Advertisements

Reversal Zits: Early Action Variations

While cranking out some Tux Cookie Cutters, I varied the Reversal settings to see what effect they’d have on a single object with a smooth perimeter. I’d previously settled on 25 rpm for 125 ms with no early action, so this series tests three different times with early action turned on.

Position 1, where the perimeter threads join. Yes, I have Jitter activated and cranked up to something like 10, but it obviously has no effect on this object:

Position 1 - Reversal 125 100 50 - early

Position 1 - Reversal 125 100 50 - early

Position 2, where the nozzle enters from the outside to start a new thread. The snot hanging off the end makes for an ugly wad:

Position 2 - Reversal 125 100 50 - early

Position 2 - Reversal 125 100 50 - early

Position 3, another nozzle entry point:

Position 3 - Reversal 125 100 50 - early

Position 3 - Reversal 125 100 50 - early

Early Reversal action simply doesn’t work well. With retraction times sufficient to prevent drooling, stopping the extruder before the end of the thread produces unacceptable gaps and starting it before reaching the thread produces hanging snots when the nozzle passes over an existing wall.

Shorter retraction times produce strands all over the object, because the extruder still contains pressurized plastic and drools.

I’d previously discovered, although I didn’t write up, that unbalanced Reversal times didn’t provide any benefit: inhale and exhale times must be essentially equal to prevent either starving the first part of each thread or serious drooling. So there’s really only one degree of freedom: the total volume of plastic = rpm x duration.

Perhaps having separate early action times would help: adjust the shutdown and startup delay times independently of the total Reversal inhale/exhale time. Right now, those delays are simply the inhale/exhale times, evidently assuming clean cutoffs and startups, which obviously isn’t the case.

And, alas, the Reversal Threshold bug remains unfixed, so you (well, I) can’t tell Reversal to not operate across short motions like the end of one thread and the not-quite-adjacent start of the next.

Advertisements

  1. #1 by Bruce on 2011-09-26 - 08:03

    I have been considering this problem off and on for a while and have been trying to come to a conceptual understanding as to why this happens. This is where my brain is at:

    When the reversal occurs the bulk of the plastic is pulled away from the nozzle outlet, however there is still a string connecting the plastic inside the extruder to the previously deposited plastic. This string gets stretched when reversal occurs and the extruder head moves away from the line end, when the counter-reversal occurs this string gets shoved back out the nozzle before normal printing resumes.

    I’m pretty sure that this is what occurs, but im still exploring the implications of it. Some conclusions I’ve been considering that require testing:

    – Early reversal in all likelihood will make the problem worse. You stop depositing plastic where it really should be, and you make this “reversal string” longer.

    – A stronger reversal action (higher rpm) will just produce a longer “reversal string”

    – A more effective action might not require much reversal at all, is it enough to move the nozzle away from the line end at a rate fast enough so as to break the “reversal string”? would this pull the intentionally deposited plastic away as well?

    • #2 by Ed on 2011-09-26 - 16:28

      still a string connecting the plastic inside the extruder to the previously deposited plastic

      That seems to not be the case, because the zits show no sign of stretching in the direction of travel. They’re basically a blob at the end of the thread, with nothing much leading to the start of the next thread.

      It’s certainly true that pulling the plastic away from the nozzle at low speed doesn’t work at all; I can get meters of hair-fine thread without any trouble at all. Pretty much by definition, the plastic inside the nozzle is molten, so it doesn’t exert much force at all on the object, which means it doesn’t “snap”, even at zero pressure. I think that means the Ultimaker “high speed snapoff” trick probably works only after most of the pressure comes off the molten pool inside the nozzle and the remaining molten plastic can’t ooze fast enough to supply more than a hair.

      I’ve taken several crappy high-speed videos of the nozzle while extruding that show some of what’s going on (in addition to showing that I need much better lighting and focus, alas) that I must pull together into another movie…

  2. #3 by Rob Giseburt on 2011-09-26 - 23:43

    About the Ultimaker “high speed snapoff” trick, this is what you get:

    IMG_1069.JPG

    :-)

    • #4 by Ed on 2011-09-27 - 07:43

      Attack of the Spiders from Mars!

      The spikes on the crown show the stringing problem very clearly. Here’s a tight crop of that section:

      Ultimaker strings

      The extruder pressure supplies a few millimeters of nearly full-size thread at the start of each high-speed motion, which then turns into a hair stretched all the way to the start of the next thread. It looks as though the extruder starts running slightly before the nozzle reaches the start of the thread, although that may be due to the hair contracting and pulling a bit of molten plastic from the next thread.

      Obviously, there’s room for improvement!