loooooooong f6, like 10+ minutes but works finally
Greetings all;
I've obviously managed to make a loop, trying to defeat a bug in cura
that wants to retract between teeth thousands of times that is not
needed. The code:
...
module innerspline()
{// PETG now on printer, PolyProp (PP?) has arrived. And its a disaster, went back to PETG
// and the nozzle troubles started, plugged up whole damned hotend. 3 times so far
// Magicgoo for PP must be very caustic warnings galore
// new example, low ratio 30/1 harmonic drive floating splined belt.
teeth = 60; // define number of teeth
arcx = ( 360 / teeth ); // arc of circle per tooth
$fr = 0.000000;
// adjust this to just barely get an overlap at inner corners of spline triangle
toothdia = arcx/1.60;// no bigger then needed
// adjust this for outer teeth fit in internal spline
toothrad=53.5 / 2.000;// make radius was 53.0, reduce backlash but dont compress it either
// adjust this to thin or thicken flex valleys
linerdia=51.62; // thins web at bottom of spline, helps flex if larger, was 52.2
height=12;
tranz=height / 2.000;
echo (177,teeth,arcx,toothrad,linerdia) ;
union()
{ // make splines first
for($fr=[ 0: 6 : 359 ])
{
rotate([0, 0, $fr])
translate([toothrad, 0, tranz]) // mv x=[$lr + $r1] y,z=0
cylinder(d=toothdia, h=height, $fn=3, center=true);
// made triangles, now smooth inner wall by making 2 rings, outer same as base of teeth
difference()
{
translate([0,0,tranz])cylinder(d=linerdia+.5, h=height+.01, $fn=60,center=true);
translate([0,0,tranz])cylinder(d=linerdia, h=height+.02, $fn=360,center=true);
} //but the innards needs smooth finish
}
}
}
The above reflects some comments I added as it was stuck rendering with
the blue line showing 999/1000 for around 10 minutes. A rerender after
makeing line one of the difference at an $fn=60 was still 6 minutes plus.
So the difference is being executed 60 times and getting the same results
everytime.
So I closed the "for" before the difference, getting exactly the same render
in .8 seconds. Maybe this old fart will learn how to do efficient code yet.
;o)
But cura is still miss-rendering it, doing a retraction per tooth, 60 teeth,
times layer height=6000 times. No wonder its 4 hours per print. Its spending
half its time doing useless retracts and restores. hell on the filament and
inviting a freezeup.
Thanks all.
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
If the teeth aren't connected by a section big enough for infill then it
will have separate infill in each tooth which would require retraction for
each tooth. That should not be a problem.
On Sun, 4 Jul 2021 at 09:12, Gene Heskett gheskett@shentel.net wrote:
Greetings all;
I've obviously managed to make a loop, trying to defeat a bug in cura
that wants to retract between teeth thousands of times that is not
needed. The code:
...
module innerspline()
{// PETG now on printer, PolyProp (PP?) has arrived. And its a disaster,
went back to PETG
// and the nozzle troubles started, plugged up whole damned hotend. 3
times so far
// Magicgoo for PP must be very caustic warnings galore
// new example, low ratio 30/1 harmonic drive floating splined belt.
teeth = 60; // define number of teeth
arcx = ( 360 / teeth ); // arc of circle per tooth
$fr = 0.000000;
// adjust this to just barely get an overlap at inner corners of spline
triangle
toothdia = arcx/1.60;// no bigger then needed
// adjust this for outer teeth fit in internal spline
toothrad=53.5 / 2.000;// make radius was 53.0, reduce backlash but dont
compress it either
// adjust this to thin or thicken flex valleys
linerdia=51.62; // thins web at bottom of spline, helps flex if larger,
was 52.2
height=12;
tranz=height / 2.000;
echo (177,teeth,arcx,toothrad,linerdia) ;
union()
{ // make splines first
for($fr=[ 0: 6 : 359 ])
{
rotate([0, 0, $fr])
translate([toothrad, 0, tranz]) // mv x=[$lr + $r1] y,z=0
cylinder(d=toothdia, h=height, $fn=3, center=true);
// made triangles, now smooth inner wall by making 2 rings, outer
same as base of teeth
difference()
{
translate([0,0,tranz])cylinder(d=linerdia+.5,
h=height+.01, $fn=60,center=true);
translate([0,0,tranz])cylinder(d=linerdia, h=height+.02,
$fn=360,center=true);
} //but the innards needs smooth finish
}
}
}
The above reflects some comments I added as it was stuck rendering with
the blue line showing 999/1000 for around 10 minutes. A rerender after
makeing line one of the difference at an $fn=60 was still 6 minutes plus.
