<peepsalot>
not sure what to do about precision differences on Mac vs others. I got different Ubuntu versions and Windows to all play nice, but don't even have a test report for mac.
fling has quit [Remote host closed the connection]
fling has joined #openscad
ur5us has quit [Ping timeout: 250 seconds]
<JordanBrown[m]>
J1A84: I think the best way to think about it is that $fa controls the maximum number of sides for large circles, and $fs controls the maximum number of sides for small circles. The dividing line between "large" and "small" depends on the values.
<JordanBrown[m]>
For instance, for $fa=18 and $fs=0.65, the dividing line is about r=2.
<JordanBrown[m]>
r>2 will have 20 sides, no matter how big they are.
<JordanBrown[m]>
r<2 will have sides less than 0.65 units long, no matter how few there are.
<JordanBrown[m]>
A non-silly combination, for instance, might be $fa=5 and $fs=0.5. Circles larger than about 5.7 will be 72-ogons, which are pretty smooth, while smaller than that will have sides of at least 0.5. because that's getting down into the smallest feature you can get on a 3d printer.
ur5us has joined #openscad
aiyion has quit [Ping timeout: 258 seconds]
aiyion has joined #openscad
fling has quit [Remote host closed the connection]
fling has joined #openscad
fling has quit [Read error: Connection reset by peer]
fling has joined #openscad
linext_ has joined #openscad
linext has quit [Ping timeout: 252 seconds]
ur5us has quit [Ping timeout: 248 seconds]
<J1A8487>
JordanBrown[m] The thing is that $fa makes sense to limit resources but for printing it doesn't help because if you print a 200mm cylinder with 72-ogons it looks terrible, while this is fine for a 10mm diameter. So for printing the unlimited $fs is the only important.
J1A8487 is now known as J1A84
<J1A84>
a $fs of 2 will be clearly visible when printing big circles
<peepsalot>
i often set $fa=0.1, effectively disabling it in favor of $fs
<JordanBrown[m]>
I think I misread the original comment.
<JordanBrown[m]>
Now I'm going to have to try printing a 200mm circle...
<J1A84>
print a ring Ü
<JordanBrown[m]>
Yeah, doing that. Well, actually, a quarter of a ring, since it will take 1/4 as long and 1/4 as much plastic.
<JordanBrown[m]>
At 5°, 2°, 1°, and 0.5°
<JordanBrown[m]>
though maybe I should have done the experiment in terms of $fs instead.
ur5us has joined #openscad
<JordanBrown[m]>
way past bedtime. g'nite. or good morning, as the case might be :-)
KimK has joined #openscad
lastrodamo has joined #openscad
KimK has quit [Ping timeout: 250 seconds]
castaway has joined #openscad
jonasbits has quit [Ping timeout: 268 seconds]
teepee has quit [Ping timeout: 258 seconds]
ur5us_ has joined #openscad
teepee has joined #openscad
ur5us has quit [Ping timeout: 248 seconds]
jonasbits has joined #openscad
Killy has quit [Quit: You have been kicked for being idle]
ur5us_ has quit [Ping timeout: 250 seconds]
fling has quit [Ping timeout: 258 seconds]
fling has joined #openscad
KimK has joined #openscad
KimK_ has joined #openscad
KimK_ has quit [Remote host closed the connection]
KimK has quit [Remote host closed the connection]
KimK has joined #openscad
epony has quit [Ping timeout: 252 seconds]
ur5us_ has joined #openscad
epony has joined #openscad
ur5us_ has quit [Ping timeout: 260 seconds]
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
snaked has quit [Quit: Leaving]
TheHouse has joined #openscad
<TheHouse>
good day all, i am new to the program, channel not the network, or code/script. i was wondering if there might happen to be a simple scisssor lifttable that i could use to jumpstart my project from?
<J1A84>
hmm
<J1A84>
i have something like that .. or hinges in the lib
<TheHouse>
ill get it i like it already i can see a million different uses/project this will help me out with
<TheHouse>
ahhhh yes
<J1A84>
you also can make animations for the movement
<TheHouse>
this is a start lol
<TheHouse>
easy now
<J1A84>
for the start you will use rotate translate difference cylinder cube
<TheHouse>
let me get a couple pieces together first
<J1A84>
so as a start make familiar with these 5
<TheHouse>
listening and taking notes believe it or not
* TheHouse
takes suggestion well from those there before him
<TheHouse>
forced into retirement and kids have lomg left home now so one thimg i do have is time
<TheHouse>
i can forsee the need for real kb too
<InPhase>
A proper keyboard will certainly facilitate OpenSCAD usage.
