how to size bowden tubing?

Is there any tribal knowledge about the subject of Bowden tube size?
I am building a larger-sized delta so the tubing is both curved a lot and longer in length than, say, a Kossel Mini.
I'm using '3mm' filaments, which can of course deviate from filament to filament.
So...it seems like the best approach is an ID just close enough to enclose the largest expected filament OD (lets just say 3.1mm for now).
This limits 'bend' hysterisis ie the filament trying to coil up inside the tube.
Too close a fit though and too much of the ID starts to touch the filament and cause excessive friction.

As for the OD of the tube, you of course have to fit the fittings at each end, but if given a choice, a thicker ID means less compression of the tube. During back-pressure the filament and the tubing will both experience compression/tension. Cant' do anything about the filament but you would halve the hysteresis if you completely eliminate the Bowden tube compression.

Which brings to mind about material? PTFE is the usual choice for obvious reasons but in my opinion is not a universal solution, if one considers the filament friction vs the tube compression stiffness. FEP tubing is like a hard teflon so I'm thinking it might be better than PTFE, although it does require more bending force and the problems that incurs like tilting forces on the effector platform, etc. It also depends on how smooth the tubing ID is. I have a piece of PTFE tubing here that has more of a satin finish, which would be less desirable with any filament that itself has a rough surface, which is almost always the case because of the hobb-marks; in my case with 3mm the hobb clamping force is fairly high so those teeth marks are very visible. So for 'fuzzy' surfaced filaments it seems like a hard smooth tubing ID would be best.

Any logic I am overlooking?
thanks

I wouldn't use 3mm filament with a Bowden system. A Bowden tube for 1.75mm filament imparts quite a lof ot sideways force to the hot end, which used to cause the head angle to vary on my Ormerod printer. With 3mm filament, this force will be around 5 times higher, because I expect the sideways force to increase with the cube of the tube diameter.

---

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].

Yeah I am kinda of stuck with the 3mm components for now.
However, I found that if the tube isn't too short the torque on the effector isn't too bad. The stiffness of the tube seems to dominate over the stiffness of the filament. I also put a string of rubber bands between the top of the rear tower and and the bowden tube about 8 inches up from the effector. This does a surprisingly good job of countering the pull of the bowden at all positions within the chassis and converting it into upward tension which is taken up by the motors. (at botto
BTW, I thought beam-bending force was 4th power of beam thickness? Been a long time I learned so maybe I mis-remember. I had to build an apparatus at work that graphed the bending stiffness of a hair!

I am guessing you'll soon find out why most people recommend using 1.75mm filament with a Bowden arrangement.

As for the question that is the subject of this thread, the simple answer is: as short as possible!

AndrewBCN: My biggest problem with my Kossel Mini was a super-lame hobb assembly not designed for the extra back pressure of 3mm filament. I ended up buying a beast of an extruder that fixed my drive problems. I also admit that I want to preserve the ability to use soft materials like Ninjaflex, which I hear is basically impossible with 1.75mm in a Bowden.
It also seems like there would be a proportionate increase in filament compression stiffness, which reduces hysterisis, a good thing.
Also, since I am designing a 300m-bed system with larger overall dimensions, including the Bowden tube, then it almost seems preferable to have a heavier filament.

I don't think there is a 3D printer that can do everything perfectly, otherwise we would have heard about it. So, every 3D printer is a compromise in one way or another. Same applies to extruders.

For example, most people who use a Bowden arrangement give up on flexible filaments, and on using 3mm filament/hotends. And most Kossel variants use a Bowden setup. I think you are swimming against the flow, here. Entertaining, but it wastes a lot of energy, time and patience. Well, good luck all the same!

Quote
dc42
I wouldn't use 3mm filament with a Bowden system. A Bowden tube for 1.75mm filament imparts quite a lof ot sideways force to the hot end, which used to cause the head angle to vary on my Ormerod printer. With 3mm filament, this force will be around 5 times higher, because I expect the sideways force to increase with the cube of the tube diameter.

What causes a side force on the hotend in a bowden setup?

Quote
wfcook

Quote
dc42
I wouldn't use 3mm filament with a Bowden system. A Bowden tube for 1.75mm filament imparts quite a lof ot sideways force to the hot end, which used to cause the head angle to vary on my Ormerod printer. With 3mm filament, this force will be around 5 times higher, because I expect the sideways force to increase with the cube of the tube diameter.

What causes a side force on the hotend in a bowden setup?

The force needed to keep the Bowden tube bent instead of 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].

Quote
dc42

Quote
wfcook

Quote
dc42
I wouldn't use 3mm filament with a Bowden system. A Bowden tube for 1.75mm filament imparts quite a lof ot sideways force to the hot end, which used to cause the head angle to vary on my Ormerod printer. With 3mm filament, this force will be around 5 times higher, because I expect the sideways force to increase with the cube of the tube diameter.

