Hi everyone! Here are the GBX R&D updates from the past month:
Infill Bridging Distance
The smoothness of the top surface of a 3D print depends on how successfully the solid top layers can print on the infill geometry below it. During slicing, parameters including number of top layers, infill percentage, and infill type must be chosen. In an effort to characterize the interaction between these variables to improve slicing decisions, a 3” x 3” x 1.5” calibration square was sliced with infill percentages ranging from 5 to 30%, for both rectilinear and full honeycomb infill types, and for nozzle sizes 0.4mm, 0.8mm, and 1.75mm.
Cross-section screenshot of 3” x 3” x 1.5” calibration squares sliced with a 0.8mm nozzle and 5% to 30% full hexagon infill.
Cross section images of the cubes were brought into ImageJ, where the maximum bridging distance was measured across the holes of the infill. The infill percentage was then plotted against the required bridging distance.
The infill percentage is calculated by the slicer based on the volume of infill material versus the volume occupied by the geometry with infill, including empty space. Therefore, models sliced with the same infill percentage but with different infill types will have the same mass of infill, but due to the different infill geometries, may have different bridging distances. Based on the bridging distances measured for the calibration cube, full honeycomb infill has about double the required bridging distance than rectilinear even at the same infill percentage. In practice, this means models sliced with full honeycomb infill must be sliced at a higher infill percentage compared to rectilinear infill in order to achieve the same required bridging distance and smooth top layers. Increasing infill percentage has the added effects of increasing part mass, strength, and print time, which may be a benefit or a disadvantage depending on the application of the printed part.
At any one infill percentage, there is also a correlation between nozzle diameter and required bridging distance. Larger nozzle diameters produce larger extrusion lines, and since infill percentage is a volume calculation, this results in larger bridging distances for larger nozzles. For geometries that require a lot of infill, this effect may be substantial enough to make print times for different nozzle sizes similar, since larger nozzle sizes may require higher infill rates and therefore longer print times in order to achieve a low enough bridging distance.
These graphs include power series trendlines, which can be used to calculate required bridging distance from infill percentage and vice versa. The below equation describes this relationship:
Required Bridging Distance = A*(Infill Percentage)^B
Where the empirically-derived constants A and B are in the below table depending on the infill type and the nozzle size:
|
Infill Type |
Rectilinear |
Full Honeycomb |
Rectilinear |
Full Honeycomb |
Rectilinear |
Full Honeycomb |
|
Nozzle Size (mm) |
0.4 |
0.4 |
0.8 |
0.8 |
1.75 |
1.75 |
|
A |
44 |
151 |
105 |
337 |
279 |
736 |
|
B |
-0.958 |
-1.07 |
-1.05 |
-1.08 |
-1.16 |
-1.13 |
Part Cooling Testing
One of the subtasks of re:3D’s NSF Phase II grant is to develop a part cooling solution for Gigabot X. During the last month, work has been done to develop testing procedures and models to characterize the effect of part cooling on 3D printed parts, with the eventual goal of using the testing procedures to optimize the part cooling solution.
The first set of tests developed was bridging, in which a line is extruded over an unsupported distance in the model. Examples of geometries that require bridging are horizontal holes, overhangs, and top layers, which must bridge over infill.
A test model was created with a top surface that is one layer thick. One side of the model is straight, and the other side is set at an angle, which varies the bridging distance from 0mm to 25mm throughout the length of the model. Tick marks along the side of the model denote 5mm increments in the bridging distance.
Sliced model for testing part cooling’s effect on bridging over bridging distances from 0 to 25mm.
The model was printed sequentially with part cooling set at 0% (off) and at 100% (on). Using Ultrafuse rPET pellets, five trials of each part cooling condition were printed with a 0.8mm nozzle, and three trials of each part cooling condition were printed with a 1.75mm nozzle.
0% Part Cooling, 0.8mm nozzle (top) and 100% Part Cooling, 0.8mm nozzle (bottom)
Visual inspection of the resulting 3D prints show variance in sagging of the bridging layer depending on bridging distance, nozzle size, and part cooling conditions. To quantify the sagging, a height gauge was used to compare the minimum height of the bridging layer to the perimeter of the model, which is the expected height of the bridging layer under ideal bridging conditions. This sagging distance was measured at 5mm increments in the bridging distance.
