Bowden tube and Filaments frictions influence on feeder capacity.

[Torgeir 288]

Hi Folks,

This issue is something that is never been taken to the very end and solved!

Searching this problem in here do not help very much, but when Google from the outside, you get inside! (Here.)

So you could say I've been seeing lots of people addressing this problem, -but been "stopped" by some comment that should never been said/asked!

I have to say, I'm very sorry for that!

OK., I've made a simple drawing of the setup with this issue, just something to start some discussion if you want. Yes, this is about how friction can work and make problems.

Friction is a coefficient that is "more or less" constant for a given material and can never be higher than 1, or less than 0.  

What is the problem then, -it is the force that is needed to overcome the friction!

Now we can start the discussion.

The issue is that; the track after the feeder wheel is following the outside arc of the bowden tube and will create more friction than the clean filament.

As the pressure from the feeder is directed into an "180 degrees" bowden tube, we will have a situation like this:

PS. Borrowed the picture from here: http://www.fabbaloo.com/blog/2016/1/31/bowden-style-3d-printings-critical-element-the-tube

To help a little here, we have some very good tools; Solid Works and Fusion 360.

Here you can do the calculation of the friction and you'll see some pretty interesting numbers!

I do not think much more is needed in here?

Ok. You're next.  

Thanks.

Torgeir.

Edited December 2, 2016 by Guest

[rigs 735]

Work in progress....

proposal ultimaker

[neotko 1,417]

I did aome findings about what happens without knowing what was really happening on the Bowden at that time.

https://ultimaker.com/en/community/19330-esteps-and-ultimaker-o-and-the-magic-missing-10

Clearly the filament also, apart of friction, flexes like a spring.

That's why when I calibrate my 1.75 feeders I * the feeder math supposed real number by 1.055. For 2.85 the number is almost a flat 10%

Edited December 2, 2016 by Guest

[Torgeir 288]

Work in progress....

proposal ultimaker

 

Hi rigs,

This is great info, thanks.

Regards

Torgeir.

[Torgeir 288]

I did aome findings about what happens without knowing what was really happening on the Bowden at that time.

https://ultimaker.com/en/community/19330-esteps-and-ultimaker-o-and-the-magic-missing-10

Clearly the filament also, apart of friction, flexes like a spring.

That's why when I calibrate my 1.75 feeders I * the feeder math supposed real number by 1.055. For 2.85 the number is almost a flat 10%

 

Hi neotko,

Yes, I've been reading your posting about this issue with interest and compared it with the "MK8 experiment" i did sometimes ago. I’ve had to do the "reverse" engineering on the requested number of e-step for the UM2 in order to find this "theoretical diameter". So the effective diameter for the 8 mm knurled feeder wheel found using this approach appear to be approx. 7.2 mm(?). I'll think you are right in that the feeder is compressed and build/accumulate up pressure by making a sort of S pattern inside the bowden tube.

Torgeir.

(Right now down Sout in Spain. )

[neotko 1,417]

Indeed Torgeir. To avoid the S (called Hysteresis) compression I use a 6mm-2mm dia Bowden (although I could just use a 4-2mm). Today I finished (almost) installing a Bondtech+Um3 feeder mix (using the cheap drivegears of bondtech and reusing the um3 feeders) and after finishing fixing stuff my 5.5% did result in close to perfect esteps again. So I think is quite accurate to get good results. I been using this since I did that post using MK7's and the same when I moved to bondtech cheap DIY kit. Ofc I think that you can only get esteps at 'certain temps'. So, when you get your perfect spot, you can move up temp/speed to avoid changing the amount of pressure back. I'm quite obsessive about getting perfect toplayers so I try to control as many parameters as possible to get constant repetition.

[Torgeir 288]

Hi Folks,

As we now can see that the additional force needed to overcome friction in the bowden tube can be an issue in order to have enough pressure to assure proper extrusion, lets see how this thing can influence our print object.

If we study the movement of the extruderer when moving at different places on the bed, you might see that the feeder inside the bowden tube twist a little when moving around the bed. The worst twist difference occur from right inner corner to left outer corner, or the other way around.

