Hello everyone. Here are the R&D updates for GBX from the last month.
GBX Klipper Configuration
Klipper configuration for Gigabot X has continued this month (see last month’s forum updates for an introduction to Klipper). Klipper was initially configured on an Azteeg board, and last week it was uploaded onto a 32 bit board, which is what re:3D is transitioning to. Configuration for the 32-bit board has begun to build upon the configuration established with the Gigabot X with an Azteeg. Configuration includes:
- Setting pins
- Establishing preheat macros for PLA, PETG, PC, and Dynapurge
- Calibrating motor speeds and step values
- Establishing purging macros to purge for 2, 5, and 10 minutes
- Writing a macro to aid in Z calibration by homing in Z and positioning the extruder at 101.3mm in Z in the center of the bed
- Update thermal runaway and max extrusion settings for the Gigabot X extruder
- Remove any filament-related functionality
- Optimize UI configuration for mobile, desktop, and tablet platforms
In order to meet short-term deadlines, configuration for basic functionality will be prioritized to bring the Gigabot X Klipper firmware for a 32-bit board on par with GBX Marlin firmware for an Azteeg. After basic functionality is achieved and 100 hour of testing are completed, the first two Gigabot X machines with Klipper and 32 bit boards will be shipped out to beta customers, and further development of Klipper can be pursued to go beyond the functionality previously available via Marlin firmware.
rPET Material Testing
This past month, rPET material testing on Gigabot X was revisited with an rPET sample pulled off the processing line of Phoenix Technologies, a company that recycles rPET. The material was tested in collaboration with the Engineering Mechanics Corporation of Columbus (EMC2) and, it underwent extrusion and print optimization testing. The results are reported below:
Particle analysis picture of the Phoenix Technologies rPET sample.
|
ImageJ Particle Analysis Data |
|
|
Average Particle Area (mm2) |
10.05 |
|
Median Particle Area (mm2) |
9.06 |
|
Particle Area Standard Deviation (mm2) |
5.50 |
|
Print Settings |
||||
|
Extrusion Test |
Standard Extrusion Rate |
Max Extrusion Rate |
||
|
Bottom Temperature (℃) |
240 |
240 |
||
|
Middle Temperature (℃) |
235 |
235 |
||
|
Top Temperature (℃) |
200 |
200 |
||
|
Test Extruder Speed (mm/s) |
3 |
8 |
||
|
Average Extrusion Sample Mass Over 5 Trials (g) |
5.12 |
3.58 |
||
|
Standard Deviation of Extrusion Sample Mass (g) |
0.59 |
0.92 |
||
|
Max Extrusion Rate (kg/hr) |
0.276 |
0.516 |
||
|
re:3D Drying Recommendations |
Dry to PET specifications and store in a sealed container with desiccant. Re-dry once a week. |
|||
The material showed inconsistent extrusion for both test conditions: extruding at a rate comparable to printing with a 1.75mm nozzle at standard settings (Standard Extrusion Rate), and extruding at a maximum speed (Max Extrusion Rate). However, the inconsistency increased for the latter. There was slight crystallization at the nozzle tip when left heated at extruded temperatures, but not enough to clog the nozzle, which is a common issue with other rPET.
The inconsistent extrusion persisted in the calibration cylinder, but was greatly mitigated by using a vibration motor. The material has evidence of contaminants, but most pass through the 1.75mm nozzle without impeding flow. The material underextrudes overall, which was compensated for by adjusting the extruder motor calibration. The temperatures were also adjusted to achieve a smoother surface quality.
Calibration cylinder printed from rPET flake with a 1.75mm nozzle.
The Moai printed well on the second trial with vase mode. However, additional Moai couldn’t be printed due to a piece of metal contaminant in the nozzle, combined with increased crystallization after a week since the material was last dried.
Moai test model printed from rPET with a 1.75mm nozzle.
|
Recommended Settings |
||
|
Nozzle Size (mm) |
1.75mm. Smaller nozzle sizes may have issues with contaminants clogging the nozzle, and larger nozzle sizes may have issues with reaching a high enough extrusion rate. |
|
|
Bottom Temperature (℃) |
245 |
|
|
Middle Temperature (℃) |
240 |
|
|
Top Temperature (℃) |
180 |
|
|
Bed Temperature (℃) |
60 |
|
|
Extrusion Rate (M92 E Value) |
78 |
|
|
Print Speed (mm/s) |
40 |
|
|
Enclosure (Y/N) |
N |
|
|
Bed Adhesive(s) Recommended |
PVA Glue |
|
|
Bed Adhesive(s) Not Recommended |
N/A |
|
|
Additional Settings |
Using a vibration motor is essential to achieve consistent extrusion. |
|
The rPET sample exhibited the following issues:
- Extrusion inconsistency was mitigated by using a vibration motor, which indicates a difficulty with feeding at the extruder inlet.
- Overall underextrusion was indicated by the fact that the extruder motor was calibrated to 78 steps/mm, which is 30% more than the baseline 55mm/s.
