<Jack21>
and precisely on time opened .. zzzzzzzzzz
ferdna has joined #openscad
noonien has joined #openscad
noonien has quit [Client Quit]
JakeSays has quit [Ping timeout: 256 seconds]
noonien has joined #openscad
noonien has quit [Client Quit]
noonien has joined #openscad
JakeSays has joined #openscad
JordanBrown has joined #openscad
<JordanBrown>
Why is [2]*[2] equal to 4, rather than [4]?
<JakeSays>
well. that was fun. i was literally sitting here staring at my printer trying to decide if i should cancel this 30 hour print, and the power went off in my house. lol
<JordanBrown>
Ah, because those are one-dimensional vectors rather than two-dimensional vectors?
<JordanBrown>
I put my 3D printer on a UPS.
<Alicia>
I'm getting [4] when I try it
<JordanBrown>
You are trying echo([2]*[2]); ?
<JordanBrown>
I get "ECHO: 4".
<Alicia>
oh, I misread the * as +, * does indeed give me 4
<JordanBrown>
But I bet that it's because it's two one-dimensional arrays, and I don't know what it means to multiply two one-dimensional arrays.
<JakeSays>
JordanBrown: i get the same
<JordanBrown>
I'm sure it's consistent. I just wasn't figuring out why it was getting the scalar result.
<JordanBrown>
Of course, pretty much as soon as I said something I realized the dimensionality issue.
<Alicia>
it seems to perform the operation index pair by index pair and then sum the results, [2,3]*[2,5]=19
<InPhase>
JordanBrown: It's a dot product.
JakeSays has quit [Ping timeout: 252 seconds]
<JordanBrown>
Now if only I understood dot products :-)
<JordanBrown>
(I have the Wikipedia page up in front of me, but I don't really understand it.)
<JordanBrown>
I mean, I could do the math that it says to do, but I don't understand what it geometrically *means*.
<InPhase>
JordanBrown: Geometrically, the dot product takes one vector and drops a projection (shadow) onto the other vector, and multiplies those two values. So it is a magnitude-scaled similarity metric.
JakeSays has joined #openscad
<InPhase>
So if they're pointing at right angles, it's zero. If they're pointing the same direction, it's the product of their lengths (norm(v) in OpenSCAD). If they're pointing in opposite directions it's the negative of if they're in the same direction.
<ccox>
Or the cosine of the angle between the two vectors.
<InPhase>
When broken down as orthogonal components, this is done as the sum of pairwise multiplication. If you have vectors as a magnitude and direction, it's the product of the magnitudes times cos(angle) for the direct angle between them.
<JordanBrown>
so kind of sort of dimly an intersection-like operation.
<ccox>
very, very useful in geometry, and computer rendering
<InPhase>
Yes. An intersection-like multiplication.
<InPhase>
This shows up all over physics.
<InPhase>
It's the key way to do multiplications that only act when things are in the same direction.
<ccox>
(well, so does the cross product - physics likes it's math)
<JordanBrown>
this is what happens when you jump straight out of high school into a programming job, and don't take high-end math and physics.
<JakeSays>
first snow of the year. people must be sliding in to power poles/boxes all over the place
<ccox>
yep. I've had to use a lot of advanced math in programming work, which of course the users never see...
<JordanBrown>
We don't allow snow here in Los Angeles.
<JordanBrown>
We keep it in the mountains where we can visit it when we want to.
<JordanBrown>
I've done relatively little in the way of geometric and graphical programming, so haven't run into it.
<JakeSays>
ccox: i did at the start of my career, but then i enterprised myself.
<JakeSays>
accidentally, i might add
<JordanBrown>
My limits are transformation matrixes, norms, and some light trig.
<JordanBrown>
Anyhow, thanks!
<ccox>
JakeSays: I stuck with desktop software most of the time, though some of it got turned into libraries for cloud software.
<lf94>
Day 10 - thank god.
<ccox>
Now I'm semi-retired and writing benchmarks, compiler tests, and some open source fun.
