<Virindi>
wow nice. I tried guso78k's csg branch and freecad import gets a lot further with my test model. the result even looks like the model! still a boatload of errors but I am not sure if they are meaningful https://paste.ec/paste/6RkMaSHk#PAUSa+6224tnlM-k+9HHqRWiGQJ5jVCIBeYE3dsbJ0D
R2robot has quit [Quit: Pull the lever, Kronk. Wrong leverrrrrrr!]
<Virindi>
interestingly some arc faces got converted to an arc face, others became a pile of tiny faces
<Virindi>
in the middle of some arc faces there is randomly a break
<Virindi>
where nothing should be different at that point
R2robot has joined #openscad
<Virindi>
but it looks like it might be usable
<Virindi>
I'm impressed :)
<Virindi>
I know it is just a missing print param but that was a fast fix
teepee has quit [Ping timeout: 260 seconds]
teepee has joined #openscad
Church- has joined #openscad
<Church->
guys, can you advise on https://pastebin.com/Pya0QG2a why i need to increase lenght by constant called weird :) , for sake of connecting sides? even without it formulas seem fine, hate adding manual adjustments for parts "to fit"
Guest68 has joined #openscad
Guest68 has quit [Client Quit]
mmu_man has joined #openscad
<J24k47>
Church- ?
ccox has joined #openscad
<Church->
i just googled up how to make curved rectangle, and trying to use it to convex in hex sides by specified amount
<J24k47>
dh need to be the correct length then you don't need "weird"
<J24k47>
in general i would use a difference with circles and then extrude
<Church->
dh i want to be adjustable, to choose how much convex sides of hex i want
<J24k47>
you can but you need to calculate the connection point via the size and so determine the radius (or other way around)
ccox_ has quit [Ping timeout: 260 seconds]
<J24k47>
if you want a hex the circles need to translate so they meet at the edges
<J24k47>
depending on the radius the angle of how much of a circle will fit into a side of a certain length
<J24k47>
Church- if using a difference you simply can overlap the circles and get the concave radius you like still need to set the size of the form
<Church->
damn, forgetting all learned geometry in school three decades ago, bites hard, when trying to adapt random googled snippets :)
<J24k47>
and if you need a thin rim .. use the form and difference with itself and offset
<J24k47>
else you will need some sin() to calculate the proper distances
<J24k47>
check for calculating a chord as the side of the hex need to be the chord of your circle
<Church->
hmm, thanks. your way seems simpler in several ways
zauberfisch has quit []
zauberfisch has joined #openscad
Virindi has quit [Ping timeout: 255 seconds]
J24k47 has quit [Quit: Client closed]
J24k has joined #openscad
Virindi has joined #openscad
snakedGT has quit [Remote host closed the connection]
kintel has joined #openscad
mmu_man has quit [Ping timeout: 255 seconds]
pca006132 has quit [Remote host closed the connection]
pca006132 has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mmu_man has joined #openscad
mmu_man has quit [Ping timeout: 268 seconds]
<Church->
j24k47: thanks again for hints. What i ended up with: nutwidth=10;recess=0.3;rad=nutwidth/sqrt(3);rad2=rad^2/(8*recess)+recess/2;difference(){ circle(r=rad,$fn=6); for(a=[0:5]) rotate(30+a*360/6) translate([nutwidth/2-recess+rad2,0]) circle(r=rad2,$fs=.2,$fa=1); }
<Church->
now if i adjust nut width/recess depth parametrically, works as i wished
<Church->
main thing i had to google for was calculation of circle radius based on chord length and arc height
<J24k>
great!
<J24k>
if you only need it for hex you are lucky as the radius is equal to the side length.
<J24k>
I put functions for these in my lib to use
mmu_man has joined #openscad
<InPhase>
Church-: Is this for a torx inset?
GNUmoon2 has quit [Remote host closed the connection]
TheAssassin has quit [Remote host closed the connection]
fling has quit [Remote host closed the connection]
califax has quit [Write error: Connection reset by peer]
<InPhase>
Church-: If so you probably want offset(roundby) offset(-roundby) in front of that difference. And also a bit more curve in than you have seems to be standard.
<InPhase>
teepee: Is there a right click menu on any of those fonts?
<teepee>
but good point with a sensible sample, I'll add that to the issue todo list
<teepee>
no right click right now. what would that do?
<InPhase>
teepee: I was just thinking that this really enhances the utility of fonts. But one core issue is people having different fonts on their systems. Which can be alleviated if people are sharing fonts. But if you're using system fonts, that means you're going to need to copy to the local directory. Which means you'd need to know where the font was loaded from on your system.
<InPhase>
Of which the first thing that popped to mind is "copy full pathname".
<teepee>
there's actually a full path column too, but it's hidden
<teepee>
right, and open path
<InPhase>
I figured displaying it would be too noisy, yes.
<teepee>
like on the tabs of the editor
<InPhase>
Open containing folder would be great too.
<InPhase>
I don't need this personally, as I have "locate" installed like a good Linux user. ;) But for the workflow of the majority.