- Small contaminants easily passed through the 1.75mm nozzle, but a large chunk of metal completely blocked extrusion during testing.
- Crystallization was present at the nozzle when the extruder was left heated and idle for at least 30 mins, and the extent of crystallization increased to the point of fully clogging the nozzle when it was at least a week since the material was last dried.
In summary, although this material showcased some issues with inconsistent extrusion and crystallization, it performed better in those respects compared to other rPET water bottle flake tested by re:3D. Diligent material dehydration and reducing contaminants will make this material more reliable to print with. Further testing beyond the Moai test print is recommended, especially larger and longer prints.
Vibration Motor Changes
In the January GBX forum update, the vibration motor design was improved in an attempt to decrease the decibel level of the vibration while it was running. The free limited version of the Clear Wave app was downloaded and used to take decibel measurements since the previously used decibel measurement app had a limited number of decibel readings in its free trial. That means that the decibel measurements in this forum post may not be comparable to past decibel measurements.
The below decibel readings were taken with the vibration motor running at the established PWM-controlled average voltage setting of 9V, which was set via the gcode line M106 P1 S85. Motor Mount Rev5 was tested to assess if it improved compared to the previous design, and foam was placed in different positions in an attempt to identify where the sound was coming from.
|
Trial |
Average dB |
Max dB |
|
Control with no vibration motor running |
36 |
|
|
With Motor Mount Rev5 |
65 |
73 |
|
Motor Mount Rev5 with wires taped in place and foam added around the connectors |
65 |
73 |
|
Motor Mount Rev5 with foram added around motor |
64 |
64 |
The decibel results from testing Motor Mount Rev5 were inconclusive. Therefore, Motor Mount Rev6 was designed with the top wire cover part completely removed. Additional versions of a Motor Mount Rev6 were designed (Revs 6.2, 6.3, 6.4), and each version had different alterations to test what decreased the decibel reading.
The Rev6 versions of the motor mount were tested with the same procedure as the Rev5 motor mount.
Overall, Revs 6, 6.2, and 6.4 performed better than Rev5, with a decrease from 65 dB to 50 decibels. However, there was not much difference between the Rev6 versions. Rev6.3 (with the mounting face shaved off to increase clamping) actually showed an increase in decibel level from 49 to 59dB compared to Rev6.
Since Revs 6.2 and 6.4 did not decrease the sound level, Rev6 was selected as the best design since it retains the motor-holding nub removed in Rev6.2, and the cover for the eccentric weight removed in Rev6.4.
The Motor Mount Rev6 was used to continue material testing. However, after 10 more hours of material testing, the vibration motor feed throat split in half along the layer seams.
Broken Vibration Motor Feed Throat Rev2.
To strengthen the Vibration Motor Feed Throat and test out some more noise reduction measures, a Rev3 version was made with the following design changes:
- Add a fillet in the area that broke
- Remove the top part of the vibration motor mount section to replicate the improves of Rev6 over Rev5 of the vibration motor mount.
- Make some internal fins to hold the vibration motor to reduce sound
- Slice with solid perimeters on the feed throat walls.
- Trim the sides and bottom of the feed throat tongue because that part usually doesn’t print well
Vibration Motor Feed Throat Rev3 with strengthened motor mount geometry and other design improvements.
The Vibration Motor Feed Throat Rev3 was printed and installed on in-house GBX 554 with a Motor Mount Clamp Rev6. The resulting decibel readings were a significant improvement, with an average of 40dB and a maximum of 47dB, down from an average of 49dB and a max of 63dB.
The next step is to take the “fin” structures holding the motor in place in the feed throat and replicate the geometry in the motor mount piece, then test to validate if the decibel level decreases even further. However, anecdotally I (Helen) no longer get headaches from having to work next to a loud vibration motor with the current configuration.
4.2.4.2 Marlin GBX Bugfix Firmware
Although Gigabot X firmware is transitioning from Marlin on an Azteeg board to Klipper on a 32 bit board, a bugfix version of the 4.2.4 GBX firmware was created as 4.2.4.2 to make some essential bug fixes available. The 4.2.4.2 GBX Bugfix firmware is available on re:3D’s github. The fixes include:
- Updating the GBX Regular build volume in the X, Y, and Z dimensions
- Updating the GBX XLT build volume in the X, Y, and Z dimensions
- Adjusting the minimum temperature for the heatsink fan from 18C to 60C
Conferences, Competitions, and Events
In the coming months, there are a variety of events and competitions that Gigabot X is involved in:
- Research related to Gigabot X was submitted as abstracts for proposed papers to the 2022 International Conference for Additive Manufacturing (ICAM) and the 2022 ASME Industry Summit.
- Gigabot X was also submitted to the 2022 TCT Awards.
- Helen will be speaking about an open-source material testing procedure for GBX at the 2022 Open Hardware Summit on Friday April 22nd.