<InPhase>
JordanBrown: If you want pairwise multiplication of any matching dimension list structures, this should do it: https://bpa.st/MG4A
<JakeSays>
i did a lot of image processing and handwriting recognition, document scanning, etc. that was so much fun. working with neural networks in the early 90's was interesting.
<ccox>
lf94: oh, good, a drink coaster.
<lf94>
yeah lol
<JordanBrown>
I've been more of a device driver and network protocol person.
<ccox>
Oh, I did NN in the 80s through 90s, all the stuff that ML kiddies are rediscovering today.
<JordanBrown>
And all-around trouble shooter/maker.
<JakeSays>
JordanBrown: i did a lot of that kind of stuff as well. was also a lot of fun.
<InPhase>
Much of my programming career has involved softwaring the tricky math so that other people don't have to think about it.
<JordanBrown>
InPhase: the question originated with a program that I was looking at that did rands(x,y,1)*[1], and I didn't understand why it worked.
<ccox>
InPhase: yep. The problem has been trying to get management to understand that the math (and logic) is not nearly as simple as they think.
<JordanBrown>
I was failing to see that [x] is a one-dimensional array, not a two-dimensional 1x1 array.
<InPhase>
JordanBrown: Ah. :) A strange spelling for index by [0].
<JordanBrown>
And then I didn't understand what it meant to multiply two one-dimensional arrays.
<JordanBrown>
Yes, a very strange spelling for that.
<JordanBrown>
These days my biggest interest (which is not to say that it's exactly what I get paid for) is in how to write programs that are as easy as possible to understand and are bug-resistant.
<JordanBrown>
When I first encountered C, I thought it was magic that you could express strcpy as "while(*d++=*s++);". Now I look at that kind of thing and shudder.
<lf94>
Yeah that's pretty terrible.
<JordanBrown>
I do maybe 40% of my work in JavaScript, and it's a ton of fun to work with because it's very flexible and powerful...
<JordanBrown>
... and I think it's a truly awful language for doing real work in because it's almost immune to static analysis.
<JordanBrown>
Similar comments about Python.
<JordanBrown>
(But it's only maybe 10% of my work.)
<JordanBrown>
These days I enjoy thinking about things like "how would you design this library so that a bug in how it's called is detectable by the compiler?"
<JordanBrown>
and "These two pieces X and Y are required to be consistent. How can I make it so that if they are ever inconsistent the compiler will detect it - or, better still, so that it is impossible for them to be inconsistent."
<JordanBrown>
Speaking of which, I guess I should go back to stuff that I get paid to do...
JordanBrown has quit [Read error: Connection reset by peer]
<JakeSays>
that's no fun
<InPhase>
JordanBrown left before I could applaud his focus.
<lf94>
Modern JavaScript is pretty excellent.
ur5us has joined #openscad
<lf94>
What I really hate is webdev in general.
<InPhase>
The best thing about JavaScript is that I don't have to use it very often.
<InPhase>
Maybe WebAssembly and the 3D support in browsers will change some things for me, but I just in general haven't been able to get too excited about webdev, other than making a few things I need work.
<JordanBrown>
InPhase: Sometimes it isn't clear whether it's focus or navel contemplation.
<JordanBrown>
But when you're working on software that works 99.9% of the time, where the bugs that you expect are the ones that you only see because you've been running for six months, ordinary testing isn't enough.
<JordanBrown>
You want ways to analyze every path through the program, even (and especially!) the ones that your tests don't exercise, the ones that you didn't realize were there, that you can't even figure out how to exercise, and find the bugs that are lurking there.
<JordanBrown>
Not that I work on those analyzers. I'm a decent language lawyer, and I've written a couple of small-to-medium languages, but I'm not a compiler guy.
<JordanBrown>
But I *use* them, and think that as much of a pain in the neck they are to satisfy, the result is worth it.
<JakeSays>
language lawyer?
<JordanBrown>
Somebody who can look at a complex piece of a program and explain why the compiler is doing the apparently-strange thing that it's doing.
<JakeSays>
lawyer would not be the word i'd use for that.
<JordanBrown>
Like why, in C, it's really bad to compare a negative number to an unsigned number.
<JordanBrown>
The negative number gets converted into a very large unsigned number, and the result is never what you expect.