So the difference is being executed 60 times and getting the same results
everytime.
So I closed the "for" before the difference, getting exactly the same
render
in .8 seconds. Maybe this old fart will learn how to do efficient code yet.
;o)
But cura is still miss-rendering it, doing a retraction per tooth, 60
teeth,
times layer height=6000 times. No wonder its 4 hours per print. Its
spending
half its time doing useless retracts and restores. hell on the filament and
inviting a freezeup.
Thanks all.
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
On Sunday 04 July 2021 04:30:13 nop head wrote:
If the teeth aren't connected by a section big enough for infill then
it will have separate infill in each tooth which would require
retraction for each tooth. That should not be a problem.
But it is, its doing a worthless build with yards of string hangin off the part and jammed up the hot end only about 2mm up in a 12mm build. asking the teeth big enough to share/connect the infill is not an option because then its too thick and rigid to survive more than a minute or so.
And thats about a 4 hour teardown to clear that as there will be solid plastic in the heat sink that will have to be burned out of the heat sink and heat break tube. That gets old, very quickly as I just did that yesterday. And the only way to fix it is get rid of the bowden tube so you aren't retracting far enough to pull hot plastic back up into the heat sink. Screw it, a biqu bx will be here Wednesday, has the h2 head, auto bed level, silent motherboard, yadda yadda. Not as big a bed as this CR10-S Pro v2, but still like the ender 3, big enough for most projects. And likely only 1 z screw. But with the H2 head, no bowden tube, so a 2mm retraction is a good starter. At some point, I may see if the direct drive head from another cheap, Chinese disappearing maker printer can be transplated to the CR-10S pro v2. This CR10-S has done a good job, but has very high maintenance needs. The cheep chinky printer, only printed half a very small job, badly when the mainboard died. I also put 2 grand in dremels best, but it was the highest trouble maker I've had to date and less than half its features actually worked.
I've ranted enough for one day. Thanks for reading. I put the flag out since its the 4th, but I feel a nap is the next thing on my agenda.
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
I have never used a bowden tube myself. All I can say is with direct drive
it can retract as many times as it wants without any jams or stringing. It
does add a bit of time but with 1mm retracts done as fast as possible it
isn't much.
Seems like you are retracting too far if the retract causes a jam.
With some slicers you can specify to not retract when moving small
distances. That might give a little stringing but the idea with bowden
machines is the head can move so fast it doesn't need a retract. That only
works for moving head machines, which I think are the only ones where
bowden makes sense. With a bed slinger the bed weighs more than a direct
drive extruder so not much point in using bowden to make the head lighter.
On Sun, 4 Jul 2021 at 12:04, Gene Heskett gheskett@shentel.net wrote:
On Sunday 04 July 2021 04:30:13 nop head wrote:
If the teeth aren't connected by a section big enough for infill then
it will have separate infill in each tooth which would require
retraction for each tooth. That should not be a problem.
But it is, its doing a worthless build with yards of string hangin off the
part and jammed up the hot end only about 2mm up in a 12mm build. asking
the teeth big enough to share/connect the infill is not an option because
then its too thick and rigid to survive more than a minute or so.
And thats about a 4 hour teardown to clear that as there will be solid
plastic in the heat sink that will have to be burned out of the heat sink
and heat break tube. That gets old, very quickly as I just did that
yesterday. And the only way to fix it is get rid of the bowden tube so you
aren't retracting far enough to pull hot plastic back up into the heat
sink. Screw it, a biqu bx will be here Wednesday, has the h2 head, auto bed
level, silent motherboard, yadda yadda. Not as big a bed as this CR10-S Pro
v2, but still like the ender 3, big enough for most projects. And likely
only 1 z screw. But with the H2 head, no bowden tube, so a 2mm retraction
is a good starter. At some point, I may see if the direct drive head from
another cheap, Chinese disappearing maker printer can be transplated to the
CR-10S pro v2. This CR10-S has done a good job, but has very high
maintenance needs. The cheep chinky printer, only printed half a very small
job, badly when the mainboard died. I also put 2 grand in dremels best, but
it was the highest trouble maker I've had to date and less than half its
features actually worked.
