Welcome! Log In Create A New Profile

Advanced

Retract length ...

Posted by lolocaledo 
Retract length ...
November 08, 2015 02:59AM
Hi all,

On your Delta printer with bowden extruder, how much is your retract length ?
I've read somewhere that 2mm was the max length on direct extruder but I think it could be more on bowden due to the flexibility.

Have a nice day
Laurent
Re: Retract length ...
November 08, 2015 03:20AM
i get decent results with 6mm retract at 60mm/s
Re: Retract length ...
November 08, 2015 05:22AM
Depends on hotend.
E3d, clone, j head, etc.....

But normally i do between 2 and 4.
Just play with the settings till no stringing.
Re: Retract length ...
November 08, 2015 06:00AM
I have a 600mm Borden tube and I have to use 8mm retraction to avoid blobbing. This is with a genuine E3Dv6 hot end.


Delta printer calibration calculator, mini IR Z probe, and colour touch screen control panel: [escher3d.com]

Large delta printer, and other 3D printer blog postings: [miscsolutions.wordpress.com]

Disclosure: I have a financial interest in sales of the Panel Due, Mini IR height sensor, and Duet WiFi/Duet Ethernet [www.duet3d.com].
Re: Retract length ...
November 08, 2015 06:11AM
You may find this E3Dv6 help page very useful - amongst other things it recommends the amount of retraction for different kinds of extruders.
Re: Retract length ...
November 08, 2015 10:02AM
10 mm at 150 mm/s, 5000 mm/s² with some old unbranded head. 8 or 9 mm might be enough too but 10 is a nice number smiling smiley

Remember it is also important to retract quickly enough. If you are retracting slowly you get a blob (plastic leaks till the retraction is finished). It is good if you can retract in 0.1 s or quicker.
Re: Retract length ...
November 09, 2015 12:19PM
I retract around 3.5mm at 100mm/s. Wheather or not it ever reaches 100mm/s is another subject. I do not get any blobbing. There are things you can do to eliminate some of the slop and reduce retraction length. Remember through, if you retract too much it may cause jamming. Some hot ends have a very fine heat break, others do not.


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

Modicum V1 sold on e-bay user jaguarking11
Re: Retract length ...
November 09, 2015 12:56PM
Be careful with too much retraction, as if your hot-end is not optimal, or has a very short transition between hot and cold, you can cause jams by having too long of a retract. This can be hidden until you print gcode with many retractions in a small timeframe.

I find that there are ways of combating the drawbacks of a bowden setup with slicer settings, like the coast and wipe functions of simplify3d, or the similar options in KISSlicer and other free alternatives.
Re: Retract length ...
November 09, 2015 01:29PM
Quote
n8bot
Be careful with too much retraction, as if your hot-end is not optimal, or has a very short transition between hot and cold, you can cause jams by having too long of a retract. This can be hidden until you print gcode with many retractions in a small timeframe.

I find that there are ways of combating the drawbacks of a bowden setup with slicer settings, like the coast and wipe functions of simplify3d, or the similar options in KISSlicer and other free alternatives.

The thing is that you can eliminate a bunch of this by simply eliminating gaps in your bowden structure.

I may do a video of this at some point. But the points I have been able to identify to reduce slack are the following.

The hot end fitting connection. You generally get a 3-4mm gap between compression fitting and hot end. This can be eliminated by fabricating a slug that fills in the space with a 2mm hole in it.
The same thing goes for the extruder side. I got around that by designing my extruder with a built in slug and very little clearence between gear/bearing and feed position.
The quality and type of compression fittings. You should also ensure you seat the tube properly to prevent working the springs.
One more thing that cant be overstated is cutting the bowden tube properly to ensure a flush seating surface.

Bowden tube locks are optional as far as I am concerned.


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

Modicum V1 sold on e-bay user jaguarking11
Re: Retract length ...
November 09, 2015 03:42PM
You are exactly right about the loose points. I'm speaking even after accounting for these effects, there is a spring effect to the bowden system. This is really unavoidable, as the tube needs to be flexible and the diameter of the filament cannot be exactly the same as the bowden tube or it would bind. I'm even running a 1.8mm ID bowden tube, which is quite tight and requires a geared stepper to drive filament through, and oversized filament will not fit in it. I still see the effects of bowden tube springiness. (Hysteresis is the technical term, I think.)