<JordanBrown>
Or, as a somewhat trivial example, given a typical implementation of MAX(), why it's a bad idea to say MAX(func(), 10).
<JordanBrown>
even worse would be MAX(i++, 10).
<JakeSays>
so why is it a bad idea?
<JordanBrown>
because MAX is often defined as a macro
<InPhase>
JordanBrown: I completely agree with the things you've said, and that's basically how I've been focused on software for the last decade-plus.
<JordanBrown>
If you want a disk server with a couple of petabytes of spinning media, a few dozen terabytes of flash cache, and a couple of terabytes of RAM cache
<JordanBrown>
and you have a spare million or two dollars
<JordanBrown>
we will be happy to sell you one.
<JakeSays>
lol i'm sure
<JordanBrown>
(and yes, that's terabytes-with-a-T of RAM.)
<InPhase>
JordanBrown: Well I have a $70k budget and need a new storage system. Does it scale down? :)
<JakeSays>
InPhase: lol
<JakeSays>
it's oracle..
<JordanBrown>
Originally, yes. Today, I don't think we go that low.
<JordanBrown>
I think the bottom end is in the ~$250K range these days.
<JakeSays>
i really wish someone other than oracle had bought sun
<InPhase>
JordanBrown: Well good. I'm still unhappy with Ellison, so I will be happy to look elsewhere. :) But zfs flash-based systems are what I'm investigating.
* JakeSays
has a flashy zfs system
<JordanBrown>
You can certainly build something out of a Solaris system, and there you can go pretty low.
<JakeSays>
well, it used to be until i disabled the hipster rgb leds
<JordanBrown>
Though I fear that you'd have to buy Oracle hardware, and I get the impression that's pretty pricey.
<JakeSays>
zfs on linux is the way to go
<InPhase>
We have the other hardware in place working adequately. We just accumulate too much data and need to expand space and speed for that.
<JakeSays>
by flash do you mean nvme?
<JordanBrown>
I'm not sure.
<JordanBrown>
With respect to our products.
<InPhase>
I'm open to options, just seeking to avoid seek time lags.
<JordanBrown>
Part of our answer to that is to throw massive RAM at caching.
<JakeSays>
i have 2tb nvme ssd's under zfs and damn it's fast
<JordanBrown>
I'm not a performance specialist, but they talk about routinely getting 80-90% cache hits.
LordOfBikes has joined #openscad
<JordanBrown>
My work is generally in the vicinity of the administrative software for the storage appliance, though I'm also involved with Windows interoperability and LDAP.
<JordanBrown>
And whatever fire needs to be put out today.
<JordanBrown>
Like: if you can avoid ever having to deal with Turkish ı and İ, avoid them
<JordanBrown>
Run screaming in the opposite direction.
<dalias>
:)
<JordanBrown>
back to pretending to do useful work...
JordanBrown has quit [Read error: Connection reset by peer]
TheCoffeMaker has quit [Ping timeout: 240 seconds]
<ali1234>
reddit: "trigonometry is hard, just use minkowski"
<lf94>
How do I calculate the normal for the top left,right and bottom left,right sides of this?
<lf94>
InPhase: maybe you can help
<lf94>
I'm trying to make the prism to cut out for the coaster
<lf94>
I just dont know the kind of math I need
<ali1234>
is that even a 3d object?
<ali1234>
it's kinda hard to tell what shape it is
<ali1234>
but if you know the vectors of two edges of a side (or any two non-colinear vectors tangental to the surface) then the normal can be calculated by taking the cross product of those two vectors
<ali1234>
you have to be careful though: the direction of the result flips if you swap the two vectors in the input
<lf94>
yes it's 3d but the 3d doesnt really matter
<ali1234>
oh but it does... a 2d surface can't have a normal (or alternatively, the normal is trivial)
<lf94>
I guess the 2D normal is just... a coordinate?
<lf94>
I need to find the coordinate of the middle of those lines then
* lf94
back to highschool geometry
<ali1234>
the "center" of the polygon in 3d space?
<ali1234>
i've got bad news for you: there's like 100 different definitions of what that means...
