Crazy flying extruder idea #92
-
Hi DC, wondering if you've put any thought into the fourth axis idea to keep a forth carriage floating somewhere around the same z height as the effector? I've got the mechanical bits all sorted and am going to test it with the setup of just mirroring another tower, but it'll still require the bowden to be pretty long to reach the effector when it's very close to that tower or pretty far from it. I'm tossing around some other ideas too like basically winding it up a long worm gear so it can be suspended from the top and centered closer to the effector without the weight on the carriages which would require the firmware to actually know the z height of the printer to know how high the effector is, but for now I'm just checking in on our earlier 4th carriage ideal! Thanks!
My original plan was to run the extruder drive on a carriage on an additional tower, probably situated midway between two of the other towers. The "diagonal rod length" to the additional carriage would be configured to be a little less than the length of the Bowden tube, and possibly somewhat shorter than the lengths of the real diagonal rods.
The 4th carriage would need its own homing switch. During homing, the motor current for the 4th carriage drive could be reduced, so that it its position is a long way off where it should be, the motor can skip steps in either direction. Then the motor current could be restored before fine homing.
Works for me. How do it do it? haha
-
I saw this post, and I'll throw in my 2c. I implemented a Z traveling tower on my delta – (my delta runs smoothieware) -- with my own code to control the independent axis for Z -- it's original purpose was to be a place where the effector would go to wipe during color changes. I have a silicone sponge there, and in general it worked for the volcano chimera setup, on each tool change the effector would go to the wiper, do temperature change -- cool off the inactive tool and heat up the active tool, and then wipe over the silicone pad. However I ran into an issue which is inherent to chimera style setups -- the inactive head would hit the printed part and cause a layer shift. So basically unless you can get the 2nd nozzle out of the way somehow, there is no way to make this system work. I've since switched to developing my quad idex, which is nearly finished -- doing a true first 4 color print -- a 4 color benchy. But going back to the delta. I wanted to still print multi color on the delta being that I have 2 bondech QR extruders, a flying extruder on rubber bands is out of the question, so I wanted to use the extra tower I have to mount my dual bondechs, and use the distech automation multi material heat break for color changes. I'd be able to shorten my currently 1.2m bowden tubes going into Y coupler ( I ditched the chimera in favor of a single volcano) -- so after brainstorming for a few months this summer, I came up with an idea to implement this "extruder platform" -- but it would not be using the printer firmware -- for 2 simple reasons, the printer doesn't always know where the effector is. It needs to home, and homing the 4th point which is attached to the effector is a recipe for disaster. So my idea involves essentially running the platform stepper using an arduino pro mini (or regular arduino, doesn't matter) and a hall effect sensor to keep track of the distance between the platform and the effector -- there will be a rigid carbon fiber rod going to the effector from the platform, fairly short -- around 100-200mm, basically the length of the bowden tube with the Y coupler and it would be moved by the stepper as the hall effect sensor indicates -- I may need 2, either way the platform must maintain a certain distance from the effector, having it done in printer firmware would be impossible without the additional sensors.
This sounds unduly complicated. You see, smoothie doesn't have a DC42 to help come up with the easiest way to implement his customers ideas
Of course the printer knows where the effector is, otherwise the G-code wouldn't mean anything. G1 Z+1 works exactly because the firmware knows where that is lol. Nonetheless, DCs idea will work out fine, I'm confident! Homing shouldn't be a problem, if it needs to be done at all. Even without a homing switch just push the thing where you want it relative to the effector and hit it with a G92. This would work better if we could use the actual effector Z height instead of a 4th axis rod solution, but it'll still get it close enough, doesn't need to be that precise.Which is another question for you David. would it be possible to average all 3 existing towers? That plus or minus some offset should get us close to the height of the effector!
-
Start with machine off, the firmware has no clue where the effector is, if you pull too hard on the 4th axis, you will rip the machine apart. My idea will work with a delta regardless of the firmware. When you have a heavy extruder platform if you pull to hard on it – it will rip it off the effector, David's idea of using low power is interesting, but it assumes that you will have little differential between the bed and the extruder platform, basically allowing the machine to skip for a bit is not the most elegant idea IMO
-
Start with machine off, the firmware has no clue where the effector is, if you pull too hard on the 4th axis, you will rip the machine apart. My idea will work with a delta regardless of the firmware. When you have a heavy extruder platform if you pull to hard on it – it will rip it off the effector, David's idea of using low power is interesting, but it assumes that you will have little differential between the bed and the extruder platform, basically allowing the machine to skip for a bit is not the most elegant idea IMO
It will only need to skip if the 4th carriage is not in the right position. That would happen if the carriages don't hold position when you power down, or if you transport the printer and that jogs the carriages.
