<teepee>
some older instructions have gpg -deamor but it seems not needed anymore
<teepee>
well minus the typo
<UltimateCodeWarr>
If the goal is to just run that command, just say RUN THIS: wget...
<UltimateCodeWarr>
What is the purpose to show all that Key ID stuff?
greenbigfrog has joined #openscad
<UltimateCodeWarr>
I'm thinking, ok, I got to cut and paste this into some file and import it...
<UltimateCodeWarr>
Why muddy the waters by listing it
<teepee>
because some people care to check that the downloaded file matches the fingerprint shown on the page
<UltimateCodeWarr>
Then say that
greenbigfrog has quit [Ping timeout: 252 seconds]
<teepee>
nope
<UltimateCodeWarr>
so a bit of a cliffhanger here
<UltimateCodeWarr>
added the url to the openscad.list .... and then what?
<UltimateCodeWarr>
How do I get it to pull down the most recent version?
<teepee>
how would I know? it's your distribution
<teepee>
i do "apt update" and "apt upgrade"
<teepee>
you do whatever your system has setup
greenbigfrog has joined #openscad
UltimateCodeWarr has quit [Quit: Client closed]
teepee_ has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
greenbigfrog has joined #openscad
teepee has quit [Ping timeout: 260 seconds]
teepee_ is now known as teepee
LordOfBikes has quit [Ping timeout: 265 seconds]
UltimateCodeWarr has joined #openscad
paddymahoney has quit [Ping timeout: 276 seconds]
LordOfBikes has joined #openscad
paddymahoney has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
greenbigfrog has joined #openscad
SamantazFox has quit [Killed (NickServ (GHOST command used by SamantazFox_))]
SamantazFox_ has joined #openscad
<UltimateCodeWarr>
So I couldn't figure out how to make wget work right. I tried the sudo apt-get update / upgrade nothing really happened. So I just went and downloaded it directly here:
<UltimateCodeWarr>
Got a date of 10.1.2024 nightly build
paddymahoney has quit [Ping timeout: 252 seconds]
<UltimateCodeWarr>
The F5 render works, but F6 does not
<UltimateCodeWarr>
I take that back, I had to switch a dropdown in the preferences from Rendering Old to Manifold. I think it used to be a checkbox
greenbigfrog has quit [Ping timeout: 252 seconds]
<UltimateCodeWarr>
I did quite the "Hail Mary" move today switching from Windows 11, new PC, Installed Linux. Using a HP Elite 800 with i9/DDR5 seems to be about 2x as quick rendering
<UltimateCodeWarr>
Was getting around 23 seconds to F6 the model, now it's around 13 seconds
<UltimateCodeWarr>
Next Stop: Mac Pro, it has 800 GB/S vs DDR5 wimpy 64 GB/S. Of course the price is exhubernt
greenbigfrog has joined #openscad
paddymahoney has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
UltimateCodeWarr has quit [Quit: Client closed]
PolyBytes has quit [Quit: Client closed]
greenbigfrog has joined #openscad
paddymahoney has quit [Ping timeout: 252 seconds]
paddymahoney has joined #openscad
TylerTork has joined #openscad
<TylerTork>
Terminology question: In some of my scripts I find I need a parameter to adjust for the degree of precision of the 3D printer. It's basically the distance apart two parallel lines of filament need to be to not end up glued together. Is this common, or is there some other way the measurement is expressed, and what do people call it?
UltimateCodeWarr has joined #openscad
<TylerTork>
Nobody?
<UltimateCodeWarr>
Is there a way if you just joined to list the questions posed in the psat?
<TylerTork>
yes, there's a log you can get to by clicking the "i" at the top right
<UltimateCodeWarr>
Thx
<UltimateCodeWarr>
I know in other CNC systems there is a 'Kerf' that is the size of the tool cut
<UltimateCodeWarr>
So, I would think it would have something to do with the nozzle size on your printer. .0.2 mm - 0.8 seem to be what's normal these days
<UltimateCodeWarr>
Mine came with a 0.4 mm nozzle
<TylerTork>
yes, I'm not trying to determine the value of the number -- I want to know what people call it in this context and whether they think of it the way I expressed it.