<ali1234>
you're trying to remake day 10?
<peepsalot2>
a normal is not a coordinate. mainly because it has no units
Jack211693 has joined #openscad
<ali1234>
can you draw on the screenshot the thing you want to find?
<peepsalot2>
edges have normals in 2D
Jack2116 has quit [Ping timeout: 256 seconds]
TheCoffeMaker has joined #openscad
<InPhase>
lf94: A hexagon's 2D normals are just directionally at 60 degree angles around a circle.
<InPhase>
lf94: If you want the distance from center for the middle of the edges, it's the distance to the vertices divided by cos(30).
<InPhase>
lf94: Does that address it?
<InPhase>
(It's cos(30) because you drop a line down and get a right triangle that's half of those 60 degrees, so (360/6)/2.)
<lf94>
Yes, that should do it!
<lf94>
I thought it was 45 degress
<lf94>
So what's the normal in terms of 0-1?
<InPhase>
That would be a very octagonal hexagon. :)
<InPhase>
That's not a thing. Want to try that question a different way?
<lf94>
Hm
<lf94>
In Curv, there is this shape called a "half_plane"
<lf94>
This half_plane is what it sounds like: it goes off into infinite from the middle of the cartesian plane
<lf94>
You specify the distance from the center, and the normal
<lf94>
the normal is the direction the plane will be going off into
<InPhase>
Sure.
<InPhase>
So specified as a vector?
<lf94>
Yeah
<ali1234>
i keep asking for that in openscad lol
<lf94>
What would 60*deg as a vector be?
<InPhase>
[cos(a), sin(a)] for a in steps of 60 degrees.
<lf94>
ah, so cis
<lf94>
(cis (60*deg))
<lf94>
nice
<ali1234>
but this won't get you the top and bottom of the prism, only the sides
<lf94>
Well top and bottom is easy though :)
<lf94>
I got those already X)
TheCoffeMaker has quit [Ping timeout: 252 seconds]
<lf94>
Well
<lf94>
Almost got it
<lf94>
The plane is at a weird angle
<InPhase>
ali1234: So we can't create a planar segmentation geometry, but I bet we could make a PlanarCut built-in.
<lf94>
InPhase: could you download curv? :D
<InPhase>
ali1234: That would by itself have a ton of uses.
<lf94>
oh I'm a fool
<lf94>
It should be 30 deg
<lf94>
Awww yeaaaaa!
<InPhase>
lf94: Oddly enough it seems I downloaded it October of 2019. I don't remember doing this.
TheCoffeMaker has joined #openscad
<InPhase>
lf94: I suppose it changed a bit since then. :)
<lf94>
Yeah. Best is to build master currently.
<InPhase>
That was sure a list of dependencies. It's building.
<InPhase>
fatal error: tbb/parallel_for.h: No such file or directory
<InPhase>
Is that a secret dependency missing from cmake checks? Or is that an internal thing which is broken?
<InPhase>
I see the build is marked as failing right now on github, but that happens for all sorts of reasons.
<InPhase>
Also, side note, there should really be some sort of tool to apt install all the dependencies cmake is going to look for. It's silly to spend 10 minutes of time doing something a computer should be doing for me.
<lf94>
If you're using debian like me, it should work no problem
<InPhase>
This project uses `make` as a wrapper around `cmake`. What what what.
<ali1234>
that's bananas
<InPhase>
You make a shell script to do that sort of thing. It's too backwards. :)
<lf94>
You shouldnt run make at all
<lf94>
on debian, it says to create a build/ directory
<lf94>
ard then run cmake from it
<lf94>
I agree the build instructions need some updating
<lf94>
But they are accurate
<lf94>
updating -> better user experience
<InPhase>
This thing is totally failing. Maybe the weird make approach can't handle -j8? I'm trying a non-parallel build.
<InPhase>
It spiked to 2.6GB of RAM usage in the compiler, then launched an assembler for some reason, then failed with no clear error message. But I'm not sure if that was related because of the parallel.
<InPhase>
There's a generation of people raised to like those approaches, but I generally find them like wading through molasses.
