During the COVID crisis, I was trying to maximise production and therefore to maximise flow. The graph below shows the geometric flow at the extruder driving wheel, and the resulting fluid flow out of the nozzle, with a 0.8 mm nozzle and PLA at 215°C. It may seem strange to many that those flows actually differ, without any jumping or grinding at the level of the extruder. However, the difference is easy to explain by flexing of the filament when the extruder operates near maximum pressure in the print head. Due to pressure, the filament is compressed near the nozzle, and comparatively extended at the extruder wheel, and for the same absolute mass flow, the extruder has to run faster in this condition. The upper limit is attained when the extruder occasionally jumps or when it grinds the filament.
To obtain this curve, I made calibration prints : elongated blocks with a known number of lengthwise runs and a known volume, and I printed those blocks at many speeds. The extruder speed was calculated by the baseline mechanism "Extrusion multiplier = 1", and I weighted each block to measure the actual fluid flow. I looked up the instantaneous speed on the printers' control panel and checked with the speed coloring of Simplify3D.
The blocks printed at high speed showed a higher mass deficit. This particular nozzle / material / temperature combination maxes out at eleven cubic millimeters per second. The flow will not be higher, whatever speed and layer thickness & width are programmed. Of course, it is also important to note that, for a good fill, what is important, is the fluid flow, and the curve shows that the commanded flow has to be substantially higher than the required fluid flow.
If I am going to print at constant speed, I can just set the proper "Extrusion multiplier", but while printing thick 0.4 mm layers, I measured that 12 seconds was a minimum solidification delay between successive layers. To achieve this minimum layer time, "Speed Overrides" allows automatic speed adjustments to achieve a given minimum time per layer, in order to let the layer solidify sufficiently. The function alone isn't enough, because if it slows down a lot, the "Extrusion multiplier" will be grossly overestimated and there will be too much material in the smaller layers. In fact, because the geometric flow goes down proportionally to speed, the point of operation move in the nearly linear region of the curve, and the fluid flow is too high.
Of course, the slicer has to avoid going back immediately over the parts of the previous layer that have been printed last, and to ensure that for tests, corkscrew mode is great. I hope that this is always guaranteed, but in fact I am sure that it's not, because when I printed my elongated rectangular blocks I noticed that the first length of filament of one layer was printed just on top of the still hot last length of the layer below. In a real print, where print quality is important, varying the printing direction may help, but only to a point because the print head may very quickly be back on the still hot layer. Printing at a lower temperature may help, but reduces overall flow capability, and of course a Prime pillar could be used, with a very low speed to ensure that 12 seconds have passed, whatever the layer time. This delay if evidently much shorter with thinner layers. At a given flow, there i always a speed / thickness trade-off giving more or less constant printing time. Printing thinner layers faster reduces problems with slow solidification and bleeding, but excessive speed introduces inaccuracies and vibrations in the X-Y plane.
Cura does not have an automatic extrusion multiplier adjustment.
Research is needed:
[*] To separate the factors and to find out which effect is nozzle related, which is temperature related and which is material related.
[*] To find a proper way of getting the calibration factors for a new material / nozzle / temperature combination.
[*] To find a proper user interface in Simplify3D for this function (to replace the attached Excel file).
In order to refine the analysis and separate the factors related to the extruder / print head and the factors related to the material, more test results are needed, with different materials. I printed a simple cube, that I elongated using the controls in Simplify3D. I worked with 2 grams samples and a precision scale with 0,01 g accuracy. I'll use what's left of the green PLA to test at lower temperatures.
Flow is currently calculated from geometry and speed, and modified with a fixed "Extrusion multiplier" factor associated with the Extruder.
- Raw and ready to use data
- (35.13 KiB) Downloaded 4 times