<InPhase>
But basically, some workflow to get the equivalence of how pdf embedding fonts solved font issues that postscript had.
<InPhase>
But I think here the only sensible embedding is probably an accompanying file.
<InPhase>
Heh. Apparently without any effort to ever acquire this, I have 31 copies of DejaVuSansMono.ttf on my laptop, which according to md5sums of them all consists of 6 different versions of this font, with the most popular version having 12 copies.
<InPhase>
Ah. The 12 are from the 12 copies of matplotlib I have. Thanks, Python environments.
<InPhase>
I was doing that to check the filesize, which is 340k, which would be 450k if base64'd for true "embedding", or 276k even if compressed first. So this is probably a bad idea, and accompanying files makes maximal sense.
fling has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
fling has joined #openscad
fling has quit [Read error: Connection reset by peer]
fling has joined #openscad
<Church->
InPhase: nope, not for torx. Just that i saw advise regarding embedding hex-nuts in recesses of paused 3d prints, that just like with flank drive wrenches, rounding sides can reduce chance of stripping, as torque happens in this case not at narrow sharp hex nut edge, but over wide side area
<Church->
so = less chance to strip / hold more torque
<Church->
prusa employs something similar in their models for hex nut holes (except in a bit different form, by cylinder delete of corners): https://youtu.be/9-6tIkTrcwA?t=48
<othx>
Church- linked to YouTube video "5 3D Printing Tips the ELITES don't want you to know... #3DP101" => 1 IRC mentions
<Church->
Obviously square nuts would provide more torque, and real melt-in threaded inserts much more, but wanted to model with hex nut as insert as that what most will have around
<Church->
but as i wanted to play with various curvature offset and other clearances in fitment test prints + wanted to make it parametric for different nut sizes, hence the wish to get it right instead of manual translating bits in place once
<Church->
Slightly interested, if thingiverse's customizer does unicode. In one openscad model made prompts of parameters looking better using some unicode icons/glyphs .. but high chance for them to not display properly on thingiverse's
<teepee>
would using "\U01F600" help?
<Church->
i've yet to post in thingiverse. just that stumbling here and there mentions of possible differences or lack of features in it's customizer with fully featured one in openscad, i wouldn't be surprized with limitations regarding that aswell
<teepee>
also keep in mind thingiverse is on an ancient version
<teepee>
even though openscad releases are *very* slow, they are behind 2 releases at 2015.03 IIRC
<Church->
hmmm. damn, gonna need to look up changelogs then, to not get issues with model code itself :/
<Church->
any biggies to keep in mind of 2015 vs current?
<teepee>
lots of new stuff obviously but if you don't need that, it should still work ok
<teepee>
not sure if makerworld has their customizer public yet
<Church->
if i have use case to need to hull around concave objects (eg. two bolts), should i dismember them and hull separately? or should i do instead of hull projection cut and extrude it and place between them? hull as is, hulls together bolt head with threaded body :/
<InPhase>
Church-: Oh, that. I've done hex nuts in prints before, but I always just left a slot in the side to slide them in, rather than bothering with pausing.
<InPhase>
I think I had an example of that on thingiverse...
<InPhase>
Nope, maybe I never put that up.
<Church->
i saw examples of insertion from side .. but it will be two sides holding nut in place less, +insertion hole on otherwise good looking smooth outer surface :)
<Church->
and frankly i was impressed with technology itself of FDM printing in objects :) .. seems very perspective one, for example to insert not just bolts, but also carbon fiber strips, to make printed part much stronger
<Church->
oh, and then there was use case for eg. reducing support waste & reducing print time, by inserting preprinted support just below of top bridged or overhang part. peaked my interest
erectus has quit [Quit: erectus]
erectus has joined #openscad
snaked has joined #openscad
erectus has quit [Remote host closed the connection]
erectus has joined #openscad
erectus has quit [Remote host closed the connection]
erectus has joined #openscad
<InPhase>
Church-: Yeah, it's a pretty nice bonus capability. :)
erectus has quit [Remote host closed the connection]
<teepee>
ah, no not quite, just the header has 2001 but that's probably just bogus
<teepee>
did gnome exist in 2001?
<teepee>
crazy, it did Initial release March 03, 1999
<teepee>
anyway, question now is obviously, as we are talking fonts, how does it work on other platforms
<Church->
btw, is there a way to more decouple UI of openscad from rendering engine? And at very least make each openscad window to have own rendering engine unrelated to those in other windows?
<teepee>
right now: run multiple instances
<Church->
if i sometimes mess up with parameters and it runs into ultraheavy long (that probably won't complete) rendering, during which i have to click few minutes on "stop rendering" button
<Church->
and during that time other windows are also not usable
<Church->
UI should be more responsive/prioritized, if nothing else, then for sure/quick way to stop rogue rendering :). And it shouldn't block others :(
<teepee>
I'm sure everyone agrees
<Church->
if it had been separate process, then i could kill that process w/o killing UI app (in which i might have forgotten to save several changes, so sometimes not wishing to force kill app .. and in those occasions again, few minutes clicking "stop" :)