<ali1234>
you can clean up the mess by grouping nodes together into sub-units that have inputs and outputs
<ali1234>
it has strengths and weaknesses. it's definitely going to be easier for artistic types who don't know the first thing about functional programming...
<ali1234>
i wouldn't use this for functional prints though
<ali1234>
it isn't really meant for complex CSG modelling, it's designed for randomly scattering rocks on your 3d height field terrain
<ali1234>
you can also split up geometry by selecting a subset of points, and then transform only those points
<lf94>
InPhase: it's all good, I figured it out
<lf94>
InPhase: I'll push today's code if you want to play with it :)
<JakeSays>
ali1234: but for 3d printing blender seems like overkill, i think
<ali1234>
so you can do things like select all the point in the top half of a sphere and translate them all up to create a pill shape
<ali1234>
in openscad you'd use hull for that. blender does not have hull
<ali1234>
oh wait, actually it does have hull in 3
<ali1234>
also bounding box :)
<ali1234>
yeah i would not use blender for 3d printing unless i was making a purely decorative item like a vase or something, and even then i'd just model it as a normal mesh, not geo nodes
<ali1234>
unless it was a vase with random placed lumps on it, i guess
<ali1234>
blender's CSG is pretty fast but it has no separation of preview/render so actually it is really slow
<JakeSays>
and it uses python.
<ali1234>
by which i mean it generates meshes much faster than openscad's render, but much slower than openscad's preview. probably because it is not trying to be "always correct"
af has quit [Ping timeout: 260 seconds]
af has joined #openscad
<ali1234>
(and therefore the results probably aren't suitable for slicing anyway)
<InPhase>
lf94: I think the closest shape to Day 9's 4-pointed version is the Meissner Tetrahedron. But, it's not quite a Meissner Tetrahedron. It might have the same unintuitive constant width property though. It looks pretty close to it.
<lf94>
jack21 mentioned another
<InPhase>
Oh, no, there's a side orientation where it loses the constant width.
jkent3 has joined #openscad
jkent3 has joined #openscad
<InPhase>
"it is kind of a rounded disphenoid" -- Jack21
<othx>
InPhase linked to YouTube video "Meissner Tetrahedron Demo" => 1 IRC mentions
<InPhase>
Not quite either of those, but it might not have a formal name.
<InPhase>
Let's call it the Ulrich Tetrahedron until someone finds out who discovered it first in 500 B.C.
jkent3 has quit [Ping timeout: 268 seconds]
<InPhase>
lf94: Here: I back-calculated the functional form from the OpenSCAD code. This plots it in wxmaxima: wxplot3d([sin(asin(x)+acos(y)),sin(asin(x)-acos(y)), [x,-1,1], [y,-1,1]])
<InPhase>
sin(asin(x)+acos(y)) does the top, and sin(asin(x)-acos(y)) does the bottom.
<InPhase>
That should be enough to make an sdf, right? :)
<lf94>
I tried that but yes right track
GNUmoon has quit [Ping timeout: 276 seconds]
jkent3 has joined #openscad
jkent3 has quit [Ping timeout: 252 seconds]
GNUmoon has joined #openscad
snaked has quit [Ping timeout: 240 seconds]
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openscad
Jack2116938 is now known as Jack21
<Jack21>
JakeSays hm everyday something new .. this looks odd not sure if it is the filament though - you see delamination which is quite rare with PETG
<Jack21>
I could imagine that the extruder temp dropped significantly - did you use octoprint or have a log about temp?
<JakeSays>
hmm. didn't think about temp fluctuation
<JakeSays>
i do use octoprint
snaked has joined #openscad
<JakeSays>
Jack21: the sides of the box, even in areas where the print looks good, are super brittle.
<JakeSays>
almost like there's no layer adhesion at all
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openscad
<Jack21>
if it is not temp .. maybe something on the filament - do you use a wiper ..
_whitelogger has joined #openscad
<Jack21>
As they extrude Pellets it is unlikely that the chemical composition changed for just a few meter
<JakeSays>
Jack21: no wiper
<JakeSays>
i had the temp set at 230c
<peeps>
if your volumetric extrusion rate (extruder speed * nozzle area) is high enough, the nozzle heating may not be able to keep up
<Jack21>
what peeps said the heater cartridges often run near max and if they are installed without thermal compound and also the sensor there could be quite some undetected variations
<JakeSays>
you know, the state of 3d printing hardware is pathetic.