<InPhase>
Also, welcome. :)
<TheHouse>
yes for sure and thank you all a million times over was rather surprised to see irc still being used (more as i believe it was intended)
<TheHouse>
irc is the reason i dont have any other forms of social media long lomg long time now old irc user
<InPhase>
Yeah. This network in particular seems to remain a cluster of people who lean smarter and more polite than most of modernized social media. It is a great asset to have it.
<J1A84>
sometimes i wonder if polite and smart is going hand in hand - also that smart is sort of interpretation if the other said something offensive
<JordanBrown[m]>
and I never bothered to fine tune the dimensions and angles to make it be a truly accurate simulation.
<JordanBrown[m]>
So my large-circle test finished. It has a diameter of 180mm. (To fit on my 200x250 bed... I could have gone bigger once I cut it down to a quarter.) Indeed, the 5° test strip, which would be about 8mm sides is clearly a polygon. The 2° test strip, about 3mm sides, I can't convince myself that I can see the sides, but maybe I can feel them. The 1° and 0.5° strips, 1.6mm and 0.8mm, no, can't see or feel them them.
<JordanBrown[m]>
So J1A84 thanks for prompting me to do that test; I thought it would do better than that.
califax has quit [Ping timeout: 258 seconds]
<J1A84>
you re welcome .. i had the experience my self where i wondered why i had these marks in a vase print .. and turns out it was from the render
califax has joined #openscad
<J1A84>
if you print with a small nozzle these become even more prominent
<J1A84>
and you have this also on small prints if you have a big radius fillet or curve
<JordanBrown[m]>
I do wonder whether any of the slicers try to detect curves and round them out.
<JordanBrown[m]>
Ah, well, off to paying work.
<Scopeuk>
there is the "arc welder" plugin a lot of people use although that is mostly targeting using the g code commands that prescribe arcs esentially as compression
<J1A84>
nah they don't use G2 or G3 commands and have a resolution .. would be hard to tell the difference if someone want a smooth surface or is printing a diamond
<Scopeuk>
arc welder is a pattern match to turn a series of short linear moves (heuristically id's as a curve) and swapping it to G2/G3. never used it personally though
califax has quit [Remote host closed the connection]
* teepee
still things G2/G3 are useless for that kind of stuff
<J1A84>
the g codes were invented when you are not able to feed several MB code into a CNC .. and you don't want to switch floppy disk while machining .. like installing a software .. please insert disk 8/35 to continue
<J1A84>
when i check my gcodes are often over 20Mb
<Scopeuk>
they proabbly also make a lot of sense if you have rotation controls to be able to actually drive an arc
<J1A84>
when you use fuzzy surface they go through the roof
califax has joined #openscad
<J1A84>
the arc work on a 3D printer .. it just uses the max resolution like every cnc
<Scopeuk>
I've never felt the need to use it either to be honest
<teepee>
oh, they are probably quite good for "normal" cnc for drilling routing holes / circles
<teepee>
the more general stuff is G4/G5 I think
<Scopeuk>
g4 is dwell, g5 is bezier cubic spline so thats definately more general
<teepee>
ah, g5 then
<J1A84>
works if your export file is not triangle mesh based
<J1A84>
FreeCad is also converting a scad circle into an arc
teepee has quit [Remote host closed the connection]
<JordanBrown[m]>
But it wouldn't be silly for a slicer to have a (switchable) heuristic for smoothing curves.
teepee has joined #openscad
Guest8183 has joined #openscad
Guest8183 has quit [Client Quit]
<J1A84>
it would help for speed as high resolutions already slow down printer as they can't keep up with the amount of data (and their small buffer)
<J1A84>
was a big issue when a cura update switched high resolution on and all "non current ultimaker models" started to have bad prints as they had to stop
<J1A84>
while the dev team said " works fine on our newest printer model"
<Friithian>
I've had no problems with the S5 I use :D
<J1A84>
i installed the update after they fixed it - Ü
<J1A84>
iirc it was with some outdated parameter from old profiles .. so creating a new profile also helped
<Friithian>
I usually wait a bit before updating cura, although I do like using their beta for beta features
<J1A84>
but i assume it would cost a lot of slicer time to calculate the best bezier path through all mesh points for every layer. (and not sure if all printer can use G5)
<Friithian>
tree style supports already take a long time to calculate, and those curves would be a lot more than trees
<J1A84>
the arc commands would only work on very old school tech cnc designs .. but not what 3dPrinters are good for like free 3D surfaces
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
<ali1234>
what version of cura was that?
