teepee changed the topic of #openscad to: OpenSCAD - The Programmers Solid 3D CAD Modeller | This channel is logged! | https://openscad.org/advent-calendar-2021/ | Website: http://www.openscad.org/ | FAQ: https://goo.gl/pcT7y3 | Request features / report bugs: https://goo.gl/lj0JRI | Tutorial: https://bit.ly/37P6z0B | Books: https://bit.ly/3xlLcQq | FOSDEM 2020: https://bit.ly/35xZGy6 | Logs: https://bit.ly/32MfbH5
ur5us_ has quit [Ping timeout: 240 seconds]
ur5us_ has joined #openscad
snaked has joined #openscad
miique has quit [Ping timeout: 240 seconds]
miique has joined #openscad
epony has quit [Ping timeout: 240 seconds]
ur5us_ has quit [Ping timeout: 240 seconds]
Jack229137 has joined #openscad
Jack2291 has quit [Ping timeout: 256 seconds]
califax- has joined #openscad
califax has quit [Ping timeout: 276 seconds]
califax- is now known as califax
Guest60 has joined #openscad
Guest60 has quit [Client Quit]
lostapathy has quit [Read error: Connection reset by peer]
lostapathy has joined #openscad
lostapathy has quit [Client Quit]
lostapathy has joined #openscad
submariner has quit [Remote host closed the connection]
epony has joined #openscad
othx has quit [Ping timeout: 240 seconds]
othx has joined #openscad
Jack229137 is now known as Jack22
ToAruShiroiNeko has quit [Ping timeout: 240 seconds]
ToAruShiroiNeko has joined #openscad
ur5us_ has joined #openscad
lostapathy has quit [Ping timeout: 256 seconds]
lostapathy has joined #openscad
califax has quit [Ping timeout: 276 seconds]
califax has joined #openscad
arebil has joined #openscad
ur5us_ has quit [Ping timeout: 240 seconds]
arebil has quit [Ping timeout: 240 seconds]
arebil_ has joined #openscad
arebil has joined #openscad
arebil_ has quit [Ping timeout: 252 seconds]
lastrodamo has joined #openscad
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]
arebil has joined #openscad
Guest21 has joined #openscad
Guest21 has quit [Client Quit]
Virindi has joined #openscad
<Virindi> Just got a "new" laptop (used 2017 model) and of course openscad preview is like 1fps on all my main models, so it is unusable. The only machine I am able to actually use is the 5800x I purchased specifically to get the fastest single core to run openscad previews. There just HAS to be a better way :(
<Virindi> I see a lot of PRs for making render faster, but it doesn't even matter to me if render takes 10 minutes, if I get 1fps in the preview window I will never build that model in the first place
<teepee> there likely is
<Virindi> also I never use hull() or anything, and my $fn is always like 20...only unions and differences :\
<Virindi> sorry, it is just frustrating because I hate other cad types and love openscad
<teepee> did you try with the dev snapshots, there's some preview render improvments that can be enabled via preferences->features
<Virindi> I will try it. My first try was with the default packages in bullseye, which is like 2021.something
<teepee> if you can setup extra repos, you can use the nightly debian packages
<teepee> otherwise the AppImage should work for that test too
<Virindi> yeah np. I just have to reinstall the OS again because I hosed the first install messing with luks ;)
<Virindi> then I will try it
<tcurdt-> I am also a bit curious about this. Would a render have similar performance to what e.g. Blender is able to provide? So it's explicitly the way the preview is created?
<Virindi> for 10 years I have been buying hardware just to get faster openscad previews, heh
<teepee> tcurdt-: no
<Virindi> is it known what the main bottleneck is in preview speed?
<tcurdt-> teepee hm... isn't the final render result also just a mesh that is passed on to be displayed?
<teepee> the only comparison against blender is *showing* the final result of an F6 render (aka not counting the calculation time)
<tcurdt-> that's what I meant!