<UltimateCodeWarr>
So you were asking how do I space a line 1 kerf apart so they won't connect.
<TylerTork>
No, I'm asking what do they call that distance? I don't think "kerf" is the common term in this situation.
<UltimateCodeWarr>
I guess a way to attack something like that would be "Emprically" where you just run off some test patters of gradually increasing widths between lines over a run.
<TylerTork>
You seem to be not understanding the question. I thank you for trying, but I need you to stop now.
<UltimateCodeWarr>
What is the minimum (tiniest) air gap that you can print between lines?
snaked has quit [Remote host closed the connection]
snaked has joined #openscad
mmu_man has quit [Ping timeout: 252 seconds]
abff has joined #openscad
germ_ has joined #openscad
myosotis has quit [Ping timeout: 260 seconds]
greenbigfrog has quit [Ping timeout: 260 seconds]
hyperair has quit [Ping timeout: 252 seconds]
paddymahoney has quit [Ping timeout: 252 seconds]
paddymahoney has joined #openscad
<JordanBrown>
I'm not a 3DP expert, but I call it something like "gap".
<JordanBrown>
TylerTork^
<InPhase>
TylerTork: I use the term "clearance".
<JordanBrown>
And note that that distance might be different on the first layer than on later layers, and may be different in XY than in Z.
<InPhase>
TylerTork: Sometimes I hardcode a clearance, and provide a 0 centered adjustment relative to my suggested value, and then I call this adjustment "tolerance".
<InPhase>
TylerTork: What I mean by tolerance on a particular part varies, however. I use this term somewhat liberally.
<InPhase>
JordanBrown: Yeah. It is tricky when that varies a lot. I try hard to minimize this being a problem by thinking about print-friendly designs.
<InPhase>
I will very often completely reshape a design just to make it print better, while still achieving the core objective.
greenbigfrog has joined #openscad
<JordanBrown>
I am not advanced enough to figure out how to do that.
greenbigfrog has quit [Ping timeout: 252 seconds]
greenbigfrog has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
greenbigfrog has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
greenbigfrog has joined #openscad
<UltimateCodeWarr>
Under Ubuntu, if you set your Mouse Cursor Size to Large, it Shrinks back to small when you hover over the OpenSCAD window.
<InPhase>
JordanBrown: Uhm, on reflection I probably didn't earn that. There's a certain hard to explain intuitive approach to mechanical design that my grandfather, father, my kids, and I all started doing at a young age. I wasn't actually sure it was a pattern until I saw my kids doing it. But that's the part of my head I lean on when I'm thinking about print friendly design. I could list principles maybe,
<InPhase>
but I'm not sure I'd know how to really explain it.
<InPhase>
Other things I can explain, but maybe not that one so much. :)
<InPhase>
But I'll ponder it. Maybe enough reflection on the principles would reveal extractable patterns or rules.
califax has quit [Remote host closed the connection]
califax has joined #openscad
greenbigfrog has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
TylerTork has quit [Ping timeout: 256 seconds]
hyperair has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 260 seconds]
teepee_ is now known as teepee
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
<Scopeuk>
I guess you are dealing with hard constraints and then design goals, your hard constraints must always be met (that mounting point isn't moving etc) but your goals can be thrown to the wind if it's aesthetically pleasing or reduces print time etc
UltimateCodeWarr has quit [Quit: Client closed]
<JordanBrown>
InPhase part of the picture is even understanding what makes a design more print-friendly. I know some obvious ones, of course, like avoiding overhangs and supports, but it seems like you're talking about more than that.
UltimateCodeWarr has joined #openscad
greenbigfrog has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
mmu_man has joined #openscad
califax has quit [Remote host closed the connection]
califax has joined #openscad
<pca006132>
UltimateCodeWarr: your comment about Mac Pro 800GB/s vs DDR5 64GB/s is not really correct
<pca006132>
800GB/s is the total bandwidth, CPU + GPU combined. CPU cannot use that much bandwidth anyway.
