<J1A8413>
and how would that work.. you need to give to points you like to measure .. and norm(b-a) will give you the length
<JordanBrown[m]>
Theoretically it would be nice to be able to click on vertexes, faces, edges, et cetera, and draw measuring lines between them. But that is technically quite difficult.
pie_ has quit []
<J1A8413>
click? .. you mean using the mouse like in conventional CAD software?
<JordanBrown[m]>
yes
<tcurdt>
yeah, I guess right now there is not much of that interaction part
<JordanBrown[m]>
Here's what my cylinder-with-black-marks measuring stick looks like:
<J1A8413>
why would i use scad then? when i need the mouse
<JordanBrown[m]>
How do you currently rotate the model to look at all sides?
<J1A8413>
if you design your things proper you just use the variables and some math to echo your measurements
<tcurdt>
well, you need a mouse to position the view anyway
<JordanBrown[m]>
Mostly I agree that if you do the math right then things end up where they are supposed to be.
<JordanBrown[m]>
But sometimes I am using the math to get something to match up with the real world, and measuring lets me confirm that I got the math right.
<J1A8413>
ctrl 4-9
<JordanBrown[m]>
You only use the primary six views?
<tcurdt>
that wouldn't work for me
<Scopeuk>
peepsalot it looks like you made some big leaps with the windows ci stuff congratulations
<JordanBrown[m]>
I mean, I suppose, to each his own, but while I want the *model* to be expressed in text, I am happy to use the mouse to *look* at it, and I would consider at least some cases of measurement to be looking.
<J1A8413>
it is all you need for design in orthographic view .. for a nice picture you can rotate a bit
<tcurdt>
what JordanBrown[m] says
<JordanBrown[m]>
For instance, here's one of my models that I think I ended up using a ruler on:
<tcurdt>
if the part gets more complicated the standard views are not enough for me 🤷♂️
<othx>
JordanBrown[m] linked to "My House, with furniture by jordanbrown" on thingiverse => 2 IRC mentions
<peepsalot>
Scopeuk: yes, well the action you linked to worked perfectly, since it defaults to 22.1.7.2 But trying to make it download and install the latest mesa 22.2.2 instead was not as successful
<JordanBrown[m]>
That's not the simplest piece of furniture, but it's not the fanciest either.
<JordanBrown[m]>
It's all in 1mm=1in scale and represents real furniture, so the dimensions matter.
<tcurdt>
all in openscad? crazy ;)
<JordanBrown[m]>
There are, I think, two objects that are 3D scanned.
<JordanBrown[m]>
Though I don't think one of them is in that Thingiverse post.
<JordanBrown[m]>
The elephants are scanned. The chair, however, is modelled.
<peepsalot>
Scopeuk: are you following the issue on the msys2 packages repo? or just looking at the results of the various CI runs?
<Scopeuk>
Both
<Scopeuk>
At the very least you've proved a viable alternative setup
<Scopeuk>
I wonder if there is a reason that action hasn't been updated to a new default
<Scopeuk>
Looks like the ci reboot just hit the same behaviour as earlier
<Scopeuk>
S/reboot/retry
<peepsalot>
yeah it is frozen on the exact same test even: cgalpngtest_child-modifier
<peepsalot>
was just making sure it wasn't a one time fluke of CI server being weird.
<Scopeuk>
I guess this is equally narrowing rapidly to a reproduceable test case
<Scopeuk>
Yeh that's always a good idea particularly going upstream
<InPhase>
JordanBrown[m]: Wow. That set of models of your house is both way over the top, and incredibly well done. You did very intricate detail there.
<JordanBrown[m]>
Thanks.
<InPhase>
JordanBrown[m]: Also I note it was posted at the start of lockdowns. :)
<JordanBrown[m]>
Oh, it was in the works for a long time before that.
<InPhase>
Ok. :)
<JordanBrown[m]>
People occasionally say "you should do this for other people, for money" and I say "they couldn't afford it".
<JordanBrown[m]>
There's hundreds of hours of work there.
<JordanBrown[m]>
I've gotten better at it, over the course of it, but my standards have also gone up.
<JordanBrown[m]>
And we just did a remodel to the master bath and laundry room, and so I need to redo that area...
<InPhase>
Yeah, it's clear that would be ridiculously expensive to have done.
<JordanBrown[m]>
And yes, there's a lot of detail, really far more than is necessary. I realized at one point that one particular subcomponent that I'd spent some time on would print at around one cubic millimeter.
<InPhase>
:)
<InPhase>
Also, while you've probably moved on to a new belt by now, I bet a TPU version of that belt loop would be invincible.
<JordanBrown[m]>
There's a cap on top of a half-wall. It's an overhang, so I designed in a 45° support under it. That's not so silly - it's several inches long, and probably 1-2 millimeters across.
