Sunday, May 28, 2006
0.5mm Mk 2 extruder trials/tribulations
Getting the 0.8mm nozzle on the current 3D axis/extruder setup showed that we're basically on the right track, but the trail deposited by the 0.8mm nozzle was far too big. Also, the rather flat profile of the 0.8mm nozzle tended to drag plastic trails around the surface of the workpiece.
So I made a 0.5mm nozzle with a more conical profile. This extrudes a 0.8mm filament, which with care and attention might make a trail narrower than 1mm. Maybe even not too far away from the Stratasys' 0.6mm minimum reliable extrusion width. Interchangeable nozzles continue to be a great idea.
But lo, the pressure generated inside the extrusion mechanism is higher with the smaller nozzle, even when the motor speed is decreased to allow for the reduced throughput. This means that the feed mechanism is now much less tolerant of dodgy feedstock, and I have had a great number of jams with my last batch of somewhat lumpy feedstock.
After a few attempts and unjammings, the entire heater barrel assembly was extruded from the PTFE rod. It just popped out. Now, I'll freely admit that the original tapping of the PTFE wasn't the world's best, but all the same I wasn't happy. Fortunately I had a section fo 16mm diameter PTFE and a clamp modified to hold same. As this wouldn't fit in the Afghan lathe, I drilled it manually - hence the 3mm off-centre hole at the far end.
Instead of tapping 15mm of M6 thread to hold the heater barrel, I tapped 20mm. Not only that, but I used a pilot hole of 4.5mm rather than the normal 5mm hole used for M6 tapping. I took the heater barrel off the old concrete extruder because it had a longer bit of thread sticking out the back of it and re-assembled.
After a few more jams, the new thermostat algorithm started playing up, and the extruder heater would inexplicably just turn off. I've asked Simon if we can re-implement the old "bang-bang" algorithm, which seemed to be more reliable and resulted in far fewer pauses for the element to re-attain working temperature.
I never managed to complete a single layer of the test articles using the new nozzle, but I'll try again later in the week once the bang-bang thermostat is re-instated. The output that was produced looked good as far as it went, and once I get the layer's thickness sorted out it should produce some tidy results. Might make some neater feedstock too.
Vik :v)
So I made a 0.5mm nozzle with a more conical profile. This extrudes a 0.8mm filament, which with care and attention might make a trail narrower than 1mm. Maybe even not too far away from the Stratasys' 0.6mm minimum reliable extrusion width. Interchangeable nozzles continue to be a great idea.
But lo, the pressure generated inside the extrusion mechanism is higher with the smaller nozzle, even when the motor speed is decreased to allow for the reduced throughput. This means that the feed mechanism is now much less tolerant of dodgy feedstock, and I have had a great number of jams with my last batch of somewhat lumpy feedstock.
After a few attempts and unjammings, the entire heater barrel assembly was extruded from the PTFE rod. It just popped out. Now, I'll freely admit that the original tapping of the PTFE wasn't the world's best, but all the same I wasn't happy. Fortunately I had a section fo 16mm diameter PTFE and a clamp modified to hold same. As this wouldn't fit in the Afghan lathe, I drilled it manually - hence the 3mm off-centre hole at the far end.
Instead of tapping 15mm of M6 thread to hold the heater barrel, I tapped 20mm. Not only that, but I used a pilot hole of 4.5mm rather than the normal 5mm hole used for M6 tapping. I took the heater barrel off the old concrete extruder because it had a longer bit of thread sticking out the back of it and re-assembled.
After a few more jams, the new thermostat algorithm started playing up, and the extruder heater would inexplicably just turn off. I've asked Simon if we can re-implement the old "bang-bang" algorithm, which seemed to be more reliable and resulted in far fewer pauses for the element to re-attain working temperature.
I never managed to complete a single layer of the test articles using the new nozzle, but I'll try again later in the week once the bang-bang thermostat is re-instated. The output that was produced looked good as far as it went, and once I get the layer's thickness sorted out it should produce some tidy results. Might make some neater feedstock too.
Vik :v)
