Issues with bowden feed...

Does anyone know of an alternative way of connecting the bowden tube to the hotend? The Fisher method of drilling out the PTFE tube and forcing a thread on the end doesn't provide a smooth passage for the filament to enter the hotend. From day one I find I have to sharpen the filament end like a pencil in order for the filament not to catch on the upper opening of the nozzle tube. Even if the process is done carefully this seems a very sellotape and scissors solution. Using a high powered microscope I can see the drilling leaves fine PTFE burrs just waiting to cause an obstruction in the future. I use a compressed air gun to ensure there is no loose debris from the drilling process remaining. These burrs not only create a rough surface but will eventually loosen and fall into the nozzle thru friction with the filament causing partial or complete blockage which I suspect is the culprit of my current extruding problems. Having cleaned out the nozzle and 'smoothed' the PTFE tubing as much as I could I find the filament catches every time I try to insert it.

Is there a brass coupling that is compatible with the Fisher hotend that would provide a much smoother passage for the filament to traverse the bowden-hotend connection?

RRP use the same drill process with bowdens on all their machines, on my Mendel it works very well however on the Fisher they have a different hot end assembly where the bowden tube screws into the heat sink rather than directly into the nozzle, and yes it can be difficult to feed new filament, and yes I've had to dismantle my hot end before now to change the filament!

If you have to use a microscope to see your PTFE swarf then I would not expect it to clog a 0.4mm nozzle.

I suspect that the problem is not so much the tube insertion, but the removal of the old filament, RRP recommend that you first heat the nozzle up to 200C and extrude a little filament then cool the nozzle to 100C before extracting it, this should pull the material out of the nozzle cleaner than just pulling when at 200C, which as likely as no leave more material in the nozzle

---

RepRapPro Mendel 3 Tricolour
RepRapPro Fisher
-Carbon Arms
-Easy adjust Carriage+effector
-axis stiffness mods
HE3D -600 delta
-Duet 0.8.5
-PanelDue
-DC42 Height probe
-RobotDigg metal components
Simplyfy3D
RS Design Spark CAD

You are right that the inner surfaces of the hotend + bowden should be smooth and aligned, manually drilling the PFTE tube will not guarantee a concentric hole or a flush connection to the hotend tube. If those surfaces are not flush then filament will get stuck when changing filaments, or it may cause a blockage if molten plastic makes its way up that high.
I think it would have been better if RRP had precut and pre-drilled the threaded part of the PFTE tube. A special tool to cut the threads and at the same time ream the inside to 2mm would have been better, but would have increased the price considerably.

Quote
Jelle
I think it would have been better if RRP had precut and pre-drilled the threaded part of the PFTE tube. A special tool to cut the threads and at the same time ream the inside to 2mm would have been better, but would have increased the price considerably.

Reamers don't work too well on soft plastics, nor to taps and dies, you get a lot of heat and very little else, far better to use a good quality sharp drill bit, I actually use PCB drill bits to make my bowden tubes, they have a much sharper edge and a more angled geometry. Regular jobber drills tend to be a compromise between sharpness and wear resistance for cutting hard metals, and are not the best for cutting soft plastic.

I think sometimes people forget these are kits, the general rule of kits is don't spend time and resources doing something the builder can reasonably do themselves.

---

RepRapPro Mendel 3 Tricolour
RepRapPro Fisher
-Carbon Arms
-Easy adjust Carriage+effector
-axis stiffness mods
HE3D -600 delta
-Duet 0.8.5
-PanelDue
-DC42 Height probe
-RobotDigg metal components
Simplyfy3D
RS Design Spark CAD

RepRapPro used to tell us to screw the PTFE tube into the brass ends, then drill the ends back out to 2mm. Now they tell us to pre-drill the ends to 2.5mm. I wonder why they changed? It seems to me that the new method will leave a step in the tube for the filament to catch on, where the old method wouldn't as long as the drilling was straight.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

On the Fisher, there is no brass end on the hotend side, PTFE tube is directly screwed in the radiator. So it is not possible to access the tube hole after screwing it. You have to drill the hole before screwing it. Having the same method for both ends is simpler.
At startup, I screwed too hardly and that was totally blocking the filament. Then, next trials showed that the filament friction is very sensitive to the screwing torque, so now, after firmly screwing in the hotend radiator, I unscrew a half turn. That may create a gap between the nozzle and the tube, but normally this is a cold area, so you are not supposed to have melted filament here. As I am using a much higher temperature than most (hot PETG) and as I never get any trouble, I think this 'unscrew' method is safe.
Having a brass end on the radiator side may be an improvement, but will significantly complexify the radiator which is yet an ultra simple part.