<JordanBrown[m]>
But the itty-bitty piece at a corner, where two of the 45° supports meet... that's silly small.
<JordanBrown[m]>
I do still use that belt, and have to replace the belt loop every year or so.
<JordanBrown[m]>
I haven't tried making a TPU version, partially because the only TPU I have is white and it's a black belt.
<JordanBrown[m]>
But also I don't have a lot of experience with TPU and so am not very comfortable with it.
<InPhase>
JordanBrown[m]: Well, my standard metric for explaining TPU is that a properly printed 1mm thick sheet of TPU is not destroyable by human hands without tools.
<JordanBrown[m]>
OK.
<InPhase>
If someone wants to claim otherwise, I want to see an unedited video of it. ;)
<InPhase>
Because that stuff is tough.
<JordanBrown[m]>
Though I'm not completely sure that I could destroy a 1mm slab of PLA with my hands either.
<InPhase>
Oh if you gave it enough flex, it would snap.
<JordanBrown[m]>
I suppose it would depend on how big it is, and so how much of a lever arm I could get.
<InPhase>
But you can't snap TPU. And you can't tear it.
<JordanBrown[m]>
In my very few experiments with it, I found that I had to print it very slowly.
J1A8413 has quit [Ping timeout: 260 seconds]
<JordanBrown[m]>
And I needed a special adapter stuck in the filament path to keep it in line between the feeder gear and the melting chamber.
<InPhase>
Bowden extruder?
<JordanBrown[m]>
no, direct
<InPhase>
Oh, weird.
<InPhase>
I always just stick it in and print.
<JordanBrown[m]>
but it has a maybe 3mm gap between the feeder gears and where it goes into the hot end.
<JordanBrown[m]>
and that's enough for it to pop out to the side.
<InPhase>
Initial feed is sometimes tricky, but I make sure to start with a straight tip, and I have a sort of pull and rotate trick going for me.
<InPhase>
That tube at the top takes away your control when inserting.
<InPhase>
Mine is pretty structurally similar, but I'm putting it right into the hole at the top, and using the manual tactile feedback to put it in.
<InPhase>
I suppose it took me a few tries to get a feel for that, but now I do it without even thinking.
<JordanBrown[m]>
No, I'll get it all the way down into the hot end and extruding, but then after a while it'll pop out to the side.
<InPhase>
Does it melt up too far?
<InPhase>
Or it bends out?
<JordanBrown[m]>
Well, yes, it bends out to the side below the feeder gear.
<InPhase>
Hmm, ok.
<JordanBrown[m]>
Whether that's because it gets too hot and soft there... don't know.
<InPhase>
I had never heard of that happening at that location, but I guess it's indeed possible.
<JordanBrown[m]>
The vendor designed a little adapter that pops into that gap and keeps the TPU on the straight and narrow.
<InPhase>
I wonder if that spacing gap is larger than mine...
<InPhase>
Ah, yeah. The insert assembly has a 3D printed guide just like that metal piece. You can see it right here on this assembly video: https://youtu.be/O5S8lFf15UY?t=950
<othx>
InPhase linked to YouTube video "HICTOP 3DP-11-ATL 3D Printer Assembly Part 3 - X & Z Axis final build" => 1 IRC mentions
<InPhase>
And yes, very old printer. I kind of figured it would break by now so I could justify replacing it, but after a few early issues and upgrade-repairs of those, it kind of stabilized.
<JordanBrown[m]>
Mine is a bought-fully-assembled Craftbot Plus.
<JordanBrown[m]>
I think I bought it in 2015, so it's a similar age.
<InPhase>
Well the current versions of those at least look pretty nice.
<JordanBrown[m]>
I like the fully enclosed design, gives me warm fuzzy feelings about moving parts.
<InPhase>
Yeah, mine looks like it should by all rights fall apart, with all the wires hanging all over the place. I printed an enclosure for some of them, which snapped open and is half hanging enclosed... I didn't bother to fix that.
<JordanBrown[m]>
Also I like that it doesn't try to move the bed in XY... I can't understand how you can get decent speed if you're moving the bed.
<InPhase>
Well it just moves the bed very fast.
<JordanBrown[m]>
Yeah, but beds are heavy and inertia.
<JordanBrown[m]>
it's bad enough slamming the print head around.
<InPhase>
Yes, for sure.
<JordanBrown[m]>
I've never tried a Bowden - actually, never tried anything other than this one - but the idea of having the head be lighter is appealing.
<JordanBrown[m]>
You know, physics :-)
<JordanBrown[m]>
But I don't think I'd ever thought about jerk before seeing it in 3DP parameters.