<pca006132>
and this GPU bandwidth is not special, recent highend GPUs have bandwidth higher than this (but less VRAM)
califax has quit [Remote host closed the connection]
<Deneb>
can I change the background in OpenSCD to a custom colour?
<Deneb>
I would also like the editor background to be dark or near dark. Is that possible?
<InPhase>
JordanBrown: I suppose one way to express the approach is to not necessarily avoid overhangs, but just accommodate how they behave. Like if you functionally needed space for a cylinder, and the top of that cylinder would be an overhang, add a dome cap and you gave the overhang a tolerance space to be defective without altering how a cylinder insertion behaves. And then once you know you can do this
<InPhase>
for a cylinder, if a tight fit is required for that cylinder hole, you should really work to orient the whole design so that this is the cylinder's print orientation, even if that wildly alters how the printed item behaves. One can make similar considerations for gap merging, stringing, strength lines, delamination propensity, and so on.
paddymahoney has joined #openscad
<buZz>
Deneb: there's themes you can select for the 3D render inside the preferences
<buZz>
and also themes for the editor
<buZz>
i'm not sure, but i assume you can edit the themes too , or add one
<Deneb>
yes, I found the themes in preferences. Some of them are duplicated, for example, Nature is white, the same as Tomorrow. BeforeDawn and Ocean look identical, so do Cornfield and Solarized.
paddymahoney has quit [Ping timeout: 265 seconds]
<Deneb>
Just found the option for the editor background in the editor tab. Thanks.
<buZz>
Deneb: they arent really the same ;) there's more colors than just the background defined in them
<Deneb>
Was hoping there was a config file that defines the colours somewhere.
<buZz>
i assume those themes do originate from somewhere, sure
<Deneb>
buZz: yes, now that I have drawn one, I see that the cube changes color as well.
<buZz>
worstcase, find em in the sourcecode ;)
<Deneb>
I kind of expected DeepOcean to have a dark blue background for some reason.....
paddymahoney has joined #openscad
<Deneb>
Anyway, I found a combination that is not so tiring to the eyes. Not ideal, but it will do. Just going through the tutorial.
<Deneb>
I downloaded through apt get so don't have source code, but yes, that would be another option I guess. Could compile it from source.
<InPhase>
JordanBrown: Some of the reasoning process that follows is illustrated in this design: https://www.thingiverse.com/thing:3024698 It was like, I needed a cord wrap, but I can achieve functional equivalency with the base of the cord wrap all on the build plate, as it will hang in the air which removes concern for the non-ideal wrap having its axial point at the base. And I need functional stability of
<othx>
InPhase linked to "Toothbrush / Toothpaste Holder, Philips Sonicare and More by rcolyer" on thingiverse => 7 IRC mentions
<InPhase>
the wall mounting against downward pressure, so the mounting point needs to be below what I just established as the build plate, so I need a secondary mounting insert. And then I need the insert and the hole it inserts into to have a mutually compatible printing shape for the two printing orientations, where I can hide the defects, so I give a little tolerance for the mounting bracket's overhang to
<InPhase>
hide the defects inside, and put the pretty side up to not pick up visible defects, as I already left space for invisible ones inside at insertion.
<buZz>
Deneb: i ment mostly to see where they originate from ;)
mtm has quit [Ping timeout: 248 seconds]
mtm has joined #openscad
<InPhase>
JordanBrown: The "The holder conveniently lifts up off of the mount for ease of cleaning" is kind of a retroactive lie. It does do this, but the justification for this outcome followed from my print friendliness reasoning about matching up things like that cord wrap requirement and the lower anchor point requirement. Removability was kind of an accidental bonus byproduct of printability consideration.
paddymahoney has quit [Ping timeout: 252 seconds]
greenbigfrog has quit [Ping timeout: 260 seconds]
TylerTork has joined #openscad
paddymahoney has joined #openscad
TylerTork has quit [Client Quit]
greenbigfrog has joined #openscad
mmu_man has quit [Ping timeout: 252 seconds]
<J24k83>
Deneb the colors are defined in json within the render folder https://imgur.com/a/WuqnrYR you can add your own design if you like
Guest32 has joined #openscad
<Guest32>
hello, Ihave a problem with import of a .stl:
<Deneb>
J24k83: thanks. Yes, I found those json that directory.
