Why 16x micro stepping is not a good idea for mixing hot ends

As title really. Just written a post on my blog which, if you use a mixing hot end, might be of some use. Link here [somei3deas.wordpress.com]

Edited 1 time(s). Last edit at 06/08/2018 07:52AM by deckingman.

---

[somei3deas.wordpress.com]
[www.youtube.com]

Very interesting read!
I used to use Raps128 drivers for my Diamond3_hotend. Using 0.9° steppers for extrusion is a bad idea IMHO, especially with Bowden tube. They tend to be pushed backwards when pressure builds up in the Bowden.

I have an ( overkill ) idea to get best microstepping AND retraction: You could use two different gear ratios, decoupled with freewheeling bearings.
While rotating in extrusion direction, freewheel 1 is engaged with a gear ratio that adds steps/mm. (ie. 1:16 muSteps and 1:4 gearing )
When retracting, the other freewheel engages and you can pull back very fast with a 1:1 ratio.
Only issue with this scenario is: max. unretract speed would be lower and retract/unretract would need different steps/mm.

Another way to increase steps/mm is: using thinner filament. Have you seen my 0.8mm filament experiment? I mainly started it, because I wanted to use smaller extruders for direct drive, but they address your problem as well. It's steps/mm is ~4 times higher. Unfortunately, there is no mixing hotend for 0.8mm filament ( smaller melt chamber )

Edited 1 time(s). Last edit at 06/09/2018 01:56AM by o_lampe.

Better still would be to have effective pressure advance strategy. That way you could negate the need for retraction, or at least reduce it so much that retraction speed would be less important. Unfortunately, although pressure advance with Duet firmware works really well on straight moves, I've never been able to cure the problems that it introduces when doing segmented arcs, so I can't use it.

---

[somei3deas.wordpress.com]
[www.youtube.com]

Wouldn't that require a different object description than .stl? (new slic3rs and all that...)
RRF always claims to be segmentation_free, but the .stl files aren't....

Quote
o_lampe
Wouldn't that require a different object description than .stl? (new slic3rs and all that...)
RRF always claims to be segmentation_free, but the .stl files aren't....

Not sure I understand the question. Slicers deal with arcs and circles by chopping them up into small straight line segments. On my machine with my hot end etc etc, duet firmware pressure advance works really well in dealing with the build up of pressure over longish moves and almost negates the need for using retraction, but I need to use a high value. The problem is that when I set my machine to use a high value of pressure advance, I get really jerky behaviour when doing circles and arcs. This behaviour used to be limited to only when I use multiple extruders but later firmware versions give me problems when using just a single extruder.

---

[somei3deas.wordpress.com]
[www.youtube.com]

I meant, that the .stl files already inherit the segmentation of arcs and circles. On CNC machines arcs are described with G2&G3 moves. If we had a slicer that could produce such Arc moves ( based on other object format than .stl) , it would be seen as one long move from RRF and the pressure advance would work better.
Maybe it's an option to use pressure advance from one retraction-command to the other instead of track_by_track? Jerk setting would have to be pretty high then....

Quote
o_lampe
I meant, that the .stl files already inherit the segmentation of arcs and circles. On CNC machines arcs are described with G2&G3 moves. If we had a slicer that could produce such Arc moves ( based on other object format than .stl) , it would be seen as one long move from RRF and the pressure advance would work better.
Maybe it's an option to use pressure advance from one retraction-command to the other instead of track_by_track? Jerk setting would have to be pretty high then....

Ahh now I understand. Yes I thought that G2 G3 arc moves would be a cure too, but I've had this same conversation here on these forums and was told that even with G2 and G3 arc moves, the firmware would segment them. So it won't help.

---

[somei3deas.wordpress.com]
[www.youtube.com]