I use retract length of 3.1mm, 60mm/s and a coast value of 0.8mm in s3d. This works well for ABS, after accounting for the things you mention and having the smaller ID tube.
Re: Retract length ...
November 10, 2015 02:47AM
n8bot, about coast parameter in S3D, what is the purpose ? I user S3D and I've never used this parameter
Re: Retract length ...
November 10, 2015 03:09AM
Quote
lolocaledo
n8bot, about coast parameter in S3D, what is the purpose ? I user S3D and I've never used this parameter

It is one of the main reasons I use S3D, haven't seen such an option in another slicer.

It's a function that stops pushing filament through the nozzle before the end of a printing move, so the last bit is filled with what oozes out of the hotend relieving the pressure in the nozzle before the retraction.
This helps against stringing och blobbing and to me this has helped the most of all the different options out there.
Takes some tuning to get it right, don't want missing plastic on the layers, but is well worth it.
Re: Retract length ...
November 10, 2015 03:54AM
Github version of cura has it too. That said, the github version of cura often crashes on my machine grinning smiley
Re: Retract length ...
November 11, 2015 07:04AM
Quote
n8bot
You are exactly right about the loose points. I'm speaking even after accounting for these effects, there is a spring effect to the bowden system. This is really unavoidable, as the tube needs to be flexible and the diameter of the filament cannot be exactly the same as the bowden tube or it would bind. I'm even running a 1.8mm ID bowden tube, which is quite tight and requires a geared stepper to drive filament through, and oversized filament will not fit in it. I still see the effects of bowden tube springiness. (Hysteresis is the technical term, I think.)

I use retract length of 3.1mm, 60mm/s and a coast value of 0.8mm in s3d. This works well for ABS, after accounting for the things you mention and having the smaller ID tube.

The springiness of the tube is definitely something we cannot cure, but we can minimise with proper extruder placement. In my particular setup the tube is about 50cm or so. Most people mount the extruder either at the top or bottom of the machine. It should be mounted about 40% of the way up on the machine to minimize bowden length while still maintaining a proper arch and reduced friction. At least in my findings.

One more thing that severely effects retraction length is nozzle size of the hot end. I find with the .5mm and up nozzles the retraction length needs to be longer, ~3.7mm is good in my setup, however while not ideal, it works. For .4mm nozzle ~3.0mm is about right in my setup.

One more thing that causes hysteresis or slack is the extruder itself. No matter what kind of geared extruder you have it will always have some lash. Its the nature of gears to have something like that. In automotive applications gear lash is a desired thing as it reduces wear. In our application it increases retraction length and reduces response time. To eliminate this I designed a belt drive extruder with a very minimal belt length. This gives me in effect a 2:1 gearbox with no observable lash and the same response time as a direct drive with no real hit on speed. Coupled with a MK7 gear (NOT the MK8 which is more optimal for a direct drive extruders) it gives me power and speed, so far I can push more plastic than cura thinks is possible with thicker layers and bigger nozzles. The downside is that gear ratio of my design has really two options 2:1 and 2.5:1.... With that said, Its my firm belief that anything higher ratio is a waste as it reduces extruder overall speed (high rpm from a stepper is not part of a stepper design). In my past experimentation with cast gears, not 3d printed ones, I could still feel some lash. These were slanted gears yet they still had lash, a 3d printed set would have been worse.


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

Modicum V1 sold on e-bay user jaguarking11
Re: Retract length ...
November 11, 2015 08:06PM
Thank you for all these good informations.
After some tests, retract for my delta with E3D is 3.5mm at 100mm/S and cost value at 0.4

Have a nice day
Re: Retract length ...
January 10, 2018 04:40AM
Would you mind explaining what do you mean by :

* Hotend side: "This can be eliminated by fabricating a slug that fills in the space with a 2mm hole in it. "
and
* Extruder side: "I got around that by designing my extruder with a built in slug and very little clearence between gear/bearing and feed position. "

?
Sorry, only registered users may post in this forum.

Click here to login