I've ranted enough for one day. Thanks for reading. I put the flag out
since its the 4th, but I feel a nap is the next thing on my agenda.
Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
I believe that if the extruder has a PTFE tube and that tube can move
vertically, retractions can pull the tube up, which can fill the space
below the tube with melted plastic, which can cause an extruder jam.
Eliminating retractions can also help in this case, although switching
filaments by a huge retraction could cause the problem even without
retractions during printing.
On 7/4/2021 8:14 AM, nop head wrote:
I have never used a bowden tube myself. All I can say is with direct
drive it can retract as many times as it wants without any jams or
stringing. It does add a bit of time but with 1mm retracts done as
fast as possible it isn't much.
Seems like you are retracting too far if the retract causes a jam.
With some slicers you can specify to not retract when moving small
distances. That might give a little stringing but the idea with bowden
machines is the head can move so fast it doesn't need a retract. That
only works for moving head machines, which I think are the only ones
where bowden makes sense. With a bed slinger the bed weighs more than
a direct drive extruder so not much point in using bowden to make the
head lighter.
unsubscribe send an email to discuss-leave@lists.openscad.org
On Sunday 04 July 2021 08:14:57 nop head wrote:
I have never used a bowden tube myself. All I can say is with direct
drive it can retract as many times as it wants without any jams or
stringing. It does add a bit of time but with 1mm retracts done as
fast as possible it isn't much.
Seems like you are retracting too far if the retract causes a jam.
With some slicers you can specify to not retract when moving small
distances. That might give a little stringing but the idea with bowden
machines is the head can move so fast it doesn't need a retract. That
only works for moving head machines, which I think are the only ones
where bowden makes sense. With a bed slinger the bed weighs more than
a direct drive extruder so not much point in using bowden to make the
head lighter.
With a big bed such as the cr10 has, that that thought sure has come up
as a question.
In re the problem nop-head explained as causing the retractions. Which
cura set at 5mm as default but that leaves stringing during travels, so
I reset it to 6mm. A method that specs the max ID that would result in a
good fit on the armatures bearings, then adjust the tooth size to get
sufficient overlap at the base of the triangle to do away with the
retractions while leaving the ID fixed. And with 60 flats. That would
leave this part at the minimum web thickness which might be enough for
decent life. IDK until I've tried it.
I do have room to expand the outer rings yet. Then using that tooth size,
make new outer rings with the tooth point faceing inwards. I also have
the code written but commented out that puts locking keys in those parts
which might slip if not enough friction. But I am concerned with this
part, that the keys will serve as stress concentrators so they've
remained as // comments so far. So I think I'll fire up the husky and do
something about the back yards jungle. 'bout a week overdue.
Thanks muchly.
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
On Sunday 04 July 2021 04:30:13 nop head wrote:
If the teeth aren't connected by a section big enough for infill then
it will have separate infill in each tooth which would require
retraction for each tooth. That should not be a problem.
It was a huge problem, 2 days in a row. So I enlarged the teeth so there
was a bigger x section at the base of the tooth where they overlapped
with the next one, until the retractions disappeared. And I just now
burned out the jam, which went all the way back to the extruder,
breaking the filament where it entered the lever on the extruder, put
all new capricorn in and saved an hour+ in render time. Then it
autozeroed about 50 thou off the plate, redid the homeing for the third
time, got it too close so raised it .1. Now its printing, without the
retractions, but I've not a clue if its going to be usable sizewise, or
if with that thick a web, it will be flexible enough to survive an hour
running without breaking it.
That was after all, the general idea, to make it thin enough that it
could flex without near instant breakage.
I suspect that if I get the inner fit correct, that I'll have to make new
outer splines, and possibly enable a dozen keys around the edges of
those to prevent slippage which would screw up LinuxCNC real bad.