<teepee> Virindi: yes, pretty much. depending on operation, it's different reasons
<tcurdt-> and since the creation of the final result usually takes quite long there is the preview which "fakes" it - and that itself has it's limitations
<Virindi> obviously hull() or minkowski() is incredibly slow and I just avoid them because of it, I really mean for simple union() difference() intersection()
<tcurdt-> is my above summary correct like that?
<teepee> hull itself is actually very fast, it's the mostly invisible union happening inside that can make it appear very slow
<Virindi> I was thinking though...
<Virindi> when I am making a model, most of the time I am making an assembly. What is displayed on the screen is many different parts, and most of them have no change when I am working
<teepee> tcurdt-: yes
<Virindi> different parts of the top level union could maybe be rendered in the background and then displayed like a render (so, fast), and only "preview" things which haven't been rendered yet
<Virindi> 99.9% of models have a top level union which includes many independent parts
<teepee> preview has no automatic way to optimize that, it's possible to help in the design using render() which defines a point of actual mesh caluclation that can be cached and then reused for preview quickly
<teepee> why work on details with full preview?
<teepee> I tend to switch off the other stuff, so for example the "zoom to fit" works trivially for the part I want to focus on
<Virindi> if I am working on a part in an assembly, I want to have the full assembly turned on so I can see how features I am adding might interact with other parts
<Virindi> I have simplification rules for it, for example, I have a flag which turns aluminum extrusions into cubes rather than displaying the actual model
<Virindi> or a flag which disables gear teeth
<Virindi> it is enough on the 5800x but not on the 7600u
<teepee> right, that should help a lot
<Virindi> the one usecase I always come back to for having everything visible is the travel of moving elements. It is a continuous job to ensure you have not accidentally added a collision which reduces travel, so it really helps to just keep everything nearby visible to keep it in your mind
<Virindi> I have many designs where moving parts intricately 'slot together' along their travel, allowing the moving part(s) to pass just barely
<teepee> yeah, makes sense. I don't have any designs which need something like that
<Virindi> I know every 'sane' person uses fusion for this, but I am not sane, I am not using some proprietary crap and I have never been able to get comfortable with freecad (plus, openscad is just vastly superior at making parametric designs) :)
<tcurdt-> by that definition I am probably not sane either I guess :)
<teepee> and I disagree using fusion is the only sane option
<teepee> but now lets see if I can print PETG :)
<Virindi> fan to minimum :P
<teepee> ah, ok, that's easy, it's just an external 120mm fan sitting on a goose neck
<teepee> I can just point it somewhere else :D
<Virindi> with petg if you have too much fan, the layer bond is weak and the print looks opaque/cloudy rather than shiny. other than that and filament drying, it is easy :)
<teepee> cool, I want more transparent as it's the casing part for a led signal tower :)
<teepee> so shiny would be nice I guess
<Virindi> looks like a factory machine safety light :)
<Virindi> the lens would probably be a lot more clear if you made it one perimeter thick and used vase mode
<teepee> yes
<teepee> maybe if I ever manage to get a 0.8 or 1mm nozzle
<Virindi> but I guess some amount of diffusion is desired here
<teepee> yep, that too, it does not need to be fully transparent, even the PLA looks ok, but that specific filament is annoying
<teepee> so I'm trying the one from DasFilament which is nice so far but their transparent PLA is very yellow-ish
<teepee> hence the PETG which is nice "cold/blue-ish" transparent
<Virindi> yeah that makes sense
<teepee> also my signal tower is much better than those crazy expensive commercial ones, because RGB LED rings! 5V! WLAN! ;-)
<Virindi> IoT IoT IoT IoT Cloud IoT Cloud
<teepee> well, to be fair, with max power 5 levels would eat 3.6A
<teepee> but I've tuned it down so it maxes out just a tad over 500mA so it's ok for normal USB connectors
<teepee> one will probably end up as customer monitoring in my company
<Virindi> usb-c is good to 5A with a decent cable, you could always use that. Or just use a barrel jack :)
<teepee> true, but this way it would just trivially work on pretty much any usb port, e.g. of a desk monitor or so
<teepee> I guess an USB-C power delivery chip would allow for full power, but I don't think that's really required at this point
<Virindi> if you use PD, might as well use higher voltage
<teepee> I'm almost sure that would be the only option get the nearly 4A @ 5V internally
miique has quit [Read error: Connection reset by peer]
linext_ has quit [Ping timeout: 256 seconds]
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]
linext has joined #openscad
Zauberfisch has quit [Quit: No Ping reply in 180 seconds.]
