What's this called (Rep2)?

[David Kessner]

What is the black plastic thing on the Rep2 extruder called?  I'm talking about the plastic piece that is above and to the left of the filament stepper motor.  The filament feed tube sticks into it. 

I'm designing a new one, with some cool features, but I just don't know what it's called.

Thanks!  -DK

[Dan Newman]

On 29 Aug 2013 , at 5:38 PM, David Kessner wrote:

> What is the black plastic thing on the Rep2 extruder called?

MBI called it the "Wire Filament Extruder Guide" in their BOM and SolidWorks files
for the Replicator 1,

Wire-Filament Extruder Guide.SLDPRT

from one of the downloadable zip files at

http://www.thingiverse.com/thing:18813

Dan

[DHeadrick]

The mount for the duct fan with the cheezy cover?

http://www.thingiverse.com/thing:113595

[David Kessner]

Thanks guys!

My version is going to have some important differences from the original, as well as the one you made DHeadrick.

The most important difference is that it will integrate a switch that will close when the extruder motor is pulling.  The purpose of this is to allow for an auxiliary stepper motor to help the primary extruder motor so it doesn't have to pull so hard.  It's for this project:  http://davidkessner.wordpress.com/2013/08/28/3-d-printer-chamber/

The second improvement is that it will allow for some air to escape out the left side of the extruder heat sink.  Any help that heat sink gets is a plus!

But, thanks again guys!  You have helped me organize my designs in a more meaningful way.

-DK

[Jetguy]

Already did it, just print and use http://www.thingiverse.com/thing:107775

Filament feeder and P-stop trigger.

[David Kessner]

That is similar to what I am going to do, but I need the sensor to be on the extruder.  In my case, it is mostly the filament tube itself that is providing the resistance due to its additional length and its more twisted than in a normal Rep2.  If I put the sensor at the spool-end of the tube then I run the risk of not always knowing when filament needs feeding.

Because the sensor is on the extruder, I am also trying to make it as small and light as possible, so the motors are not burdened trying to move more mass.  

[Jetguy]

Yes, but the problem then is you now add wires and complexity to the extruder.

It's just one more thing to go wrong, instead of something to increase reliability.

 

If you want to take it to the next level, basically substitute the 555 timer in mine and drive the step pin on your driver the feeder. This way  it moves exactly in tandem with the extruder.

The trick here is to also then use the enable pin in a manner like I am using it.

This way if your feed mechanism ever has a problem, the motor freewheels instead of being locked. That should be a key element of any implementation.

By all means, mod it and go nuts, I'm just trying to give some pointers as I put a lot of thought into making one.

We threw out the idea of putting anything on the head for a long list of reasons, mostly being it could affect print quality or reliability in some way.

I also did not use the step pin method because I was trying to make this universal and completely external.

 

I also played with the idea of using a microcontroller for a much more controlled system of feeding. For example, when the feed is required switch or whatever you use it tripped, then the stepper could be pulsed a know distance rather than just moving until the trip point was cleared in my very simple 555 design.

You could even implement a variable speed based on how much needed taken up and so on.

 

The best of both worlds might be to use a microcontroller and sense the step pin, but also have the feed detect or tension switch (either on the head as you describe or one like mine at the feeder, so that both could call for filament. That way, 99.9% of the time, we assume the 1:1 ratio of feeder to extruder= no tension, but if there is slippage, then the feeder can push a few extra pulses until there is no tension.

Really there is no limit here, we can make this super complex code that senses the pressure and the distance, monitors the motherboard step pulses, even put an encoder somewhere in line and calculate filament slippage.

It comes down to cost and time. The other factor I look at is simple= reliable, complex may actually cause failures. But that's why I say use the enable pin in whatever logic setup you use so in a worst case scenario, you have the fact the feeder can freewheel rather than being locked. The other factor is that if learn nothing else from other failures, stay way from cheap DC motors as feeders. Most of them cannot take any side load on the shaft and will fail when used in a pinch wheel mechanism, where a stepper can take a reasonable load, is cheap, and has ball bearings.

[David Kessner]

I think that I can make the new piece plus the sensor for not much more weight than the original plastic piece.  Other than the plastic bits, the only thing I am adding is a small microswitch (https://www.sparkfun.com/products/97) and the smaller than usual spring from a pen.  The wire is 28 guage, but I'll look at the local electronics shop to see if they have anything smaller but still stranded.  At the moment, the plastic parts are about the same weight as the original.

I thought about making the two stepper motors run in tandem, without any sensor at all.  The problem with this is that it will be impossible to keep the two motors feeding exactly at the same rate.  Variations in grub wheel dimensions, schmutz on the wheels, variations in roundness of the filament will all add together and cause the two motors to feed very slightly differently.  I would think that it could start slipping as early as 5 to 10 minutes into the print (I define a slip as one tooth on the grub wheel).  While a slip would not end the print, it could cause small imperfections.

