Improvements to The Extruder
recycling
polymers
filament extruder
As the extruder extrudes, the open-source nature of the machine gives me improvement ideas and easily allows to test them out. Some of the tests have a negative result, but those that have a positive result I try to add into the machine.
The extruder has become a constant prototype, always with little changes “let me adjust that, before the next extrusion”.
Here is a test, extruding with the fan from below, to see if the airflow direction can control the output diameter better from a certain angle (given turbulance from the other two fans as well).
Diameter sensor #
In the above picture on the left, I mounted the improved diameter sensor, to check the finished filament diameter. It actually has two sensors for the diameter, one after another. The second one is also rotated by 90° along the filament diameter, to also measure ovality of the filament by taking the difference between both sensors.
Cooling and Stretching Filament #
A small improvement: the fan shroud is now a transparent acrylic instead of an old piece of paper out of the recycling. This also makes the filament catenary visible. In some pictures I still have another old version where it’s translucent plastic, not as good as this transparent current version.
Also, the fan has a pivoting linkage for adjusting it and swinging it out of the way when starting the extrusion.
Granules Flowability #
Another mini-experiment is the orange printed piece: a little vibration motor attached to the funnel. This should shake the input material around.
Especially at the end when the funnel empties out: the problem is that the last granules bunch up on the walls of the funnel and just air gets into the extruder. This then causes the filament to be too thin for printing at that point along the length, and needs to be interrupted in the best case. When unlucky, the thin new filament also sticks to itself in an unwelcome loop and the machine jams.
I would prefer for all the granules until the end to be used up, so that a good diameter filament can be produced as far as possible.
(One of) the symptom(s) of badly flowing granules. The machine thinks it’s funny to do that and likes to surprise me sometimes.
Catenary #
Mathematically, a catenary exactly describes the shape a rope takes on when hanging by two points. In my case, the freshly extruded filament is very close to that shape (note: the differential equation has the same solution even with hang-points at different heights):
An unobstructed view of the nice catenary shape. The new filament comes out of the nozzle at the left, and it’s still viscous enough to hang by its own weight. The little metal wire is a sensor to measure the height of (the tip of) this loop and adjust the pulling speed.
This system is a part of the original machine design and a very clever way to measure the filament directly when it exits the nozzle (without touching it) to adjust other parameters for a consistent diameter.
I added the stepper motor there, connected with the yellow cylinder, to move the sensor assembly automatically up or down by small amounts, this can help as a longer-timescale feedback controller for better diameter tolerance.
But after multiple hours of use, I noticed that the exact height of the sensor does not control the filament diameter as much. Other parameters like temperature influence diameter more consistently.
With a bigger hopper, designed to prevent granules from getting into the high-torque extruder screw mechanism at the bearing: they should only get in at the feed-in point.
Previously, a granule would sometimes slip into the torque-bearing sideways. Because it’s plastic, the motor still continues turning but makes a plastic skreech noise at every rotation (for multiple hours).
Now that problem is solved.
PLA+ #
I was extruding PLA+ that day, it’s not the same as regular PLA. Some manufacturers make filament similar to PLA but with some additions into the polymer. The intention is to keep the PLA aspects of simple and friendly to print, but increase strength. They do achieve that, though at some expenses, like e.g. more hygroscopic filament, so drying it is more important before printing than normal PLA.
During collection PLA and PLA+ were separated. This means I can now compare: the viscosity is much lower for PLA+ than PLA, even though extrusion temperatures are almost the same. A lower viscosity means it flows further down and faster; which translates to a much more aggressive material flow rate (~ +30%) to prevent it from thinning too much.
But it’s possible to extrude into well toleranced filament, so then here goes:
** wait for 3d printing noises **
Fresh off the printer, it prints great! Once the settings specific to PLA+ are dialled in (+15C compared to usual recycled PLA mostly for good layer adhesion), when dry it prints just as well as standard PLA. Just a hint of stringing to be expected from recycled polymers, which goes away with a heatgun.
Here is also a view of a flat face in the print. The other faces I “painted” with fuzzy skin in the slicer. It’s a cosmetic effect of adding +- 0.2mm bumps on the surface to make layer lines less obvious, and not felt when touching the object.
This fuzzy skin should only be applied to cosmetic surfaces, never on precise fits or faces of your model that have a reason for being flat.
Otherwise, the fuzzy skin is a great way to make the printed object look nice and, in a way, more interesting and organic.
Firmware recompile #
One last, but not least, improvement I made to the extruder has been
in the firmware. I changed the temperature handling loop to
represent temperature as a float instead of an int in all places (the task was easier thanks to the sensor
already being accurate enough that it was already measuring temperature as a floating point value).
This goes all the way to the display controller: setting a target of 189.6°C is now possible, where before only 189°C or 190°C were possible. Temperature has a reasonably big impact on filament diameter in a steady state, I estimate 20-30μm per 1°C, so finer steering (maybe even 2μm??) by 0.1°C should keep diameter consistency even better.
Future plans #
Currently in the works are some more firmware recompiles (yayy C macro preprocessor compile errors!). The goal is to connect some more hardware to the motherboard and control that through Gcode on a USB connection. Such hardware includes the diameter sensors in high resolution (input), additional cooling fans (output), vibration motor on a regular timer (output) and maybe some others.
The computer connected on the other end also should log all extrusion parameters, and though Gcode be able to control the machine (set temperature, extruder rpm).
The scope of these changes should reach all the way to the diameter sensors: I don’t like that there is a full separate microcontroller involved, and am replacing it by just a high-resolution ADC.
Also PWM (percentage) control of the fans! Stay tuned…
This recycling project is supported by