Guest2 has joined #openscad
Zauberfisch has joined #openscad
Guest2 has quit [Client Quit]
oldlaptop has quit [Ping timeout: 245 seconds]
oldlaptop has joined #openscad
siege has quit [Ping timeout: 252 seconds]
siege has joined #openscad
othx has quit [Ping timeout: 256 seconds]
othx has joined #openscad
<gbruno> [github] t-paul pushed 3 additions 2 modifications (Merge pull request #4040 from cwalther/usetransform
zingos has quit [Ping timeout: 245 seconds]
zingos has joined #openscad
ndnihil has quit [Ping timeout: 252 seconds]
ndnihil has joined #openscad
ndnihil has quit [Changing host]
ndnihil has joined #openscad
gunnbr has joined #openscad
othx has quit [Remote host closed the connection]
othx has joined #openscad
gunnbr has quit [Ping timeout: 240 seconds]
Xeha has quit [Ping timeout: 252 seconds]
Xeha has joined #openscad
<gbruno> [github] t-paul closed issue #2736 (SVG: Respect display: None). https://github.com/openscad/openscad/issues/2736
<gbruno> [github] t-paul pushed 3 additions 5 modifications (Merge pull request #4041 from cwalther/displaynone
<teepee> yay! print finished
<ecraven> whee, the air assist seems to work ok ;D
<InPhase> Of?
<InPhase> Oh, I missed a conversation didn't I.
<teepee> looks like it :)
<teepee> first PETG print
<teepee> the "clear" tubing part for the signal tower
<teepee> I also got some TPU, but that may need to wait a bit
<teepee> so will it get off the printbed now ;-)
<ecraven> does anyone use openscad for designing stuff for laser cutters?
<teepee> yes, no idea how many
<teepee> there's some restrictions as you can't have single cuts, but it seems to be quite ok in general
<teepee> wasn't there a hackaday post just a couple of days ago
<teepee> ok, it says "cnc cutting" in general but I guess the 2d output should just work for laser too https://hackaday.com/2022/01/02/an-openscad-library-for-all-your-cnc-cutting-needs/
<ecraven> thanks, I'll read that!
<ecraven> (I'd love a general 3d cnc thing, but right now, I only have a 2d laser cutter ;)
<InPhase> teepee: I wonder if you can make a hue-based clock.
<InPhase> teepee: One level for 24 hours, one for 6 hours, one for 1 hour, and maybe another one for day of week, cycling through a periodic rainbow hue.
<InPhase> teepee: With a little practice you could tell the time and day of week by color at a glance.
<InPhase> I geuss you could do a 15 minute one as well if you have 5 layers, and then you'd get pretty good minute-level time precision.
<teepee> hmm, seems possible, but I suspect I'm too lazy trying to read that :)
<teepee> I even consider adding a digital read-out to the 60led ring clock :D
<InPhase> I figure you probably want ratios around 4-6 between the levels, matching up to the real world items of importance. I think 15 min, 1 hr, 6 hr, 24 hr, 1 week hits that nicely. With continuous color change rather than discrete jumps, and it would be a subtle background thing.