Although an interesting option for this method is that the filament feed tube is not required at all and could eliminate slippage.  If you start with enough "slack" in the filament, the two motors should work together to maintain the same amount of slackness and making the tube redundant.  Any mismatch in the filament feed rate would result in the amount of slack changing.  But it would take much longer time for any feed rate mismatches to result in slipping, maybe 5+ hours instead of 5+ minutes.  I will have to think about that method more.  It is not a slam dunk for several reasons, but the possibilities are very attractive.

I am planning on using an ARM Cortex M4 to do the stepper motor control.  Overkill for sure, but the boards are only $13 and more importantly:  I already have several.  The cool thing about this is that I can monitor the feed rate and report errors caused by a jam, tangled spool, or the filament running out.

I refuse to use a 555, just on principle.  Of course, I'm an EE by day so I'm kinda snooty that way!  :)

Thanks.

-DK

[Obfuscated]

BTW.. Does the P-Stop command simply stop the printer and leaves the hotend where it was or does it go into the filament change position? My concern is that if it just stops it where it is that the hotend will melt the plastic it's near.

Also I'm wondering if this system could not only be used to pause the printer if it detects there is no more filament or also if the filament is not feeding or otherwise jammed and prevent an airprint?

[Dan Newman]

On 30 Aug 2013 , at 11:27 AM, Obfuscated wrote:

> BTW.. Does the P-Stop command simply stop the printer

It pauses the printer just as though you manually paused it. Only works
with Sailfish. If you use Sailfish, run it and pause it and you'll see
exactly what it all entails: clear the build platform, run a countdown
timer on how long you will leave heaters on (assuming you have set
"heat during pause", etc.).

Dan

[David Kessner]

Jetguy,

So after talking it over with fellow engineers at my office, here is what I decided to do:

The whole "tapping into the step-pulse signal" on the MoBo, running it into a Microcontroller, and then driving the aux stepper motor with that is very attractive.  The problem is that we just don't know how matched the two motors/gears will be.  Ideally it should be able to do a huge print without needing manual intervention, and we just don't know if that is possible.  The main issue is with the ratio that the two motors run at.  In theory, it should be 1:1.  But I suspect that it will be close to 1:1 but not identical.  Of course the microcontroller can support any ratio, but the issue is figuring out how much of a variation there is in the real world, and how much variation we can tolerate.  I will call this the "Synchronous Feed" solution. 

On the other hand, I am fairly confident that my original approach will be functional.  I might need to slow down the printer to accommodate any additional weight on the extruder, but it will work.   I will call this the "Sensor" solution.

However, the sync-feed solution is very attractive since it reduces or eliminates several other issues-- mostly loading on the primary filament motor due to friction in the tube.  Removing the tube could also improve prints due to less restriction on the carriage when moving.  

So I'm going to do both solutions!  I will start with the sensor solution, but still hook up the step-pulse from the MoBo.  That will allow me to have a functional solution while at the same time gathering long term statistics on the ratio between the two motors.  If it turns out that the ratio is fairly constant with different filaments and such, then I could easily pull out the sensor and filament tube and do the sync-feed solution.  Even better:  I can then report the findings so everyone else in the community can benefit from my experiment.

Of course, this requires extra work.  But the was never about reducing the amount of work, so long as I'm having fun and learning while doing it.

-DK

[David Kessner]

My version, with more airflow:  http://www.thingiverse.com/thing:143502

This one does not include the sensor that I've been talking about.  It is simply a replacement for the existing Guide that is lighter and has more airflow.  

[David Kessner]

Close but no cigar.  My first attempt at an extruder sensor was very close, and shows huge promise.  Here's some pics:

This overview shot shows the new Filament Extruder Guide, which is a pivoting arm with a microswitch and spring on it.  In the nub behind the white filament feed tube is the sensor.  You should notice that the front of the guide is open, which allows for air to exit from the left of the heat sink and thus cooling it better.

This second pic shows the arm flipped over, and the spring.  This spring was taken from a Pilot G-2 ball point pen.  It is slightly wider and shorter than your typical pen spring.

This pic shows the spring with attached plunger removed.  Also visible is the microswitch at the "top" of the hole.  If you go back to the first picture and look at the nub where the sensor is, you will see the four silver colored pins of the switch sticking out the top of the nub.  Currently the wires are not connected, but they would just be soldered directly to the switch.  I plan on using very thin wire, so weight should not be a major concern.

And now, a video of the thing in operation:  http://youtu.be/9wff4l_vB00  

In this video I am messing with the filament at the spool.  I first prevent the filament from feeding (the arm moves closer to the stepper motor), and then I force some extra filament in (the arm raises up, away from the motor).

There is one major flaw in this design, however.  In the first picture, you can see the head of a flat-head screw that the arm pivots on.  That screw, and the block of PLA that it is in, hits the frame of the Rep2 when the extruder moves to the front-left at the beginning of the print.  It is not a major flaw, but enough to keep me from trying this out right now.

Anyway, I thought you'd be amused at the progress.

-DK