-
That is the point – this intended for a large delta, like mine. My carriages -- under normal conditions park at the top -- engage breaks (servos driven by an arduino) and printer powers down, if the breaks don't engage, the effector will fall down. If assuming normal conditions, we are assuming perfect knowledge in all cases -- I can't always assume that. If I have a hall effect sensor -- -- and I have 1 small tube and a large tube -- and a universal joint at the effector -- I can track small tube going in and out and know if I have to move the extruder axis up or down -- if you can bake that into the firmware - -great, but I plan on implementing it as a standalone feature -- like the piezo board. It will use the sensor to keep the rigid arm between the effector and the extruder carriage at a predefined distance -- it's actually fairly simple, your output of the hall effect sensor must be 512 -- you have a magnet at one end of the tube, and as the value decreases, that means the distance between the extruder and the effector is decreasing, so you move away -- which is up...as the value is increasing, which means the distance is increasing also, so you move the effector down
-
That is the point – this intended for a large delta, like mine. My carriages -- under normal conditions park at the top -- engage breaks (servos driven by an arduino) and printer powers down, if the breaks don't engage, the effector will fall down. If assuming normal conditions, we are assuming perfect knowledge in all cases -- I can't always assume that. If I have a hall effect sensor -- -- and I have 1 small tube and a large tube -- and a universal joint at the effector -- I can track small tube going in and out and know if I have to move the extruder axis up or down -- if you can bake that into the firmware - -great, but I plan on implementing it as a standalone feature -- like the piezo board. It will use the sensor to keep the rigid arm between the effector and the extruder carriage at a predefined distance -- it's actually fairly simple, your output of the hall effect sensor must be 512 -- you have a magnet at one end of the tube, and as the value decreases, that means the distance between the extruder and the effector is decreasing, so you move away -- which is up...as the value is increasing, which means the distance is increasing also, so you move the effector down
My delta is 1m tall with a 350mm bed, and my carriages don't fall when I turn my printer off.
-
That is the point – this intended for a large delta, like mine. My carriages -- under normal conditions park at the top -- engage breaks (servos driven by an arduino) and printer powers down, if the breaks don't engage, the effector will fall down. If assuming normal conditions, we are assuming perfect knowledge in all cases -- I can't always assume that. If I have a hall effect sensor -- -- and I have 1 small tube and a large tube -- and a universal joint at the effector -- I can track small tube going in and out and know if I have to move the extruder axis up or down -- if you can bake that into the firmware - -great, but I plan on implementing it as a standalone feature -- like the piezo board. It will use the sensor to keep the rigid arm between the effector and the extruder carriage at a predefined distance -- it's actually fairly simple, your output of the hall effect sensor must be 512 -- you have a magnet at one end of the tube, and as the value decreases, that means the distance between the extruder and the effector is decreasing, so you move away -- which is up...as the value is increasing, which means the distance is increasing also, so you move the effector down
My delta is 1m tall with a 350mm bed, and my carriages don't fall when I turn my printer off.
My delta is 1.5m tall with 500mm bed, and my do – your point is?
-
It's all about mass and detent torque in motors. If carriages fall consider a good quality metal key chain retractor from top of frame to either effector or wiring bundle which helps guide the wiring into a nice loop above the effector.
-
Just a note, way back when I requested this same idea. I've since given it some thought and decided not to do it, for a couple of reasons. 1 - the bowden angle going into the hotend will be 90 degrees, unless you build slack into it, or actually move the platform at a slightly higher Z height, and turn the extruder on its side, etc. and 2 - the length of the bowden probably isn't any shorter on most printers. If your bed is 300mm in diameter, your bowden will need to be that + enough for the bend at each end (plus some slack, as in (1). and 3) if you do this, when your effector is nearest the extra tower, you now have 300mm+ of bowden hanging in between, possibly getting in the way of the effector.
Unless your delta is very tall compared to the bed diameter it's unlikely to shorten your bowden at all. I've suspended my extruders on my delta, so they hang just above the max homed height of the effector. This is very simple, and keeps the bowden as short as possible.
-
I could see this working with an extruder on its side, on a pivoting bracket so it can swing/turn otherwise you don't really cut the bowden down enough to be worth the trouble. This would get the mass off the effector and towers.
-
Note that the extruder doesn't have to be directly fixed to the carriage…. the fourth tower could reach out over the print area. If if it was just an upside down L that held the extruder fixed downward over the center of the bed and just high enough to be out of the way of any printer moves you could have a fairly short Bowden with no mass on real carriages. The firmware capability is the tricky part. I haven't bothered thinking through exactly how I'm going to work out the mechanics until I have a way of moving a fourth axis pretty in tune with the effector. My other idea is a long worm-gear looking thing that spans across the top of the printer and suspends the extruder, as it rotates the string/wire/whatever you're suspending with follows the path around the gear (at a constant steps/mm) effectively reeling the extruder up the print goes.
-
It's all about mass and detent torque in motors. If carriages fall consider a good quality metal key chain retractor from top of frame to either effector or wiring bundle which helps guide the wiring into a nice loop above the effector.
I have servos which engage pinion gears at the top after the print is done via a custom gcode command – it inserts a break under the carriage, so they won't fall down. I have 400 step steppers, sufficient torque to move the effector rapidly with no issues, and v-wheels, there is no play in the carriage wheels, they move very smoothly -- and they will fall, not fast, but they will fall. The effector mass isn't a lot, but not insignificant, I have a volcano hotened with an aluminum effector, and 3 40mm fans (1 for the heatsink, 2 for part cooling) as well as piezo mount + a Y coupler near the effector -- all contributing to the mass. The carriages themselves are also aluminum, so they're mass is not insignificant. I solved the effector falling issue at the outset of the build using servos. It works very well, start gcode enables the steppers, disanges the breaks, and proceeds with homing. End gcode homes the printer -- waits a beat, engages the servo breaks and disables the steppers. Similar behavior is incorporated in cancel print function.
-
…engages a pinion to effectively apply brakes to carriages...? Sounds pretty trick and cool! Mind to attach some photos to show off this neat design? Looking forward to seeing it.
Have a nice day to all!
-
FYI I've just started extending the delta support to allow up to 6 towers, all with different "rod" lengths.