<InPhase> Well it would be like resistor decoding. Rough at first, but then it gets obvious after you get used to it. :)
arebil has joined #openscad
ToAruShiroiNeko has quit [Read error: Connection reset by peer]
<teepee> the "no fan needed" part is also nice
<teepee> not seeing much difference between the PLA and PETG part
ferdna has joined #openscad
ToAruShiroiNeko has joined #openscad
<Scopeuk> teepee i built a 60 led ring clock (technically tape feeding into the end of glue sticks as light pipes), it reads pretty easy
<Scopeuk> http://scopeuk.mypicture.info/clockSilent.3gp still not finished properly, segments of the frame need reinforcing
<Jack22> the flickering makes it look like neon tubes
<Scopeuk> The flickering is a camera artefact, its not visible in person (the led's are ws2812's)
<Jack22> scopeuk doesn't mean you could make that flickering visible - Ü
<Scopeuk> Yeh, at the moments the led's only refresh once per second, could easily put an effect over it with more redraws
<teepee> Scopeuk: that's a cool effect
<teepee> hmm, something similar, but 3d printed with clear filament?
<teepee> there's an idea
<teepee> can't do more signal towers right now, led rings are out of stock ;-)
<Scopeuk> I've wondered about 3d printed light pipes, the hot melt glue works so well for it though
<teepee> maybe not just light pipes but even the numbers, or at least 4 of them
<teepee> or use a bigger 12 led ring to produce the hours inside the 60 led ring that does minutes
<Scopeuk> I have also pondered designing a grid holder for glue sticks that would clip over one of the n by m ws2812 grids and then cut them all to different lengths just to do pretty patterns. I wonder if you could do the numbers as blacked out sections on the sticks and then turn on an entire segment in the hours colour
<Scopeuk> To show a number silouet, not sure how well it would work
arebil has quit [Quit: My keyboard has gone to sleep. ZZZzzz…]
mitte has joined #openscad
<mitte> meta-k /input grab_key_command
mitte has quit [Quit: WeeChat 3.4]
mitte has joined #openscad
mitte has quit [Client Quit]
<JakeSays> Scopeuk: https://imgur.com/a/MiBfCGp
<JakeSays> Scopeuk: oh wow those glue sticks look awesome!
<JakeSays> that is a really cool idea
<Scopeuk> Blinken litghts have to love em
<JakeSays> they're awesome!
<JakeSays> Scopeuk: i'm gonna steal your stick idea if you don't mind
<JakeSays> my next ws2812 project is going to be heart shaped light that tracks my heart beat in some fashion
<Scopeuk> Absolutely go for it, i picked it up from my bike lights https://fibreflare.com/
mitte has joined #openscad
* JakeSays orders glue sticks!
<JakeSays> Scopeuk: hmm. i'm not sure i'd want numbers on it.
<JakeSays> i'm an "ish" kind of person when it comes to time so it'd work well w/o numbers
<Scopeuk> Yeh that was a different discussion as to how it could be done
<Scopeuk> I've no intention of doing it on mine
<JakeSays> those #'s i made will be house numbers so doordash can find me
<JakeSays> they're solderless
<JakeSays> Scopeuk: are you using ws2812's?
<Scopeuk> That one is yes, 60 led per meter tape
<JakeSays> ah. i used "buttons" - not sure what they're called, but they're individual ws2812's on a little breakout board
<Scopeuk> I've seen similar things
<JakeSays> Scopeuk: these https://amz.run/5D71
<Scopeuk> Fair enough
<JakeSays> next year i'm going to print all of my christmas tree lights using them
<teepee> heh, sounds like a candidate for the next advent calendar then :)
<Scopeuk> I had pondered a piece a day 80's style robot to print and assemble https://i.pinimg.com/originals/ba/09/8a/ba098abd133820361d075ac7b149f59d.jpg style, but i'm not sure it would nicely breakdown into enough interesting bits or if "thing a day" style is more eciting
<JakeSays> teepee: hey that could be cool
<InPhase> teepee: Typically the only thing PETG will give you over PLA is a slight bit more flex when you squeeze it.