This mean that the feeder track go offset from the top of the bowden tube, due to this the friction factor go a little down, so now the effective pressure at the feeder increase – so we will have more filament extruded out of the nozzle. So, we now actually realize that the feeder pressure might vary at different places on the heat bed. This is not very visible when the extruder moves around all the times, but will be obvious when the extruder work for some time at a details on a large print object at different places.

If we assume the “worst” case, thinking that this movement can create under extrusion, but probably more often seems like -just before under extrusion or some opening in lines..

Then to the issue when we try to extrude more than the printer can do. In this case we’ll compress the feeder inside the bowden tube (yes it’s really work as a compressed spring), if we try to rice the temperature the printing improve some, but it’s not good. In this situation, when trying to adjust in order to improve – any normal “small” adjustment do nothing. Retraction do not work very well and suddenly the extrusion stop.

(Actually in this above case, we might not know that the printer is reduced by high friction force when we try to setup for a default speed.)

Yes, I know there is many things that can be the reason for this scenery, but here we look at the feeder working against friction forces and what “can” happen due to this.

You know, the filament for a 3D printer is like the petrol for a car!

What is interesting here to see is that feeder friction can lead to problem that put the focus away from this issue, because so many other thing occur due to this one issue.

Thanks.

Torgeir.

[neotko 1,417]

Really really interesting link posted by @rigs on twitter

http://www.genapart.com/2016/11/14/quel-diamètre-de-fil-utiliser/

Ofc is a 2.85 vs 1.75 but the data about pressure could be interesting for the subject.

Edited December 4, 2016 by Guest

[Torgeir 288]

Hi neotko,

Very interesting.

This might be "our" rigs..

I'd use to think, I like both (2.85 and 1.75).

But you have a UM3, so..

Torgeir.

[Torgeir 288]

Hi Folks,

To resume a little here, we know that the pressure is very important to be able to control precisely our other printing parameters. If the pressure is at the correct level, other parameter changes will be “easy” to adjust, detect and evaluate.

This is like a clock frequency in a computer or a reference voltage for controlled processes. So pressure is a very important parameter for us to use and control.

As we now have looked into how friction in between the bowdwen tube and the filament can vary the effective pressure to the extruder, it’s now in place to see how we can improve/reduce the friction in order to have a more “constant” extruding pressure.

The first obvious thing is to change the feeder track line, I.E. move it from the top of the feeder.

This is easily done by mirroring the feeder unit (here reflecting the UM2 type feeder unit). This will move the position of the knurled wheel from the outside of the filament to the inside of the filament making the track going inside of the bowden tube. This will improve (friction constant will descend a little) the system as there will be quite more “effective” force at the extruder.

However, we know there is another problem as well in this setup, -the hot stepper motor shaft were the knurled wheel is mounted. A heated feeder wheel will not have as good grip into the filament when warm versus a cold wheel. The filament may/will slipp off the feeder wheel in such situation. So we’ll need an improvement/fix to avoid this issue as well.

I’ll try to be on the topics as much as possible, but the feeder wheel/mechanism is directly involved in the feeding process and is impossible to let out.

We know that this old stock feeder (UM2 type) can grind the filament when cold (without any gear to improve the force). So why not use a belt and two pulleys of same size, say GT2 16 – 5 (16 teeth and 5 mm shaft hole.) This will insulate the stepper and knurled wheel temperature wise. Using a belt is more flexible (easy to gear) and more precise (“no” play) in between here. Further, you are using the same e-step number as before and the only thing to change is the rotation direction. This can be done in hardware by “swapping two wires of any of the two coils on the stepper motor connector” -or just do an update/change of the firmware in your printer. This latter is handy, but now you have a non standard firmware in your 3D printer.

I do like modifications if it’s improving the printer, but I don’t like modifications that’s change my firmware cause this ruin the ease of upgrade when factory firmware upgrade is a must in order to use/have new “issues”. For those having a genuine UMXX, this can be a benefit if you're selling your printer etc..

So, the improvement here is: The feeder need less force to print than before, less compression, more precise retraction and better overall performance. You can now also reduce the feeder wheel pressure, leaving less signature on the filament (deep track line) that’s improve the ease of feeding.

Using this little modification and you can leave the filament roll where it should be without any issue.

There is still some more nice thing to do for improvement, but this will be next when the time allows..

Thanks.

Torgeir.