Proposal for drip-free dual extruder system

[Tim]

hot damn!! i found it. in the video they have it on an m2

Interesting, I didn't come across that one in a Google search. It doesn't appear to have a motor drive in the extruder, so it's a Bowden drive. Plus it's using a servo and additional electronics, etc.

I have an updated picture, attached. From the comments I made in the last post, I realized that I didn't have to move the extruders so far, nor do they have to be in-line with each other, and by reducing the total travel to about 10mm, the gear rack probably doesn't even need to be curved. My new drawing looks very similar to the existing dual extruder setup, except with the motors and extruders not directly attached to the rail, but connected as before through a hinge at the top. There are a lot of little differences, though.

With the extruders only at a 5 degree angle, it could probably be prototyped with the existing hot-end mounting bracket. That would put the nozzles at a slight angle to the plate when printing, but should still work reasonably well.

[sprior]

Is there enough power to drive both filament stepper motors at the same time? The thing that makes me nervous about this setup is that you rely on the non-active filament drive stepper motor to be actively braking to keep things in place. In fact I'm guessing that there would need to be a 0 position sensor switch and this mechanism would have to zero out on initialization like the XYZ directions do. I'm also guessing that the Z adjustment procedure would involve loosening/moving away both wipe plates, moving each extruder into position and adjusting it, then parking each and adjusting the wipe plate to match. The person who forgets to move away the wipe plates step probably ends up being unhappy.

[Tim]

I considered that it might require a home position switch for the rotation. If so, it should need only one. The cable assembly that comes with the dual extruder has two unused wire pairs (I'm using one of those pairs for LED lights under the extruder fan, but still have one pair spare), and the RAMBo has additional sockets for switch maximum positions, so it's possible to implement. But it complicates the system by requiring firmware changes, which I'd like to avoid. Another possibility I thought of is that you could allow more room past the gear racks on either side, and as an initialization step, just roll the gear all the way up past the end, let gravity position it against the end of the gear rack, and then roll it back a fixed amount. There might be a mm or so of play in that maneuver, but that's a fixed offset set at the beginning only.

The main thing I'm concerned about with this design is that the position of the gear teeth of the two extruders relative to each other needs to be known exactly so that they both engage the gear racks properly for the short period that they are both engaged at the same time. If not, then you'd have to turn one motor off to let it shift into position relative to the other, but the extruder stepper motors have a high gear ratio and I don't think they will shift that easily.

[sprior]

Firmware changes would be needed for sure to implement the stepper motor for the inactive extruder being locked. I don't know if there are any gear tooth shapes which would allow a free to turn stepper engage the gear when it was disengaged.

[Tim]

Firmware changes would be needed for sure to implement the stepper motor for the inactive extruder being locked.

I'm not sure that's true; isn't the stepper motor going to remain stationary as long as it's not being used?

I don't know if there are any gear tooth shapes which would allow a free to turn stepper engage the gear when it was disengaged.

In that case, you have to know the exact position of the stepper, which might be a problem. And if that doesn't work, then the whole rotation thing would almost certainly have to be done with a separate servo, and that would almost certainly be violating some Stratasys patent. Of course, as long as it's just a DIY modification, that doesn't really matter. But I'm still trying to keep it as simple as possible. If it's too complicated, expensive, or doesn't improve the existing architecture (especially that last one), then there's not much point.

[sprior]

No, you'll notice that when your printer is idle you can move the head from side to side, but if you tell the printer to home you can't. Is be pretty sure the inactive extruder isn't being kept in the brake mode.

[jsc]

Really? I would think the extruder motors are always powered, and always holding. I haven't checked this, but that is my belief.

[Tim]

you'll notice that when your printer is idle you can move the head from side to side, but if you tell the printer to home you can't.

I know that's true for the axis motors, and it may or may not be true for the extruder motor (I expect that it is true, because otherwise it would be virtually impossible to do a cold pull), but it doesn't matter, because like homing the axes, you would home the rotation at the start of the print (using the start g-code block), and the extruder stepper motor should be active from then until the end of the print. There are no individual timeouts on the motors in the firmware, so the firmware won't idle any motors until all of them have been stopped for the inactivity time, at which point it idles them all at the same time.

[sprior]

While we're working on this problem I'd LOVE to see part of the solution include automatically sensing and setting the Z home nozzle distance automatically. But of course then I open a can of worms which includes bed leveling and aligning the nozzles to each other, though this turret idea could be combined with Z distance sensing to eliminate the need to the nozzles to be precisely aligned to each other.

[Tim]

Since I just love making drawings of vaporware, here's my latest. This version mounts the hinge low, close to the rail. That puts most of the weight above the hinge, so the motors should naturally want to rest on one side or the other, meaning that the extruder keeping it parked in one position doesn't have to continuously hold it there, but just needs to keep it from getting knocked out of position. Possibly the gear racks can be replaced by rubber strips or something similar, since positional accuracy is not particularly important. This is about as efficient as I think I can make the system, and I have checked various clearances and think everything can fit. Designing all the pieces, of course, is another matter altogether. Like most people, I have to go back to work tomorrow, so this is the last I'm going to be thinking about it for a while.