Thank you for the clue that explained the sudden appearance of all those
retractions. I hope I can work around it using PETG for those 3 parts.
With retraction reset to 4mm, was 6mm, stringing doesn't appear to be
hugely worse. PETG strings 50x worse than PLA, with strings so fine they
make a spider jealous. You can feel them in the hair on your arm, but
can't see them.
[elide code no longer valid]
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
Not sure if you know, but hair like strings can be removed easily with a
flame.
The only PET I have printed is E3D Edge. I didn't get strings with my
default 1mm retraction but it stuck to the nozzle, so couldn't bridge, and
was very weak. I have since learnt I should have dried it.
Dry filament works much better because when the extruder retracts the flow
stops completely. With very dry filament I get zero ooze from the nozzle at
the end of a build. With filament that hasn't been dried I get around 20mm
that I have to snap off before starting the next build. I use that length
to indicate how much moisture there is in the filament.
The other thing that can cause strings is excess temperature that makes the
filament too liquid.
On Mon, 5 Jul 2021 at 00:44, Gene Heskett gheskett@shentel.net wrote:
On Sunday 04 July 2021 04:30:13 nop head wrote:
If the teeth aren't connected by a section big enough for infill then
it will have separate infill in each tooth which would require
retraction for each tooth. That should not be a problem.
It was a huge problem, 2 days in a row. So I enlarged the teeth so there
was a bigger x section at the base of the tooth where they overlapped
with the next one, until the retractions disappeared. And I just now
burned out the jam, which went all the way back to the extruder,
breaking the filament where it entered the lever on the extruder, put
all new capricorn in and saved an hour+ in render time. Then it
autozeroed about 50 thou off the plate, redid the homeing for the third
time, got it too close so raised it .1. Now its printing, without the
retractions, but I've not a clue if its going to be usable sizewise, or
if with that thick a web, it will be flexible enough to survive an hour
running without breaking it.
That was after all, the general idea, to make it thin enough that it
could flex without near instant breakage.
I suspect that if I get the inner fit correct, that I'll have to make new
outer splines, and possibly enable a dozen keys around the edges of
those to prevent slippage which would screw up LinuxCNC real bad.
Thank you for the clue that explained the sudden appearance of all those
retractions. I hope I can work around it using PETG for those 3 parts.
With retraction reset to 4mm, was 6mm, stringing doesn't appear to be
hugely worse. PETG strings 50x worse than PLA, with strings so fine they
make a spider jealous. You can feel them in the hair on your arm, but
can't see them.
[elide code no longer valid]
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
On Monday 05 July 2021 04:17:45 nop head wrote:
Not sure if you know, but hair like strings can be removed easily with
a flame.
I've been shaveing it with a fresh box knife blade.
The only PET I have printed is E3D Edge. I didn't get strings with my
default 1mm retraction but it stuck to the nozzle, so couldn't bridge,
and was very weak. I have since learnt I should have dried it.
Dry filament works much better because when the extruder retracts the
flow stops completely. With very dry filament I get zero ooze from the
nozzle at the end of a build. With filament that hasn't been dried I
get around 20mm that I have to snap off before starting the next
build. I use that length to indicate how much moisture there is in the
filament.
Obviously from that as an indicator I need to rig a dryer box of some
sort to feed filament from.
Been looking for such a food container at wally's but havn't seen a
suitable candidate yet. This PETG has been out of the bag for about 90
days worth of sweaty humidity.
The other thing that can cause strings is excess temperature that
makes the filament too liquid.
Running at 235C and 50C now, I should lower that. Recently sanded
surface, bed adhesion is once again excellent.
Thanks for the newbie info.
Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page http://geneslinuxbox.net:6309/gene
From: nop head [mailto:nop.head@gmail.com]
Sent: Mon, 5 Jul 2021 18:18
To: OpenSCAD general discussion
Subject: [OpenSCAD] Re: loooooooong f6, like 10+ minutes but works finally
Not sure if you know, but hair like strings can be removed easily with a flame.
I thought you were going to show a tiny flame-thrower next to your nozzle.
--
This email has been checked for viruses by AVG.
https://www.avg.com