Trials with 0% part cooling and 100% part cooling both show increasing sagging distance as bridging distance increased for both nozzle sizes, but trials with 100% part cooling show up to twice as much reduction in sagging distance in the bridging distance range tested. 3D prints from the smaller 0.8mm nozzle showed better bridging and reduced sagging overall compared to the 1.7mm nozzle, with similar sagging distances between 0.8mm prints without part cooling and 1.75mm nozzle prints with part cooling.
Graphs showing the effect of part cooling on sagging distance, the vertical distance between the lowest point of the bridging layer and its expected position in the model, over a range of bridging distances for 0.8mm and 1.75mm nozzles.
One thing to note in the graphs is the cutoff sagging distance, which is 1.5mm for a 0.8mm nozzle and 3mm for a 1.75mm nozzle. These values were determined from the minimum infill percentage achievable from past prints (8% for rectilinear with a 0.8mm nozzle, and 15% for rectilinear with a 1.75mm nozzle), and relating those infill percentages to the required bridging distance to print solid top layers on top of the infill geometry. This means that any sagging distance below the cutoff line is recoverable, and any above need either excess top layers, transitions to higher infills, or may be impossible to achieve a smooth top surface. It is also important to note that the cutoff line is based on Ultrafuse rPET pellets for large prints that take up most of the regular GBX X and Y build space. Other factors such as material, part geometry, and heat buildup may shift the position of the cutoff line.
Part cooling is controlled by a percentage setting in Simplify3D, which translates to the Mcode command M106, which sets the fan speed via a number from 0 to 255 that controls the digital PWM output signal to the fan. To test the effect of different percentages of part cooling on the printed part, four additional trials were printed with a 0.8mm nozzle with 20%, 40%, 60%, and 80% part cooling. The resulting sagging distance data were combined with data from the 0% and 100% part cooling trials and plotted.
Graph showing the effect of different percentages of part cooling on sagging distance, the vertical distance between the lowest point of the bridging layer and its expected position in the model, over a range of bridging distances.
The graph confirms that our standard 40mm fan cannot translate a variable input pulse into a variable fan speed. Instead, the fan is either off (speed set to 0 to 50%) or on (speed set to 50 to 100%), which translates to a binary effect on part cooling. This test was not repeated for other nozzle sizes, since the fan behavior in response to slicer settings is independent of nozzle size.
With the work relating infill percentage to required bridging distance detailed earlier in this post, the sagging distance from part cooling testing can be related not only to bridging distance, but also infill percentage.
Graphs showing the sagging distance in relation to infill percentage for 0.8mm and 1.75mm nozzles, and rectilinear and full honeycomb infill types.
For the 0.8mm nozzle, part cooling resulted in viable bridging for all infill percentages tested, decreasing the minimum infill percentage from 9% to 4% for rectilinear infill and 23% to 10% for full honeycomb infill. For the 1.75mm nozzle, part cooling reduced the minimum infill percentage from 16% to 12% for rectilinear infill, and 40% to 30% for full honeycomb infill. Another way to interpret these reductions is in their effect on print time and material: the infill percentage has a linear relationship with the print time and material mass associated with the infill, which has a significant impact on models that are mostly infill.
Part cooling testing will continue to further characterize the effect of part cooling on various aspects of a 3D print, such as interlayer adhesion, material strength, heat build up, and overhang angle.
GBX Simplify3D Profiles
Further development has been done on the Gigabot X Simplify3D FFF profiles. All profiles are available for download on re:3D’s Github, where the changes to the profiles can also be tracked. Instructions for updating the FFF profiles in an instance of Simplify3D can be found on the Knowledge Base.
Improved Wire Management and GBX Extruder Cover
The [11337] GBX Motor Spacer was updated to improve wire management at the extruder. An additional section was added to accommodate up to 5 fan or heater connectors: three for heaters, one for the heat sink and motor fans, and one for the part cooling fan that is currently under development. The new motor spacer also holds the thermocouple connectors more securely, whereas the thermocouple connectors frequently slid out of the previous design.
The [12144] GBX Extruder Cap and [12142] GBX Extruder Barrel Cover Front were also updated to remove internal fins to provide more room for the wires to route into the motor spacer.
Updated GBX Channel Covers
In response to some channel covers breaking in shipping, they were updated to add fillets to the mounting points to improve strength. The heat set inserts were also changed from M4s to M3s to allow for more tolerance when mounting into the clearance holes in the hopper gantry channels.