<J24k83>
just increase the index nr if you add a new one
<J24k83>
Guest32 STL are sometimes broken
<Deneb>
ah, was wondering about that. Thank you.
<J24k83>
Guest32 what is your problem?
<Deneb>
I am still only going through the tutorial, but say I want to create a cube with a hole or recess in it, how do I make this?
<Guest32>
I want to print a free T-Rex.stl but the normal support material is not procticable for me. so I made a import("T-Rex.stl"); and put some cubes to support. when drawing the construction, everything looks god, when Rendern nothing appears anymore on the screnn and I dont receiv an error message. what can I do ? - sorry I am a newbee in this forum
<Guest32>
although I have been working with openscad since 2020 and have done many parts .
<Deneb>
Oh wait, its in the next section. Trying to run before walk....
<InPhase>
Guest32: Almost certainly you are importing an stl file which is non-manifold, and thus cannot be geometrically combined with other objects unless it is repaired or replaced.
<Guest32>
I dont know anything about this stl I want to import, so how can I proof this and if its not ok can I repair it ?
TylerTork has joined #openscad
<InPhase>
Proving it's non-manifold is tricky, as the defects can be very hard to spot. Repairing it can be sometimes done in tools like meshlab.
<TylerTork>
hey y'all is there any reason we can't add a pitch argument to roof, default 45 degrees? It's easy to implement, just scale z by the tan of the angle, but does everyone have to figure that out for themselves?
<InPhase>
Guest32: Manifold specifically means that the model is basically "watertight" in the triangles describing the surface and without any self-intersections of these.
<J24k83>
Guest32 Slicer normally have the option to "paint" supports - that is much better - another option is to add geometry (within the slicer) and define them as support - so you can remove them
<InPhase>
Guest32: However this is not a goal of many tools that output stl files that are designed only for display, and not for volumetric printing.
mmu_man has joined #openscad
<Guest32>
OK I think I understand the problem a little, I will try other stl- Files for I dont want to build a T-rex on my own and a grandchild wanted me to print it.
<InPhase>
Another stl file is certainly the easier option when that's possible. :)
<J24k83>
InPhase if a face is missing - isn't it still manifold? I thought manifold is that no point has more than three faces attached
<Guest32>
Thank your for your fast infos and help.
greenbigfrog has quit [Ping timeout: 252 seconds]
<pca006132>
J24k83: just consider a cone, one point can have many faces attached
<J24k83>
pca006132 oh right .. maybe two volumes can't share points
<J24k83>
ah it was surfaces not faces
<InPhase>
pca006132: How is self-intersection addressed by that definition?
<pca006132>
there are two things here, one is the topological definition which only consider connectivity between vertices, one is the geometric definition
<pca006132>
self-intersection is geometric, and is unrelated to the topological definition
<InPhase>
pca006132: So the manifold library will operate fine with self-intersecting stl inputs?
<InPhase>
Provided they are otherwise meeting that definition?
<pca006132>
minor self-intersection, yes
<pca006132>
see the epsilon-valid requirement
<pca006132>
in particular if points are touching, it is always fine because our epsilon is always non-zero
<InPhase>
pca006132: Well what addresses major self-intersections?
<pca006132>
we don't allow major self-intersection
<InPhase>
But that's not in the manifoldness definition?
<J24k83>
isn't an intersection violating the shared point definition
<InPhase>
Is the requirement then "manifold and non-intersecting"?