<InPhase>
My bed is attached quite robustly though. I upgraded the mount for the bed belt to help with this, and it has some pretty nicely placed guide rods. But certainly there's no way around the theory that a stationary bed will be better, if you provide the same level of support to the print head.
Junxter has joined #openscad
Guest21 has joined #openscad
<JordanBrown[m]>
I'm feeling like I'm cheating, scanning a wrench and then tracing it using a drawing tool.
<JordanBrown[m]>
Hmm. I just realized the screwdrivers didn't nest the way I expected them to; I'd have to make the whole assembly longer to let them do that.
<JordanBrown[m]>
But I'll print that and see how it seems to work.
<JordanBrown[m]>
It'll take all night, sigh.
<peepsalot>
3d printed tool holder? aren't those usually cut out of foam?
<JordanBrown[m]>
Got PLA. Not got foam.
<JordanBrown[m]>
:-)
<peepsalot>
i want a cnc hot wire machine
<JordanBrown[m]>
Organizer for a tool drawer.
<JordanBrown[m]>
If the foam is rigid you could cut it with a mill.
<JordanBrown[m]>
And then you could get 2.5D, not just 2D.
<JordanBrown[m]>
Right now I want to know why OpenSCAD happily imports one of my SVG files, but not the other.
<JordanBrown[m]>
Maybe it has something to do with one of them being 1.9M and the other being 25.8M.
<JordanBrown[m]>
I used a different scanning tool for the original scans...
__xxoxx has joined #openscad
_xxoxx has quit [Ping timeout: 260 seconds]
tristero has quit [*.net *.split]
gknux has quit [*.net *.split]
gknux has joined #openscad
tristero has joined #openscad
one-star-chef has quit [*.net *.split]
Alexer has quit [*.net *.split]
Alexer has joined #openscad
one-star-chef has joined #openscad
metasean[m] has quit [*.net *.split]
lf94 has quit [*.net *.split]
gwillen has quit [*.net *.split]
niyawe has quit [*.net *.split]
lf94 has joined #openscad
gwillen has joined #openscad
niyawe has joined #openscad
gwillen has quit [Signing in (gwillen)]
gwillen has joined #openscad
metasean[m] has joined #openscad
linext_ has quit [*.net *.split]
abff has quit [*.net *.split]
clemens3 has quit [*.net *.split]
raboof has quit [*.net *.split]
Joel has quit [*.net *.split]
mlaga97_ has quit [*.net *.split]
buZz has quit [*.net *.split]
dalias has quit [*.net *.split]
ecraven has quit [*.net *.split]
dTal has quit [*.net *.split]
Wolf480pl has quit [*.net *.split]
buZz has joined #openscad
dalias has joined #openscad
dTal has joined #openscad
buZz is now known as Guest2636
abff has joined #openscad
raboof has joined #openscad
linext_ has joined #openscad
abff has quit [Changing host]
abff has joined #openscad
clemens3 has joined #openscad
ecraven has joined #openscad
Wolf480pl has joined #openscad
Joel has joined #openscad
fling has quit [Ping timeout: 255 seconds]
Sauvin has quit [Ping timeout: 272 seconds]
Bocaneri has joined #openscad
Bocaneri is now known as Guest8611
rumgzy has joined #openscad
Guest8611 has quit [Ping timeout: 260 seconds]
fling has joined #openscad
snaked has quit [Remote host closed the connection]
snaked has joined #openscad
tristero has quit [Ping timeout: 260 seconds]
tristero has joined #openscad
snaked has quit [Ping timeout: 260 seconds]
abff has quit [Ping timeout: 260 seconds]
Wolf481pl has joined #openscad
abff has joined #openscad
JoelJoel has joined #openscad
Wolf480pl has quit [Ping timeout: 260 seconds]
Joel has quit [Ping timeout: 260 seconds]
ur5us has joined #openscad
Guest2636 is now known as buZz
tristero has quit [Ping timeout: 260 seconds]
tristero has joined #openscad
fling has quit [Remote host closed the connection]
fling has joined #openscad
__xxoxx has quit [Quit: Leaving]
ur5us has quit [Ping timeout: 248 seconds]
castaway has joined #openscad
fling has quit [Ping timeout: 255 seconds]
fling has joined #openscad
fling has quit [Ping timeout: 255 seconds]
fling has joined #openscad
Guest246 has quit [Quit: Client closed]
castaway has quit [Ping timeout: 260 seconds]
castaway has joined #openscad
TheAssassin has joined #openscad
TheAssassin has quit [Remote host closed the connection]
TheAssassin has joined #openscad
snaked has joined #openscad
<Scopeuk>
peepsalot I think I would like one too, shame the hot wire is the easy part, I guess it would be possible to make an xy bed with the usual 3d printer bits, anchor the work piece to it to then manipulate it through the hot wire
<Scopeuk>
I suppose some sort of 2d drive using Omni wheels could also work.