<teepee> no need for extra fan is also nice
<InPhase> teepee: This has been important for me when printing combs for example, as PETG has just enough flex on a comb to be comfortable with the thin comb teeth, while PLA is a little too rigid for a comb.
<InPhase> teepee: Also I used PETG on the chair feet I printed, so it would flex ever so slightly while wedging them onto the bottom of the metal chairs.
<teepee> slightly higher glass temp can't hurt either
<InPhase> Right. I think I used it for that motivation once in a replacement car part in a previous car of mine.
<InPhase> A shield designed to keep leaves out of the engine disappeared from my car.
<teepee> ah, yes, that sounds like a useful thing to have
<InPhase> It held up fine up through trading in the car.
<teepee> uh, oh, big clive live with 3 scad scripts :D
<InPhase> Probably it's still on there with some new owner.
<teepee> nice, reminds me of the tweet by Prusa offering 3d printed stuff to fix Teslas
snaked has quit [Quit: Leaving]
ferdna has quit [Remote host closed the connection]
ndim has quit [Ping timeout: 245 seconds]
ndim has joined #openscad
<JakeSays> i have a hard time printing petg
<JakeSays> i've been using pla+ and i like itttttt
<teepee> why? so far it works fine here
<JakeSays> minus all the t's
<teepee> second print almost ready, 1mm to go
<JakeSays> teepee: bad luck probably.
<JakeSays> i haven't dedicated a ton of time to it though
<InPhase> PETG tends to be super easy to print. It caught on mostly because it prints identically easily to PLA.
<teepee> the spool said 230°, bed ~75° and that works
<JakeSays> i had to go with 240/80 and print super slow to get anything out of it
<InPhase> It does take a higher temp to do PETG though. I tend to print it at 250C.
<JakeSays> i also bought a spool of PC on a whim
<InPhase> But I also do PLA at 230 after many strength tests.
<JakeSays> really?
<teepee> oh? 250?
<JakeSays> hmm
<JakeSays> InPhase: pla prints stronger at 230?
<teepee> right now I'm at 230° and the good old prusa nozzle is said to require higher temp
<InPhase> JakeSays: For me.
<InPhase> These are not exactly scientifically calibrated instruments.
<InPhase> JakeSays: I'm not 100% certain we're all getting the same real temperatures when we setup settings for them. :)
<JakeSays> true that
<teepee> indeed
<teepee> that said, I tended to need > 200° for PLA where most people said 190° is plenty
<InPhase> Oh you can print it at 190.
<teepee> but then it also depends hugely on the actual filament
<JakeSays> i use 200 for pla, 220 for pla+
<InPhase> But at 190 it delaminates super easily.
<InPhase> JakeSays: Now try PLA at 230 and compare it to PLA+ at 220. :)
<JakeSays> InPhase: what does the extra heat do to pla?
<InPhase> JakeSays: It comes out so hot that it very slightly melts the layer beneath, causing a more thorough bonding between layers.
<InPhase> But it's not so hot that it does any deforming.
<JakeSays> ah
<JakeSays> i dont think that'll make it harder than pla+.
<InPhase> JakeSays: I don't have any PLA+ to test with, but if you tried it I'd love to hear the results.
<JakeSays> i'll do some experimenting
<JakeSays> Scopeuk: did you use white or clear glue sticks?
<Scopeuk> JakeSays those are clear
<JakeSays> ok cool. i think i'll make the 5 minute ones a little longer
<JakeSays> cool. i'll have 250 of 'em tomorrow.
<JakeSays> perfect timing. the batteries in my wall clock went dead
<InPhase> Oh. I missed the glue-gun stick LED clock link above. That's awesome.
<JakeSays> it is very cool
<JakeSays> Scopeuk: what's the inner diameter of your clock?
<InPhase> Scopeuk: How long does it run on the battery between charges?