What causes a side force on the hotend in a bowden setup?

The force needed to keep the Bowden tube bent instead of straight.

OK, duh. Sorry, brain must be off today. What are the other reasons 1.75mm works better for bowdens?

Friction, the surface area is lower with 1.75mm filament, therefore allowing for better torque transmission.

Weight - the main reason for going bowden is weight. 1.75mm hardware may be negligibly lighter, but the tube is more flexible as well allowing for better working weight and less torque lost through parasitic forces.

Retraction speed may also be better with 1.75mm filament - just a hunch here, I have no scientific data to back this up.

---

My Personal Blog. Build blog.
[engineerd3d.ddns.net]

Modicum V1 sold on e-bay user jaguarking11

I've browsed for information regarding the pros/cons of the two filament sizes and I have to say that the discussion here has been much more informative and thought out than anywhere else I've encountered.

Has anybody thought of putting an second extruder with a small, low-torque stepper motor on the carriage in series with the main extruder? The idea would be that the second motor is there to take up the slack but not have so much weight moving around, and the primary extruder with the bigger motor would have enough torque to make sure the smaller motor didn't get overpowered. It could be used to deal with Nijaflex and other situations where a long bowden was problematic. The direct extruder would be built light, with maybe a 20mm NEMA14 stepper weighing about 100g, but the steps/mm and the arrangement of the drive gear would be identical to keep things in sync.

Thoughts? I have an arrangement on my delta that allows rapid changing of the effector so that I can change nozzles quickly, and the board I have has one stepper output that can be used to drive two motors simultaneously (usually Z). So an arrangement like this would only be used in special situations.

Just a crazy idea...feel free to shoot it down.

I'm building a 55cm high print volume delta and I'm facing issues with my bowden tube despite being for 1.75mm filament.

1) I had to use some ninjaflex to suspend the ptfe tube half way from the extruder to the head so it is not falling on the plate when the head is close to Z 0;
2) the tube is 60 cm long, so there are lots of friction, and the current extruder I'm using cannot fight both the friction and the hotend at once, the grip is lost when the filament rach the hot-end. I'll try a fullmetal airtripper version soon, hoping for better results thanks to a more precise grip adjusment, I hope that the stepper will be able to follow!
3) all in one, the cabling+ptfe tube is heavy, and the weight may just have an impact on the print precision

That being said, I was thinking... What if a Z4 stepper would be added? Say from top of the delta, that stepper would arange for the extruder to always be around 20cm above the print head. meaning a short 25cm to 30cm ptfe tube could be used.

That could minimize both weight and torque impact on the print head.

Now, that needs implementation. I'm using Marlin.
Ideally, that Z4 stepper should always move the extruder at Z+200. that would probably be a good solution. We could also play with either math.avg(Z1,Z2,Z3) or math.max(Z1,Z2,Z3) but neither of these two adresses horizontal distance fluctuations between the hotend and the extruder so I'd keep Z+200.

I'm adding a draft I skeched to help share the idea.

Schema

Any thoughts about this? Anybody feels like he could implement that in Marlin? I definitly can't :-/

Edited 1 time(s). Last edit at 01/07/2016 09:30AM by ImShogun.

Quote
ImShogun
...That being said, I was thinking... What if a Z4 stepper would be added? Say from top of the delta, that stepper would arange for the extruder to always be around 20cm above the print head. meaning a short 25cm to 30cm ptfe tube could be used.

That could minimize both weight and torque impact on the print head.

Now, that needs implementation. I'm using Marlin.
Ideally, that Z4 stepper should always move the extruder at Z+200. that would probably be a good solution. We could also play with either math.avg(Z1,Z2,Z3) or math.max(Z1,Z2,Z3) but neither of these two adresses horizontal distance fluctuations between the hotend and the extruder so I'd keep Z+200.

I have been considering doing something similar in RepRapFirmware. I'm sure it would be necessary to vary the extruder drive Z height with the effector XY position as well, so that the diagonal distance between the extruder drive and the effector is kept constant. However, I'm not sure it would be any better than a flying extruder.

---

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].

Google Uniqueprototyping.com. I have their cold-end extruder and it solved my filament feed issues. It has a super strong holding force! The bulk of the cost comes from the down-geared stepper. They also have multi-extruder designs.
Something to consider though: small OD filaments require more speed and less torque than big ones.

ImShogun,

Are you using 1.8mm ID tube, or 2mm tube? I recently bought some bowden tube rfrom robotdigg, and it turns out to have an ID of 1.8mm. Filament has high friction in there, especially after being squashed by the extruder assembly.
I have some other bowden PTFE tube that is more like 2mm ID, where the squashed filament moves freely.


Also, another option is something called a "Flying Extruder". I'm about to install one on my delta.

[forums.reprap.org]