<Jack21>
it is tech from the 80's
<ccox>
yep, unless you get to the half million dollar industrial stuff
<JakeSays>
ccox: exactly
<ccox>
Sadly, the software (Cura, etc.) is even worse.
<JakeSays>
i'm about a half mill short
<JakeSays>
ccox: oh it so is!!
<JakeSays>
i would love to write a slicer
<ccox>
don't write a slicer. Just take a good slicer and write a @#%@!#%@#%$ stable and usable UI around it.
<othx>
peeps linked to YouTube video "Does This make Volcano Hotends Obsolete? Bondtech CHT Review" => 1 IRC mentions
<JakeSays>
well, the fun part is the math.lol
<JakeSays>
ccox: yeah that's where i would start - with the UI
<Jack21>
ccox i don't think there are good slicer - afaik none of them calculates the actual area of the wall - only the path
<ccox>
Jake - the fun part of software is usually less than 10% of the software. The hard part (UI, localization, stability) is what takes 90% of the time and effort.
<JakeSays>
i'm only aware of two slicing engines: cura's and slicr
<JakeSays>
ccox: oh of course.
<ccox>
There are more (even Photoshop has it's own slicing engine).
<JakeSays>
interesting
<JakeSays>
what other ones can you point me to?
<JakeSays>
open source
<JakeSays>
i am aware of simplify3d
<JakeSays>
talk about a BAD ui
<ccox>
Open source is a bit more limited in choices....
<Jack21>
peeps i liked the part where he hammered cooper wire into a 1$ nozzle and get similar results
<JakeSays>
close source and cheap works too
<ccox>
SLic3r and Cura are the biggies, with forks for Prusa and other groups.
<JakeSays>
prusa's slicer's ui tolerable
<JakeSays>
*is
<JakeSays>
ccox: what does photoshop use a slicer for?
<ccox>
The 3D features and driving printers. Most under-advertised features EVER.
<ccox>
Photoshop had more accurate slicing and tree supports long before Cura. BUt their product manager is focused on web design, and thinks all users are web designers, so....
<JakeSays>
oh wow. i had no idea
<ccox>
They're taking those features out of Photoshop in an upcoming release... because: bad management.
<JakeSays>
huge missed opportunity
<ccox>
yep.
<JakeSays>
well, it probably doesn't run on linux anyway
<ccox>
especially since it was so great when texturing and finishing models, then going straight to printers (or services) from PS.
<ccox>
True, Linux never stabilized the tool chains enough for PS.
<ccox>
Though the PS library does run on Linux for cloud services.
<JakeSays>
huh. just found slicer.org
<ccox>
That is not a 3D printing slicer. That is a voxel slicer for medical imaging. (which I've used for CT and MRI simulations)
<ccox>
The 3d printing export is... eh, not the best.
<JakeSays>
yeah just realized that
<ccox>
Though the voxel export is great, if you're using a voxeljet type printer.
<JakeSays>
hmm. somewhere i have some dental xray DICOM images
<JakeSays>
icesl looks interesting but it has a really bad ui
<ccox>
also does not appear to be open source
<JakeSays>
yeah
<JakeSays>
mattercontrol is another one. it's written in c# it looks like
<JakeSays>
huh. craftware pro is interesting
<JakeSays>
even the firmware is crap
<ccox>
Have you used SLA printers? Crap firmware is pretty standard.