<JakeSays> what kind of glue did you use to glue the glue sticks? :D
<Scopeuk> InPhase not sure, it got prototyped and not pushed into use intention was to switch to a psu
<Scopeuk> The sticks are friction fit into a turret
<JakeSays> i know. i was being clever
<JakeSays> well, trying to anyway
<JakeSays> i have an esp32 i'm going to use to control it. all kinds of ideas are coming to mind
<JakeSays> Scopeuk: ah cool thaks!
<JakeSays> *thanks
<Scopeuk> Mine is running on an attiny4313 with an external clock
<Scopeuk> The tiny section size is due to a small printer
<JakeSays> with the esp32 i can make it flash to use as an alarm kind of thing
<JakeSays> or change the colors based on the weather/temp, etc
<JakeSays> i really dislike the clock i have now.
<InPhase> Scopeuk: I never realized glue sticks worked that well as diffusive wave guides though. That's basically the ingredients for a battery powered glow stick. :)
<InPhase> Or if we can buy big enough glue sticks... Or make a mold and melt some in... A light saber toy.
<JakeSays> i've molded hot glue before. it works pretty well.
<Scopeuk> That design works but is a little delicate
<JakeSays> Scopeuk: yeah i'll print the ring thicker
<Virindi> gah, I got nightly and it throws a syntax error on one of my huge files that the stable version does not
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openscad
<Virindi> hmm, I guess the ability to have unnamed parameters was removed, that is okay
<Virindi> preview is indeed faster :)
<teepee> that could be a stray comma
<Virindi> yes, it is. In someone else's files.
<teepee> which one is that? if it's not secret
<Virindi> heck, the preview is even usable on the 7600u. very nice
<Virindi> uhh, I think it was some helper files that originally came from nophead
<Virindi> it was stuff that got included that I didn't actually use
<teepee> ah, simple fix then ;-)
<Virindi> my painful preview test is a 3d printer of course, it has those files included just to get a model of an e3d v6 :)
<Virindi> I enabled vertex-object-renderers*, should I also enable lazy-union? is there any downside to these options?
<teepee> to the vertex stuff no, if the preview looks fine
<teepee> lazy union can potentially produce files that other tools don't like when it produces overlapping volumes without actually merging those
<Virindi> ah I see, so does that just affect render and not preview?
<teepee> pretty much, yes
<teepee> it delays actual union computation which can hugely speedup some stuff
<JakeSays> what's the advantage of using the vertex object render stuff?
<teepee> uploadig more data to GPU for preview
<Virindi> it seems to be effective for even onboard graphics
<JakeSays> hmm. i'll try it
<JakeSays> i should probably use a stable version though
<teepee> you can just have both :)
<JakeSays> i'm currently using a preview build from october that has the render() as a function stuff
<teepee> there's not a huge amount of changes since then
<Virindi> is openscad a QT application? my next task is to figure out why it isn't scaling (that laptop has a high-dpi screen). I didn't mess with QT yet though
<teepee> yes, Qt and it did scale fine for me on Debian
<teepee> well, mostly fine
<Virindi> I'm sure it will when I configure QT then :)
<teepee> the status bar was a bit strange
<JakeSays> Virindi: are you on windows?
<Virindi> haha no.
<teepee> it did take the Gnome/Wayland settings
<JakeSays> ah. i have scaling issues with qt on windows
<Virindi> I won't let that poison near a computer I have ;)
<JakeSays> but it works fine on linux
<teepee> just forgetting things sometimes when switching monitors, it sometimes reverted to 1:1
<JakeSays> i am required to use windows at work
<Virindi> I am using xfce, I set it it to 2x gui scaling and messed with the font dpi
<Virindi> but I read QT has to be done separately
<Virindi> since I am not using KDE
<Virindi> or whatever
<teepee> ok, I don't know much about xfce
<JakeSays> oh hmm i have vor enabled
<JakeSays> but only one of the four selected
<JakeSays> should i enable all of 'em?
lastrodamo has quit [Quit: Leaving]
<teepee> I have all 4 enabled
<JakeSays> is a restart required?
<teepee> I'm not sure