<ali1234>
i started having weird print problems at some point around 4.9 i think
<J1A84>
not sure but was caused by too low values in "maximum resolution" "maximum deviation" or "travel resolution"
fling has quit [Remote host closed the connection]
fling has joined #openscad
fling has quit [Remote host closed the connection]
<J1A84>
yeah that worked too .. i just cut that side open to check but still the opening was too small to press with the finger - so i added a "button" for the counterpart - now need to be printed upside down but that also allow equal thickness of the loop
<J1A84>
and ensures you will not try to put it wrongly together
epony has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 258 seconds]
teepee_ is now known as teepee
fling has quit [Remote host closed the connection]
fling has joined #openscad
teepee has quit [Remote host closed the connection]
<peepsalot>
do we just need to add the `if: ${{ always() }}` line? I'm assuming that means a failing return code of previous step will be ignored for executing that step?
<peepsalot>
teepee: i'm guessing if anyone knows, its you?
<Scopeuk>
I did make an update to those work flows recently to fix the warnings about soon to be dropped runner versions
<Scopeuk>
shouldn't have had a functional effect
califax has quit [Remote host closed the connection]
<peepsalot>
will the workflow update if I edit the workflow on my PR branch? or does it go off of the master HEAD instead for security etc?
califax has joined #openscad
<peepsalot>
Scopeuk: yeah I actually just noticed that the only contributors to the mac workflow so far are you and kintel
<peepsalot>
ah, and I didn't have those updates in my branch either, just merged them in.
<peepsalot>
hopefully I've got it solved, but feel free to chime in if you have any extra info :)
<teepee>
ah, github actions, yes that always() thingy makes it run even on failure of steps before
<teepee>
on the action or the decimal topic? work is a bit busy currently so I'm not catching all the discussions in detail
<peepsalot>
i just meant on the action for now, but i would appreciate some feedback on the decimal stuff eventually too
<peepsalot>
if the mac test report upload works now, then hopefully I can make some more progress determining where the differences are
<teepee>
on the action, lets see, I think that should be enough to get the upload done
<teepee>
the decimal stuff is mainly for fixing the cache issues that was discussed some time ago? that would be a very nice fix
<peepsalot>
yes, but it would also apply to echo's and other places
<peepsalot>
of displaying all digits necessary to uniquely identify a value (round trip capable).
<teepee>
I guess that might upset some people in some cases but others asking for more precision will be happy :)
<peepsalot>
yeah i suspect at the least it will increase demand for formatted number output, which JordanBrown[m] brought up the other night
<teepee>
linext_: nothing wrong with asking Irv, he very generously donated to openscad, so maybe he's interested
<teepee>
worst case the reply is "no thanks, I have someone for that already" ;-)
<linext_>
i feel like if i start selling this project, i'll end up stopping at a point where the job is completed for them
<linext_>
printables, irv shapiro, cults3d, etc.
<teepee>
yeah, format to string via fmt library is probably a neat thing, ideally limited to what will be in c++20, but then that's maybe not too critical
<linext_>
they'll probably want to pay a fixed amount for a fixed number of features, not an ongoing build that's better than what they offer
<linext_>
my goal is to be better than thingiverse and it's competitors
<teepee>
that's quite possible
<linext_>
if irv want's support on integrating web assembly, i can help with that
<teepee>
I told him it's available but he did not say either way at the time
<linext_>
dominick schroer has been helping me along the way
<linext_>
oliver ochafik never responded to any of my emails
<teepee>
I guess the main focus was on having a working customizer for thingiverse
<linext_>
yea, i already build the "pull from thingiverse' functionality demo
<teepee>
there's quite a number of extensions done in the last couple of month, I have a mail listing those somewhere, did not find the time to make it an article