[ Closed ] Filament Extruder - Convert pellets to filament (Page 51 of 58)

Ian - First question.  Was the Sam Adams box filled with empty or full cans of beer and if full were they cold?

Right now the calipers are a diagnostic tool.  I can see every bump and divot as the filament runs through the capstan.  I made around a dozen manual checks and it seems spot on.  At some point I intend to interface the caliper serial port to an Arduino to use in a control loop or for data logging.  Shut down if we get out of tolerance.  I am not pleased with +/-0.1mm tolerances I am seeing.  Most of the time it is really tight and the calipers go into standby because nothing has changed.  Right now, I am assuming I am doing something bad in my setup and I may just need to focus on tuning.  What concerns me when the filament piles on the floor and jiggles around, it is constantly changing the back pressure at the extruder which stretches or compresses it.  I may try some of the SABIC MG94-NA1000 to see if that helps.  One thing I have noticed is that if I see a section of thick filament, it will be followed by a section of thin filament.  Could be auger or PID action.  Could also be that the heavy section applies more force at the extruder, stretching the stuff coming out, making it thinner.  Recommendation for the KS is to be sure to have reproducible guide and fan placements so the recipie gets shared.

I think I can make the optical work.  Plan to put the bottom of the loop through two up rights with a cross piece at the bottom.  In operaton, it would be ok for the filament to sometimes contact the cross piece.  Opto detects when filament is touching AND just above the cross bar.  Had a nice setup today where the loop was 40" down and capstan was 40" away from extruder.  But I totally hear you about the filament is stiff and takes the shape it cools in. It is not string.  Running the capstan motor on a lab supply.  I was able to dial in the RPM to match the extruder output so that I could walk away for 3-5 minutes at a time.

I ran a quick setup in direct feed mode, no guides, 18" span 6" droop, lego pulley weighting the filament.  I was able to actively adjust the power supply to maintain the bottom of the droop to almost 1/4".  I am thinking to go back to your teeter-todder suggestion for the weight.  What I like about it is that it eliminates the stroodling of ABS as it cools over the guides.  Since most of the force is provided by the pulley weight, we are less susceptable force variations caused by filament thickness variation.  Downside is it requires a tight control loop.  I am going to re-run this manually with a teeder-todder weight and see what kind of filament variation I get.

Finally.  Has anyone tried strooding some oversize filament and running it past a series of heated dies?  I noticed:  taulman 618 Technical Specifications:
Type = 4-Station-Drawn and Extruded "3D Printing Specific" Nylon based co-polymer.

i use a 80mm but i 'protect' the nozzle and let the air flow cool the filament only 1 inch after the nozzle...

i can't see that as a bad thing ..only if the speed of the cooling is important for the final hardness of the filament..

My fan blows about half on the nozzle and half on the filament. No problems.

How does the capstan work with the caliper?  Did you make that, or is it off the shelf?  Do you think that would be reproducible as a printed part?  I like the idea of using a caliper, especially if you can get a graph of filament diameter.   I assume the caliper is driven by the movement of a spring loaded idler somehow, but it's difficult to make out from the picture.

Nick,

I sent the 80mm fans to beta testers so they could give them a shot. The 40mm fans are rated at 23CFM, which is probably VERY optimistic. You could try just reducing voltage to the fan. Also, try reducing extrusion temp. 190C is hot.

No real plans for changing the pipe nipple, stainless is 3x more and isn't needed in my experience. We'll give it a shot, though.

Did you make sure to grind the lips well?

Ian - I'll post some better pics and disassembled parts later, but no magic.  3D printed is the way to go, but will take a bit of work.  This was an older HF caliper and maybe was smoother than the latest ones which feel gritty.  But. The idea is you have 5/8" OD 1/4" id eBay roller bearings mounted on each jaw.  Tension is adjusted by putting a weight on the skinny depth metal strip that comes out of the end when you open the jaws.  You can see part of a tubing cutter weight in the photo.  Ideally maybe a piece steel pipe with a cap over the end.   Didn't take too much weight, but newer units seem to have more drag.  I cut mounting holes in the hardened steel jaws with a 1/8 carbide end mill.  For the 3D, you probably want a wrap around each jaw that picks up some of the jaw features and maybe has some clamping screws on the back.  Dremel off the two ears for measuring ID?  Motor is an $8 eBay 100RPM 12V w/gearbox.  This forced me to use a 0.3" rubber capstan from a dremel sanding drum.  Running at about 2.1V.  I would opt for a lower RPM motor and larger capstan on the do over.  One problem with printed is you need to find a way to mount the rubber cylinder on the motor shaft.  I machined a brass wheel that the rubber slips over (expands to 0.3") which attached to the gearbox shaft with a set screw.    The rubber surface is a compromise as far as gripping power.  Bicycle inner tube might be another option for rubber.   Best would have been to use a knurled finish on brass with no rubber, but that would have left ding marks in the filament.  Motor has plenty of torque so it is just a matter of how much pressure you want to use on the ball bearing that presses the filament to the capstan.  The bent piece of music wire in the pic is what is applying capstan force.  Haven't tried the serial interface yet, but there are links on the web about it and apparently Arduinio drivers.

Tim - I'll drop temp to 185C.  Also raised the nozzel height to 49"...that is high!  I was seeing my controller indicating about 2C high compared to my Omega.  Displayed controller temp also apparently has hysteresis.  Omega which has 0.1C resolution, was indicating amost +/-1C movement in head temp while the PID display stayed constant.  At higher res, you can see delayed heat pulses hitting the head every 20 seconds from the controller cycling. 

Pipe lips were cut down on the lathe, even though the auger does not directly touch them.  My auger doesn't extend to the end of the pipe.  Also like ysb, I lightly touched the inside of the pipe with an adjustable reamer to knock off some high edges of the weld seam which the auger was going to attack anyway.  Figured if it became an issue, I would do a more serious job reaming later. 

ysb- how did you polish your pipe?  Down to bare bright metal?

do you see this,? filament withd measure

http://www.thingiverse.com/thing:70775

Hey Tim,

Nice to see the KS still going great guns.  I see you're out of kits again (don't need one, I'm in already, but still watching).  This is starting to look like it could be 6 months worth f pre-orders at this rate.

Craig.

Sounds sensible....

Besides, it leaves the opportunity to make a decision abut a future business model for after the KS when you don't have to give them a % of the price.

Inspired by Tim and not capable to wait till May when I must get the kit from him I have built as a trial from junks I found around my extruder. I do not have yet a PWM speed controller, which means at this moment my motor has a fix speed of 40 rpm. What is intriguing me is the fact that the diameter does not keep constant, and filament has curlings. Also, the diameter is much bigger around 2 mm instead of 1.75 mm. The things are getting better if I increase the temperature around 245 C degrees. Could somebody give my some tips? What I must do to improve? I must mention that I tried ABS pellets. any help is highly appreciated.

Thank you.

Zoltan - Very nice setup.  What model gearbox are you using?  Looks like your heater is mounted directly on the nozzle?  On my beta, the heater mounts on the pipe coupler.  The auger does not stick out of the pipe nipple and maybe is recessed 1/4".  So there is a space between the auger and nozzle, right under the heater.  Part of the auger tip needs to be in the melt zone.  Could it be you are needing to run higher temps to get the auger tip hot enough?  Have you looked at your auger motor current?  At 40 RPM in the melt zone, you should almost be getting close to where the auger will self heat.  I am seeing 2 RPM.  Have you disassembled it to see how much of the auger is in the melt zone?  Are you running a full length solid brass nozzle as shown in the pdf?