I have come up with a design that I think addresses all of these issues. It looks like this.
 |
| Filament, gap, level & Temperature Print Test Object |
Filament
It turned out that the new filament delivery system required an equally new level of quality control in the PLA manufacturing process. That has now been accomplished. I am told that the current ABS should work equally well in a well calibrated printer. But, to accurately test whether or not the new filament is up to the task there are other factors to consider as we design a test object to attempt to print
In the above design, the objects for both left and right printing are identical and relatively simple to rule out STL design issues. If one color prints correctly we can assume the other print jet should be able to handle the design also. I will only use a cartridge in the right side of my machine that has been tested successfully in the left side.
GAP
Clogs in Cube 1 and Cube 2 were almost always caused by a gap issue. Either the user put the plate on without properly seating it or the gap was so small that the print jet touched the glass plate.
While the Auto Gp on the Cube 3 is nice to have, in my particular machine it always leaves the print jet touching the table. I have to manually adjust the gap.
Gap issues are compounded in a two print jet machine and that is especially true for the earliest runs of the 3rd Gen Cube, We have to ensure that both print jets have the proper gap, not just one. Recently delivered Cube 3 printers should no longer have this issue.
The best way to determine if the gap is correct is to print an object that is no more than a few layers thick so that we can analyze how smoothly the layers are printed. In this case, the square areas are .70mm thick and large enough to give us an accurate picture of our gap situation.
Level
In all three Cubes, one can set a perfect gap at the center of the table only to find that the print jet is too close on one corner or the other. The Cube 3 is INFINITELY easier to set the level of the print plate. And, I am testing a new firmware (V1.09A), which may be already be released by the time you read this, that refines the auto leveling process even more precisely.
The reason why the large, thin pads are arranged in the corners of the print table is to allow us to assess the level of the print plate. If we get a nice smooth result in one corner; but, an irregular result in another corner it should help us decide if we need to pay closer attention to our print table leveling.
Print Jet Consistency
In my machine the new filament works flawlessly on the left side. But, I have managed to clog one of the cartridges that worked worked perfectly on the left side when I moved it to the right side. That would lead us to believe that it is either a filament issue or a print jet temperature control issue. But, unfortunately, the two color design I was trying to print was very complex and actually was a poor choice for my first two color test of the new filament.
One of the Cube 2 machines used in our YouthQuest 3DThinkLink classes had a print jet heating failure. From the earliest RepRap days heating failures have been common. So it is not out of the question that the right print jet has a temperature control issue causing it to behave differently than the left. But, to actually prove that I need a simple print task being performed by both left and right print jets.
There is a related scenario where instead of the temperature dropping too low, the melted filament backs up to a point beyond the heated area and cools. If this should happen we would, most likely, see a separation where the extension enter the print jet tip housing. I haven't seen that as yet.
The simple circular towers are 20mm high. They provide the best test for finding most print jet temperature issues. Since both print jets are performing the same task on two different areas of the print plate we can assume that a failure of one color is an indicator that temperature control should not be discounted as a clog causation. If a color fails before completing a tower, I will return the printer for evaluation.If this test completes; but, I still experience print failures, I will extend the tower
Current Observations as the Objects are Printing.
I felt the need to redesign my test to reduce the time. The new print job is reported to be about 4 hours vs. 12 hours for my first design. This is an image of the earlier design that I aborted to start running the more efficient test. It was NOT stopped due to a failure. It was stopped to allow me to run version 2 of the test file.
Even so, it tells me what I needed to know.
 |
| Aborted Test using version 1 STL |
The thin pads printed completely and the towers printed to about 7mm tall before I stopped it.
Essentially everything looks positive. I am printing with Red in the left side and Blue in the right. Previously, the Blue has been used to create several mid-sized objects and the results were beautiful.
Because the thin pads extend outward, to the corners, I am able to observe the quality of the print. Based on the pads in the above image, I thought my gap was probably a bit too large for both print jets. But, it turned out that the STL had issues that cause the pad to look a bit rough. The new version, while not yet finished, has printed the pads perfectly. So I think the gap is actually OK.
So gap and level can already be eliminated as the cause of a clog. All that is left is the potential for temperature control or filament issues should it clog. And, while I didn't let it run to completion there was no clog at the 7mm height. All towers were of equal height.
That probably means that my earlier clog, with the untested complex object, was caused by an STL design issue,
 |
| Complex two color print - right jet failure |
Perhaps the design was too flat to be supported well enough. Here is the bottom of the piece that exhibits drooping from being too horizontally to shallow. Note how rough it is. THAT is a DESIGN issue.
 |
| Complex two color print - bottom view |
I have no idea if this condition alone could cause clogs. But, it's certainly possible.
In any case, I'm running now version 2 and it looks better than the first version. The great news is that I should be able to see a problem shortly after it happens because all towers should be equal in height for each layer's print cycle. When I see a difference of more than 2-3 mm I can abort the print job.
NOTE: Great News!!!! No more little fragments falling on the print table from the parking and purging operations! The print table is absolutely perfectly clean! Be sure to update both your client and the firmware to the latest versions.
Be sure to check back to this same article later to see the results.
UPDATE: TOTAL SUCCESS
I declaring the new PLA works! This is the finished print. No Clogs!!!!
Hi Tom,
ReplyDeleteI hope the new cartridges are the solution to some of our problems. I wonder if the pla itself is improved or just the loading of the cartridge spools to prevent overlapping and stuck filament. I had my green pla fail about 80% into the build on the right side after switching from the left but the layers look perfect leading up to it, and my brother in-law's cube did the same. I don't think it would get this far into the prints if it was a heating issue or a leveling/gap issue.
I took apart my cartridge to find that a clog was not the culprit, it was the extruder stripping the filament when the spool comes under too much resistance. I cut the filament and heated the nozzle enough to dislodge the filament and pulled it through the flexible sheath. Then I fed the cable through the cartridge's extruder opening while turning the extruder with a torx bit screwdriver and fished it up the sheath to the nozzle tip. After that it worked until another jam in the spool occurred after 3hrs of printing.
I noticed that the extruder both pushes filament and retracts it as the print heat moves from one layer or feature to another to prevent blobs from forming so if that extruder slips ever so slightly even briefly enough to avoid a full out clog it creates a imperfection in the print which as you know with FDM gets amplified with the remaining errors.
Sadly, in my opinion the cartridge design to relocate the extruder to the side of the printer and inside the cartridge is a mistake as opposed to the extruder being on the print head. There is a lot of distance and flexibility from the extruder to the nozzle and play can be generated which causes imperfections. Not to mention, when the cartridge reaches empty the filament between the extruder and the nozzle tip gets wasted because there is no motor to push the last foot of filament upward.
It was a soft surface in the old filament, as I understand it.
ReplyDeleteMy personal belief on the new extruder location is that, ultimately, it will prove itself to be a brilliant decision. The loss of filament at the end is the same as in the Cube 1 and Cube 2 designs because they could not afford to let the end of the filament enter the extruder. There was a clip on the end of the filament to keep it from leaving the cartridge.
Only time will tell us if the decision was a sound one. In the meantime, it looks as if the new filament was the primary fix we needed. But, how it works for any individual is going to be dependent on that individual's understanding of the importance of always checking the Auto Gap manually until it is certain that on that individual's machine auto gap is 100% reliable.