Edited 2 time(s). Last edit at 09/09/2015 07:22PM by PRZ.

There is a step, but its not a square step, if your using jobber drills there will be a 30 degree chamfer between the smaller and larger diameter and at the hot end its in the negative direction so filament should not catch, but there is a possibility of them catching at the extruder end. My PCB drills have a 90 degree point geometry instead of the 118 degree geometry of a jobber drill, so the step chamfer is increased to 45 degrees

When comparing the fisher design with the quick set design you notice that the interface between the PTFE tube and the nozzle is a lot further away from the hot tip on the fisher and pretty deep in the heat sink, I suspect that unless you use a lot of heat and/or the cooling fan fails the melt front never gets this high.

Unscrewing a half turn makes sense as I suspect some may over tighten the tube and cause it to contract on the bore, when assembling my fisher hot end I deliberately at first over tightened the PTFE tube, the removed it and re-drilled the bore before assembling again and only loosely tightening.

---

RepRapPro Mendel 3 Tricolour
RepRapPro Fisher
-Carbon Arms
-Easy adjust Carriage+effector
-axis stiffness mods
HE3D -600 delta
-Duet 0.8.5
-PanelDue
-DC42 Height probe
-RobotDigg metal components
Simplyfy3D
RS Design Spark CAD

Is the brass piece on the drive side a new part for the fisher or is it used too on other rrp designs? A solution might be to add a thread on the outside, so you have a separate piece that you screw onto the PTFE tube and has a lip to keep the 2mm drill centered. This threaded piece again screws into the heatsink body, providing a good seal with the top of the nozzle tube. The chamfer that is on the inside of this tube (presumably to help enter a misaligned filament into the hole) is problematic: it disrupts the smooth inner surface from ptfe to nozzle.
If the bowden is perfectly aligned, this chamfer is not needed anymore. It will make the heatsink a trickier part to produce though: the inner hole needs to have two threads: M4 for the nozzle tube, and M5 or M6 for the brass insert, making the M5 section more or less a blind hole.

Assumptions that the melt will never reach the top are wrong: you can easily induce a blockage between bowden and hotend by feeding and retracting the filament by hand. You can argue that this will never happen in 'normal' operation (whereby normal is defined as circumstances where the melt does not flow up that high?), but it might make matters worse when they do go wrong: PLA is prone to degrade a high temperatures, it will degrade into the much less viscous fluid (solidifies much more brittle as the polymers are much shorter). If the nozzle is blocked at the tip, the PLA in the melt zone degrades and can be pushed much further up the tube. In that case simply deblocking by pushing a needle or drill in the nozzle tip will not work: it will also be stuck at the bowden interface.

There's a few different ways of drilling out the Bowden tube. The older method of fitting the brass union, and drilling into it with a 2mm drill, is, I think, the best; it ensures the hole in the Bowden tube is central, and concentric with the hole in the brass union. You can do this to both ends of the tube with the Fisher, if you want.

We changed to using the 2.5mm drill method, as the Fisher and the Quick-set nozzle (on all our other machines now) screws straight onto the Bowden tube, so you can't pass a 2mm nozzle through it. Using a 2.5mm drill into the free end of the tube usually works well, though it is possible to offset the hole, particularly if the hole is offset in tube as manufactured, which creates a step. The reason for using a 2.5mm drill is that the tube compresses a little as it's threaded into the heatsink or union. We experimented, and found that any size smaller than 2.5mm is too small for 1.75mm filament, once compressed.

On the hot end/heatsink end of the tube, you also need to drill a little further, as the tube goes a little deeper into the heatsink (around 12mm) than the union (8mm).

When you construct the hot end you should always 'feel' the filament path for any constriction, by pushing a piece of filament into the Bowden tube and down into the nozzle, and sort that out before mounting the hot end.

Ian
RepRapPro tech support

Thank for this valuable thread and advice!
I remade the Bowden tube after getting a PCB drill. What a difference! Sharp and clean finish. I also pre-threaded the tube using a M4 nut before attempting to screw into radiator. Also carefully measured the length of thread and marked the guide filament so I knew when it was fully seated. 1/2 turn back suggested too. Shortened Bowden tube to have better 'fold' profile as effector moves up. Added cable tie to keep it upright at effector top.

And now a smooth insertion of filament. What a difference.

Now printing predictably. But only after resetting Slicer to default settings (too much filament in wrong places) !
Lesson for me - if having to keep adjusting retract feed and speed and temp I ought to understand the underlying issues.

Thank you all!

Edited 1 time(s). Last edit at 09/11/2015 01:40PM by rogerdiver.

I found that by drilling out the tube to 20mm rather than 10mm substantially reduced the pressure required/friction at the point where the bowden enters the hotend assembly.