In search of accuracy
Greetings all;
Serial #1 of my drive finally died, but I've not knocked it apart to do
an autopsy yet. But it ran at about 1 rpm meaning about 120 rpm at the
motor for over 2 weeks, so is promising that an even longer life might
be possible.
But I am having a huge accuracy problem. the rotor, printed in white
PETG, was told to make an 8.28mm hole for the motor armature part, and
it made a 7.98mm hole, which I blamed on over extrusion since the nozzle
is .4mm, and with the correct flow setting, one would expect a hole half
the nozzle size smaller, or 8.08mm. 7.98mm can be driven onto the motor
shaft for a nice tight friction fit, which is sufficient to, with a
larger dflat, do away with any need for an alu hub to be inserted for
strength.
But I ran out of white PETG, so I made #3 out of a translucent green
using the same gcode. I just measured #2 at the desired 7.98 mm, so 2
white ones in a row are close enough for the girls I go with to
identical.
But the one made from green PETG, measures 7.66 mm and theres no way that
is going to drive onto the shaft for a tight friction fit. So I have one
started on the printer with an 8.48mm hole.
Is this sort of behavior normal just from changing the color of filament,
same brand but different color? Or is this from changing the appimage
from a 012021 to 072021, which is working much faster and smoother with
its expanded buffers since the machine has 16 gigs of dram.
Thanks everybody.
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
Filament colour doesn't change the size of holes but you have to measure
the filament diameter each time you change a reel and change your slicer
settings to match.
If you change filament types from harder to softer or change the pinch
roller pressure you also have to calibrate the E steps per mm as it changes
if the teeth sink in further.
Plastic contracts as it cools so you need to scale XY by that amount in the
slicer. In my case about 0.5%.
Then if you want to print round holes the correct size use single outlines
and use polyholes: https://hydraraptor.blogspot.com/2011/02/polyholes.html,
easy when using OpenSCAD.
On Thu, 12 Aug 2021 at 09:33, Gene Heskett gheskett@shentel.net wrote:
Greetings all;
Serial #1 of my drive finally died, but I've not knocked it apart to do
an autopsy yet. But it ran at about 1 rpm meaning about 120 rpm at the
motor for over 2 weeks, so is promising that an even longer life might
be possible.
But I am having a huge accuracy problem. the rotor, printed in white
PETG, was told to make an 8.28mm hole for the motor armature part, and
it made a 7.98mm hole, which I blamed on over extrusion since the nozzle
is .4mm, and with the correct flow setting, one would expect a hole half
the nozzle size smaller, or 8.08mm. 7.98mm can be driven onto the motor
shaft for a nice tight friction fit, which is sufficient to, with a
larger dflat, do away with any need for an alu hub to be inserted for
strength.
But I ran out of white PETG, so I made #3 out of a translucent green
using the same gcode. I just measured #2 at the desired 7.98 mm, so 2
white ones in a row are close enough for the girls I go with to
identical.
But the one made from green PETG, measures 7.66 mm and theres no way that
is going to drive onto the shaft for a tight friction fit. So I have one
started on the printer with an 8.48mm hole.
Is this sort of behavior normal just from changing the color of filament,
same brand but different color? Or is this from changing the appimage
from a 012021 to 072021, which is working much faster and smoother with
its expanded buffers since the machine has 16 gigs of dram.
Thanks everybody.
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 Thursday 12 August 2021 05:20:15 nop head wrote:
Filament colour doesn't change the size of holes but you have to
measure the filament diameter each time you change a reel and change
your slicer settings to match.
That seems fairly consistent, a roll of red says 1.75, and this green
says 1.74. I get the impression my errors are much greater than that.
I've tried flows down to 60% in cura, but all that got me was a frozen
hot end on one of the splined parts with thousands of retractions.
If you change filament types from harder to softer or change the pinch
roller pressure you also have to calibrate the E steps per mm as it
changes if the teeth sink in further.
Makes sense for final picky bits. I think my flow is way way off, heavy.
Some of my nozzle primers are almost 2mm wide. I can't get down to an
initial layer of .2 as I believe nozzle pressure is supporting the head
when I get it that close. Bed adhesion rises quickly in that range.
Plastic contracts as it cools so you need to scale XY by that amount
in the slicer. In my case about 0.5%.
I'm still trying, not very successfully, to get to where those effects re
anything but noise.