<pca006132>
yes, we need manifold and non-intersecting (epsilon-valid)
<pca006132>
shared point => intersection, but it is epsilon valid because it exists a small perturbation to make it non-intersecting, however small epsilon is as long as it is non-zero
<pca006132>
and this is the reason why manifold's output can make CGAL mad, CGAL requires strictly non-intersecting manifold
<pca006132>
so if you have, e.g. two cubes with a shared corner vertex, CGAL will not accept it
<J24k83>
just noticed an intersection can happen without changing the faces of a valid objects - but just moving them
<pca006132>
yes
<InPhase>
pca006132: That seems a robust enough set of definitions.
<InPhase>
pca006132: I will have to try to update my vocabulary to that choice. Seems meritorious in its simplicity.
<InPhase>
It requires telling people, "Most likely your imported stl is non-manifold or self-overlapping."
<J24k83>
or both - Ü
<InPhase>
:)
<J24k83>
wonder if "or" is XOR
<InPhase>
pca006132: How does cube(10); translate([10,10,0]) cube(10); work out in the manifold library?
<J24k83>
shouldn't be manifold
<pca006132>
I think if things are self-overlapping, we will just export some non-sense
<pca006132>
your particular sample is fine
<pca006132>
not entirely sure if it is considered one thing or two things
<InPhase>
pca006132: Does it not violate the manifoldness definition?
<pca006132>
no, it does not violate that
<pca006132>
and it is epsilon-valid
<InPhase>
How about if I wrap union() around it?
greenbigfrog has joined #openscad
<InPhase>
Still okay? :)
<pca006132>
yes
<J24k83>
that connection point need to exist twice or?
<pca006132>
yes, we just duplicate the point
<pca006132>
so topologically it is fine
<pca006132>
geometrically it is touching, but we can offset them slightly to make it non-intersecting, so it is epsilon-valid
<InPhase>
So the touching edges are treated internally as two same-valued edges.
<pca006132>
no, they are different... I think the points are duplicated
JordanBrown has quit [Read error: Connection reset by peer]
<pca006132>
was working on something in parallel and my brain is not functioning correctly
JordanBrown has joined #openscad
<InPhase>
pca006132: :) Sorry for the pathological question. Just trying to parse the edge cases of that definitional choice.
<pca006132>
I guess you should just try it and look at the mesh
<pca006132>
I just remember it is fine, forgot how exactly the topology should look like
killjoy has quit [Ping timeout: 252 seconds]
hisacro has quit [Ping timeout: 252 seconds]
mmu_man has quit [Ping timeout: 272 seconds]
killjoy has joined #openscad
killjoy has quit [Changing host]
killjoy has joined #openscad
<pca006132>
I think if they have to be combined into one manifold, the resulting thing is probably not epsilon-valid because there is no global perturbation to make it non-intersecting?
<pca006132>
but if you keep them as two objects, it is epsilon-valid
<pca006132>
and yeah probably no point duplication
<pca006132>
confused with the corner touching case
<InPhase>
Well, I output as stl and I just get triangles.
<InPhase>
The representation of edges disappears on me.
<pca006132>
try 3mf
<InPhase>
:( This is a binary format.
<pca006132>
unzip it and you can see the content
<pca006132>
I think it is just xml inside
<InPhase>
Ah yes.
<InPhase>
I noticed the "PK" at the top of the file right before you said that. I hadn't thought to hex dump one of these before.
<InPhase>
14 vertices in the 3mf
<InPhase>
So no duplication.
<InPhase>
But I don't know if this happens at output and it might still be split during processing?
hisacro has joined #openscad
<pca006132>
idk, maybe I can have a look later
<pca006132>
fixing some cmake stuff now, hurts my brain
<InPhase>
pca006132: And the 3mf file contains 4 triangles with 6,7 vertices. Consequently it would not meet the 3mf file definition of manifold?
<InPhase>
pca006132: Paste-binning for later inspection, to avoid duplicated effort: https://bpa.st/2XT3O
<InPhase>
I have no problem really, as it works fine. ;) Just wondering how to interpret the descriptions properly.