<Jack21>
JakeSays in case you wanna use a wiper (from what i find on it i would recommend it - and my filament is sealed and in a PTFE tube for most of the path) https://www.prusaprinters.org/prints/77292
<Scopeuk>
I haven't used industrial machines all that much but if they are any thing like high end niche software tools they cost a fortune and are still awful
<Jack21>
When i was on a printing fair - they had printer 2×3×2meter with a print area of 30cm and they didn't print any better
<Jack21>
they are just more reliable and need less attention - use pellets
<Scopeuk>
generally what you are buying with an "industrial" machine is repeatability rather than absolute capeability
<Scopeuk>
that and the ability to survive a "workshop" environment with duty cycles and less than cooperative environmental conditions
<Scopeuk>
someone once observed that there are people who can reliably and repeatedly get the same or better results out of their own machine but its being done by knowing that machine exceptionally well. where the industrial machines will do the same across a full product line
<Scopeuk>
I think that's probably a fair assessment, the "domestic" machines can produce great results but a lot of that is dependent upon the user and tuning on the industrials the manufacturer takes responsibility for that instead
jkent3 has joined #openscad
peepsalot2 has quit [Read error: Connection reset by peer]
jkent3 has quit [Ping timeout: 256 seconds]
peepsalot has joined #openscad
<Jack21>
some are also much faster or can generate parallel - exotic parts or full color prototyping
<JakeSays>
Jack21: what's the wiper do? wipe the fillament?
<JakeSays>
ccox: not yet. only FDM
<ccox>
FDM at least has open source firmware on many machines. SLA: pretty much all closed.
<JakeSays>
my definition of "print better" is unbox printer, plug in, connect to computer, load filament, slice, print.
<ccox>
There are a couple efforts to write open source SLA firmware (and hardware) - but they got stalled by COVID.
<JakeSays>
sla seems like it'd be easier than fdm firmware
<Jack21>
JakeSays yes as dust accumulates because electrostatic
<ccox>
Eh, both have their complexities. MSLA (LCD) is easiest. Galvo+Laser SLA is most difficult (worse than FDM).
<JakeSays>
ccox: yeah i was thinking lcd
<JakeSays>
Jack21: ah ok
<JakeSays>
Jack21: that wiper is pretty cool
<JakeSays>
ugh. i have got to fix the bug in my kvm sometime. it's driving me nuts.
<Jack21>
thx
mhroncok has joined #openscad
<JakeSays>
i regret buying this ender 3
<teepee>
why? supposedly they are not that bad?
<JakeSays>
i can't get it dialed in. the z axis offset changes constantly
<Scopeuk>
JakeSays is the z end stop definitely tight? I've seen that as an issue more than once (on various machines not specifically and ender 3)
<JakeSays>
Scopeuk: you mean the limit switch?
<Scopeuk>
yes
<JakeSays>
it's not used. i have a bltouch sensor connected
<Scopeuk>
I suppose the same would still apply, the bl touch is setting z = 0; inconsistent offset seams likely to be a mechanical issue. so I guess double check carriage is moving freely near z = 0 (no binding on the v slot wheels) and that the bl touch is secure to the carriage
<Scopeuk>
failing that there are a lot of ender 3's between the patisipants of #reprap
<teepee>
that's odd, especially the Z axis should be most easy to control
<teepee>
sometimes it's easy mechanical problems, like I had ages ago on the good old Ultimaker
<JakeSays>
everything is nice and snug. it homes fine
<teepee>
at some point the Z axis set screw was not tight causing Z issues
<Scopeuk>
ok lets clarify a term here, by homes fine its going to 0,0,0 (printer defined) but z is hovering above the bed or pressing into it? by varying amounts each time?
<teepee>
ah, the ender3 uses roller skates on extrusion for Z?
<Scopeuk>
I belive so
<teepee>
right, I have no experience with that :)
<JakeSays>
i'll get everything aligned and a print going. next print i do it prints 10mm off the bed
<Jack21>
that is absurd
<Scopeuk>
10 mm is huge, that sounds almost like electrical noise triggering the controler to think z is at 0 before the bl touch has registered
<JakeSays>
Scopeuk: one time it's hovering. i adjust for that, then it touches the bed
<Jack21>
maybe discon the bltouch and use the endstop switch
<Jack21>
just to find the root cause
jkent3 has joined #openscad
<Scopeuk>
I suppose does it go down to 0 and then back off to 10mm is the other question?
