less ringing which means the "tall" tool wobbles less at high speeds . Second the part cooling is much better, overhangs are vastly improved, warping is gone. Bottom line, I can work with this.

    I started with the cooling fan at 50% (the default) but after a few layers I ramped it up to 100% (M106 P6 S255). This resulted in some funky elephants foot business but no artifacts further up so 100% is better. I may also need to tune in the bed temp, I'm using a magnetic layer and textured PEI sheet so there is loss and I had the bed temp set to 75, it wouldn't stay stuck at 65. The fan is louder than my Prusa MK3. This printer is spitting out plastic much faster and I think we have to accept that it needs more cooling to print well.

    This Benchy shows more stringing than my previous prints, not sure why? Maybe the more aggressive cooling is causing the thermistor temps to drop and its ramping up to compensate? Maybe my filament (PLA) needs drying? Maybe retracts need tuning (seems unlikely because this was sliced for the V6 bowden tool)? Anyway, its tunable.

    Tool with the assorted connector bits to make it work (please avert your eyes from the horrific fan wiring ‍♂️):

    IMG_4971.jpeg IMG_4973.jpeg IMG_4974.jpeg IMG_4975.jpeg IMG_4976.jpeg

    I'll work on fixes for the issues I have found and print another tool, maybe even with this tool head!
     
  9. phoenix

    phoenix Member

    Joined:
    Feb 24, 2018
    Messages:
    20
    Likes Received:
    10
    Hi Jai - any update or chance to share your design? Your design seems super interesting and utilizing the spare parts is genius.

    Thanks!
    Pierre
     
  10. mkudzia

    mkudzia Active Member

    Joined:
    Dec 26, 2019
    Messages:
    28
    Likes Received:
    11
    How is it going - thinking of building one tool on BMG + Mosquito...
     
  11. garethky

    garethky Well-Known Member

    Joined:
    Oct 8, 2019
    Messages:
    49
    Likes Received:
    28
    Nothing like a global pandemic to force one to tune in their print settings. My toolhead is making face shield frames for a local effort to supply hospitals. I've switched over to a 0.8mm nozzle for higher flow and faster prints (the Mosquito/Bondtech combo is made for this application!). This print takes about 1:25.

    IMG_5115.jpg

    Stuff I learned along the way:

    * I'm using the Ultistik magnetic sheet on top of the glass plate. The heatbed was too warped to get reliable first layers. The glass plate is MUCH flatter.
    * Mesh bed leveling set to probe 4 times at each point.
    * The whole stackup takes much longer to heat to temp than my MK3 heatbed. It is best to leave it at temp for 5 minutes before starting a print or check it with a temp gun. If not the probed bed shape wont be the same by the time the first layer is underway.
    * The Ultistik sheets are thicker and less warped. That's the one I'm using. It is very slick and doesn't like Windex, cleaning with alcohol and the prints still pop off but the first layer sticks well.
    * I cannot go above 95C on the heatbed thermistor. Any higher and the build plate pops up in places as the bed temp cycles. This could be due to an air bubble or the sheet loosing magnetization. But whatever it is, 95C is the limit for me.
    * I was getting a lot of fouling on the nozzle that caused some print lines to get pulled off the part and fail the print. The following 3 things seems to have cleared that up:
    * Retract distance needs set to 1mm (for the 0.8mm nozzle).
    * Retract speed is 40mm/s
    * I've adopted the reduced material flow rate tweaks in Prusa Slicer. 94% flow rate with:
    M221 S{if layer_height<0.075}100{else}94{endif} D2

    I don't have a lot of time to try every combination of settings as printer time is dedicated to production. Just doing what works for now.
     
  12. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    478
    If you can figure out a rule of thumb for the time to soak the temperature fully into print plate, you could just add that pause to the starting GCODE after the initial GCODE to set the heat bed temp. Something like this:

    Code:
    M190 S95 ; set bed temp to 95 and wait
    G4 S300 ; wait 5 mins to allow full soak
    Alternatively you can do it with probing for an accurate wait. I have a thermistor mounted to the bed, plugged into and define this as a chamber temperature. I then use M191 to wait for the plate to reach temp.

    Code:
    M140 S95 ; set bed temp to 95
    M191 P0 S80; wait for the bed to report sufficient temp to start print
    M141 P0 S0; ignore bed thermistor going forward
    I found that the thermistor attached to the bed never reports the full temperature set for the heater, which I put down to failures of the mounting method, but even so, I have consistency in the readings it gives. Thus, for me, the bed never reports more than 10 degrees below the heat set for the heater. If you go this route, you can observe your temps to get an accurate wait point.

    This is how my bed thermistor is set up in config.g

    Code:
    M308 S6 P"duex.e2temp" Y"thermistor" A"T0" T100000 B4725 C7.06e-8     ; Define thermistor
    M950 H6 C"duex.e4temp" T6                    ; Bed heater
    M141 P0 H6
     
    garethky likes this.
  13. garethky

    garethky Well-Known Member

    Joined:
    Oct 8, 2019
    Messages:
    49
    Likes Received:
    28
    Stuff I learned about the design after putting 3KG of filament through it:

    It's printing just fine up to 10mm^3/s with the pancake stepper at 800ma. Beyond that speed its going to take some tuning, I'm seeing some under extrusion but I think that should be tuneble. But the pancake is enough power to pull the filament at those speeds without it getting crazy hot.

    The stepper is LOUD at 10mm^3/s. It is quiet at a standstill and at the slower speeds on the first layer. But at higher speeds it sounds like a orbital disk sander going over a board! I've tried all of the tuning advice I can find and none of it seems to make this better. I've tried microsteps at 4, 32 (with and without interpolation), 128 and 256. I've also tried different current values and different TOFF settings. This is really, really annoying. IDK if I should try a larger stepper, or go to a Duet3 with stealth chop or what. This is totally the X/Y motors. The little 0.9 pancake is a gem. Dropping main axes current to 1.5A largely solved the problem! (just be prepared to adjust stall detection if using)

    The fan mount needs to be re-designed. It collides with the tool-changer body. I'm not sure what to do here as there isn't a lot of space and I don't want to move the fans location as that would make the fan shroud taller and more annoying to print.

    The axle for the idler drive gear can slip out of the holes its held in. The easy fix is to close off the through hole that allows the idler to be removed without disassembling the whole unit. The diameter of the "C" shaped ider hole also needs to be a little smaller so it has some room to loosen up with use.
     
    #33 garethky, Apr 7, 2020
    Last edited: Apr 8, 2020

Share This Page