<pca006132>
yeah it is fine, I am a bit curious about this as well
<pca006132>
I never really look into the definition closely
<J24k83>
Also possible that 3mf is eliminating double vertices
<pca006132>
no, 3mf doesn't do that
<J24k83>
ok .. still funny that none of those microsoft links work anymore (and nobody remove or change them)
mmu_man has joined #openscad
paddymahoney has quit [Ping timeout: 260 seconds]
paddymahoney has joined #openscad
greenbigfrog has quit [Ping timeout: 252 seconds]
L29Ah has left #openscad [#openscad]
<pca006132>
J24k83: they are in the spec
<pca006132>
(just the spec url is 404)
J24k83 has quit [Quit: Client closed]
J24k83 has joined #openscad
<J24k83>
doesn't make sense to refer to non existing urls in specs or did i miss something here ?
greenbigfrog has joined #openscad
<pca006132>
they probably shouldn't refer to urls in spec... unless it is kind of permanent like doi
<InPhase>
I guess maybe there is nothing like a Klein bottle without that self-overlapping part.
<InPhase>
I was trying to think about whether or not it was possible to make a pathological object like this which meets the manifold library definition.
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<InPhase>
I think the laws of mathematics guard against this properly.
myosotis has joined #openscad
kintel has joined #openscad
<J24k83>
if a face has flipped normal it shouldn't be manifold anymore as the order of vertices is now reversed
<kintel>
J24k83 Exactly
<kintel>
..but the repair mentioned about will likely catch that and flip it back
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<cart_>
What is the easier way to make edges rounder without making the item bigger? When I use minkowski() it enlarges the item and also sometimes does not really do much in that sense
<J24k83>
cart_ in 2D or 3D - in general every edge that is round gets a radius applied - so to keep the size you need to remove that radius - but then the rounding will make the object smaller as now the corner is outside the rounding
<J24k83>
f0lkert default colors may irritate everyone using other color-schemes (some using custom ones for command line export)
<J24k83>
Oh btw did we consider exporting colors in SVG?
muesli has quit [Quit: NO CARRIER]
muesli has joined #openscad
<buZz>
cart_: in general, i just make a 'roundbox()' function , which has a 'radius' parameter for the rounding, then subtract the radius on all sides of box, and do a offset(radius)
<buZz>
but yeah, offset() just works in 3D
<buZz>
minkowski() imho should be avoided at all costs , hahaha
<buZz>
just works in 2D*
<buZz>
dangit
<f0lkert>
J24k83: true. but I'd like to get a first version in the field.
<f0lkert>
to see if people use it
<f0lkert>
(by the number of bug-reports :-])
<buZz>
f0lkert: that only works on non-perfect code
<buZz>
so, not yours
<buZz>
;)
<J24k83>
If you keep it complex - you have the chance of getting more bug reports - Ü
<f0lkert>
buZz: ;)
<buZz>
f0lkert: tried it on a big-endian machine yet? :P
<f0lkert>
buZz: it'll work, most def
<buZz>
¯\_(ツ)_/¯
<buZz>
either way, now thats it committed, i assume its coming to openscad-nightly $soon
* J24k83
has already PovRay installed
<J24k83>
was hoping there was a "modifier" to render only certain parts .. but all i could find is to use »if end«
<InPhase>
With proper handling of all sorts of anomalous input conditions.
UltimateCodeWarr has joined #openscad
<buZz>
many roads to rome :)
<f0lkert>
buZz: I think there's quite some overlap between the two languages (oscad / povray). so the goal for my next export version would be the constructs
<f0lkert>
what do you call it
<f0lkert>
difference this that
<buZz>
f0lkert: like 'heavy' for puredata to C :)
<f0lkert>
I guess!
<buZz>
oh, yeah , the buildtree?
<buZz>
i think thats what its called
<buZz>
all the objects, instead of the resulting mesh