Then if you want to print round holes the correct size use single
outlines and use polyholes:
https://hydraraptor.blogspot.com/2011/02/polyholes.html, easy when
using OpenSCAD.
Interesting read, thanks for that. From that I get the impression I'm
wasting my time doing small circles with an $fr=360 which i am sure
impacts my render and print time. And its plain this kit built prusa is
over extruding but how far can this be cut back w/o impinging the
interline weld? IDK. My experiments have frozen the hot end is all I
know for sure, which has led me to up the speeds.
And if the part currently on the printer isn't useable, I will start over
by subtracting half the nozzle width from the target measurements so
outers only drive the head to target-.2, and inners to target+.2, and
something like an 8mm shaft hole only needs an $fr of 16 or 24. And
adjust that .2 to match the designed size, to rendered at the correct
size. That won't materially effect my big d-flat, which I'm grinding on
the motor shaft with a CBN wheel.
All of which will change if the flow is wrong. Ideally, do the .2mm comp,
and adjust flow till the hole is the correct size would seem to stiffen
this jelly up to where it might be nailed to a tree? :o)
What is the t-verse # of those test files shown? They might help me fine
tune the flow with an m221 S## (in %) in the preamble cura feeds the
printer.
Thanks nophead.
[...]
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
Gene, I think the problem you're dealing with is intrinsically part of FDM, and possibly other 3D printing methods as well. At root, the technology is using an icing bag and attempting to achieve precision. There are so many production-time variables that can affect the result that it is impossible to solve the problem at design time.
I created test blocks in OpenSCAD for different fasteners. Through holes, self-tapping holes, tapped holes, in several orientations. Before beginning the design of an object, I run the test blocks using exactly the same production parameters (temps, feeds, materials, and yes, even color) to discover what values I need to specify at design time to get the dimensions I need.
On Thu, 2021-08-12 at 10:45 -0400, Gene Heskett wrote:
Then if you want to print round holes the correct size use single
outlines and use polyholes:
https://hydraraptor.blogspot.com/2011/02/polyholes.html, easy when
using OpenSCAD.
Interesting read, thanks for that. From that I get the impression I'm
wasting my time doing small circles with an $fr=360 which i am sure
impacts my render and print time. And its plain this kit built prusa
is over extruding but how far can this be cut back w/o impinging the
interline weld? IDK. My experiments have frozen the hot end is all I
know for sure, which has led me to up the speeds.
That IS an interesting read.
All of which will change if the flow is wrong. Ideally, do the .2mm
comp, and adjust flow till the hole is the correct size would seem to
stiffen this jelly up to where it might be nailed to a tree? :o)
Personally, I would adjust flow rate to where it comes out right on
simpler objects, and fine tune that by measuring and specifying
filament diameter and small adjustments of flow rate, keeping it within
about +/- 0.1 of a flow rate of 1.
Then, as for fit, I do small tests of critical holes. Four or five
holes in a small cube around the desired size will give you an idea of
what sort of fudge factor to apply to them. And of course, nophead's
polyholes will definitely be useful in that test,
What is the t-verse # of those test files shown? They might help me
fine tune the flow with an m221 S## (in %) in the preamble cura feeds
the printer.
My experience with the Prusa MK3S is that it works flawlessly with
PrusaSlicer. If you are not getting good results with this combo, then
something is physically wrong with the printer. Once fixed, it should
work well. I would adjust flow rates until you get good, solid prints,
with excellent layer adhesion. Then I would print prototypes to nudge
your design to get the print that you need. My experience is that
simply adding 0.1 or 0.2 mm to an inner diameter is enough to ensure
proper fit. I have printed threaded rods on my Prusa which work with no
problems. Not quite the tolerances that you are seeking, but I know you
can get there from here.
Jon
On 12 Aug 2021, a little later, Ned Young posted:
I apologize for the condition of the script--it isn't ready for the public. It lacks a great deal of explanation and guidance necessary that others may use it without an excessive amount of effort.
There are sets of parameter values for each fastener. All sets except the one in use during a particular run are commented out. Those parameter values are tweaked in use to discover the settings needed for a particular run.
The notation of the form "// PETG 0.3mm" does not mean those values have been verified and you can go straight to design for that material. It means only that the last time I tested that fastener, those values worked for the particular circumstances at that time.