<Scopeuk>
i.e. is this potentially some offset setting somewhere or is it that the sensor is returning "hit bed" when it hasnt
<Jack21>
print from SD ..
jkent3 has quit [Ping timeout: 252 seconds]
ccox_ has joined #openscad
ochafik has joined #openscad
ccox has quit [Ping timeout: 265 seconds]
jkent3 has joined #openscad
jkent3 has quit [Ping timeout: 268 seconds]
Jack21 has quit [Quit: Client closed]
Jack21 has joined #openscad
jkent3 has joined #openscad
jkent3 has quit [Client Quit]
jkent3 has joined #openscad
jkent3 has quit [Client Quit]
emxn99 has joined #openscad
<emxn99>
guys can someone give me a hard bulding. wich takes time. i need it really quick for school. and im very bad at openscad plz
<emxn99>
a creation wich can give me a very good note
<emxn99>
just copy and paste it here
<Alicia>
Practice and get better. Don't ask others to do your homework
<emxn99>
i only have 2 days . im gonna learn it but for now i need it
<Alicia>
you better get started then
<emxn99>
i have another project and for that one im gonna learn openscad very good
<emxn99>
im looking at the cheatsheet ad trying to get better
emxn99 has quit [Quit: Client closed]
<teepee>
o_O
<Jack21>
and we have the advent calendar .. stupid - it is not that there are not enough great examples but if he is not able to use internet ..
<Jack21>
maybe i should frame that "takes time ⇒ need it really quick / im very bad ⇒ give me a good note"
* InPhase
chuckles.
ochafik has quit [Remote host closed the connection]
ochafik has joined #openscad
ochafik has quit [Ping timeout: 240 seconds]
PovilasCNC has quit [Read error: Connection reset by peer]
<Jack21>
InPhase after you gave the Idea .. may i present the MONO or FLOP or FLIP https://imgur.com/a/9Ry9WHG
<InPhase>
Jack21: Do you want to get bees? Because this is how you get bees.
<Jack21>
off label use as insect hotel?
Jeringa has joined #openscad
<Jeringa>
hola a todos
ochafik has joined #openscad
Jeringa has quit [Client Quit]
ochafik has quit [Ping timeout: 268 seconds]
ochafik has joined #openscad
ochafik has quit [Ping timeout: 268 seconds]
ochafik has joined #openscad
ochafik has quit [Ping timeout: 268 seconds]
ochafik has joined #openscad
ochafik has quit [Ping timeout: 252 seconds]
<Jack21>
ahhh 20h print for one MONO
ochafik has joined #openscad
mhroncok has quit [Quit: Leaving.]
ochafik has quit [Ping timeout: 250 seconds]
<lf94>
mono?
<lf94>
InPhase: what was that functional form you posted again
<InPhase>
"<InPhase> lf94: Here: I back-calculated the functional form from the OpenSCAD code. This plots it in wxmaxima: wxplot3d([sin(asin(x)+acos(y)),sin(asin(x)-acos(y)), [x,-1,1], [y,-1,1]]) <InPhase> sin(asin(x)+acos(y)) does the top, and sin(asin(x)-acos(y)) does the bottom."
<lf94>
Perfect :)
<lf94>
Why are there 4 params?
<InPhase>
What 4?
<InPhase>
x and y.
<InPhase>
Yielding z.
<InPhase>
asin means arcsin
<InPhase>
Or sin^-1 depending on how you learned it. :)
<InPhase>
wxmaxima doesn't by default preserve the aspect ratios in its plotting because that's not what it's designed for, but: https://i.imgur.com/CY2k5tj.png
Jack21 has quit [Quit: Ping timeout (120 seconds)]
Jack21 has joined #openscad
<Jack21>
lf94 MONO the forefoot isometric flipflop i made with double honeycomb load distributing structure
<InPhase>
lf94: Oh, that's just the plot function's syntax. It's taking a list of functions to plot and ranges to apply to them as bounds.
<lf94>
ooooh, ok.
<InPhase>
lf94: It does some type sniffing I think.
<InPhase>
That was the first time I ever plotted 2 3D functions at once, and the error message told me to use that syntax when I got it wrong the first time by using the syntax for 2D plots. (Yay inconsistencies of interface.)
<lf94>
InPhase: what's the range for the first value?