* Scopeuk
is good on ideas but VERY weak on execution
linext_ has quit [Quit: Leaving]
linext has joined #openscad
rumgzy is now known as Sauvin
teepee has joined #openscad
<J1A841379>
just add a arm parallel to x with a hotwire .. (on reprap style printer).. or replace the hotend with a hotblade
<othx>
J1A841379 linked to YouTube video "CNC Thermocol Cutting - Multiple Hot Knives" => 1 IRC mentions
<Scopeuk>
It would probably be easy ish to move the rails on a bed slinger out to the side of the work area on your clamping (work piece clamped where bed would sit) but it would restrict the sort of shapes you could cut without compromising the workpiece such that your clamping fails
<Scopeuk>
Any rigid hot wire, I was envisaging the standard home brew nichromium wide in a bow frame for tension
fling has quit [Remote host closed the connection]
fling has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
linext_ has joined #openscad
linext has quit [Ping timeout: 260 seconds]
castaway has quit [Ping timeout: 252 seconds]
drkow has quit [Ping timeout: 248 seconds]
teepee_ has joined #openscad
teepee has quit [Ping timeout: 255 seconds]
teepee_ is now known as teepee
drfff has joined #openscad
Guest26 has joined #openscad
Guest26 has quit [Client Quit]
rvt has quit [Ping timeout: 272 seconds]
rvt has joined #openscad
drfff has quit [Ping timeout: 260 seconds]
<peepsalot>
teepee: I moved the msys package listing into the yaml on the PR. i think it is only called in that one spot. but do we want to keep msys2_install_dependencies.sh functioning for end-users to build through msys2?
<peepsalot>
I think I can easily move the package listing back over. or just delete the script
<peepsalot>
for now though, the msys2_install_dependencies script won't work for end users, without using that separate github action for installing mesa
<peepsalot>
not sure how long it will take to resolve the issue on msys2 side. so I was planning to just use that giuthub action which is able to successfully install mesa 22.1.7
<teepee>
not working means it's not able to get headless openglß
<teepee>
?
castaway has joined #openscad
<Friithian>
I love makerbot cloud print. I print something and I can't tell if it printed or not because the camera sometimes just doesn't update
<peepsalot>
that was using an alternative method of installing mesa, it doesn't go through msys2
<peepsalot>
then I tried using that alternative method to install the latest mesa and it had problems once again. so I plan to roll back to the version that worked...
<peepsalot>
I'm just wondering how much we care about maintaining msys_install_dependencies.sh specifically for non-CI usage (random end-users wishing to do msys2-based builds)
castaway has quit [Ping timeout: 260 seconds]
<teepee>
IIRC it's mentioned in the build instructions in the manual
<teepee>
if it's just about the mesa topic, I think we would not need to bother much
J1A84137975 has joined #openscad
Trieste has joined #openscad
drfff has joined #openscad
J1A841379 has quit [Ping timeout: 260 seconds]
<Scopeuk>
I think it's two questions, do we want to keep the script in some form, yes we do. Does it matter if we drop Mesa from it, probably not that reall yis a limited environment work arround most people won't hit outside obscure server hardware, looking at you super micro, or ci environments where they would just use our workflow
ur5us has joined #openscad
ali1234 has quit [Remote host closed the connection]
ali1234 has joined #openscad
ali1234 has quit [Remote host closed the connection]
ali1234 has joined #openscad
ali1234 has quit [Remote host closed the connection]
ali1234 has joined #openscad
Trieste has quit [Ping timeout: 260 seconds]
Trieste has joined #openscad
<linext_>
is it possible in OpenSCAD to output a file that includes color and opacity data?
<linext_>
is there a way yet to display .CSG in the web browser?
<peepsalot>
.csg is not a lot differrent from .scad, it basically has all loops unrolled and expressions evaluated
<peepsalot>
so, maybe one of the WASM versions being developed could display it
<peepsalot>
i haven't looked at them much
<peepsalot>
teepee: do you know the reasoning for installing single pacman package at a time in the script?
<peepsalot>
I don't really understand all the steps its doing, but when it had all the packages in the yaml, basically installing all at once, it seemed to go 3-4 minutes faster
<peepsalot>
i wonder if there was some reason for it, which is no longer relevant
<JordanBrown[m]>
Git blames me, but all I did was reformat and maybe update a little.
<JordanBrown[m]>
For my use - that random developer you mentioned - small differences in performance don't matter much, because I run it once and then use it to build for weeks or months.
<JordanBrown[m]>
It's been a bunch of separate invocations since teepee introduced it in 2018.
<JordanBrown[m]>
ISTM that it would be fine to make it list all of the packages on one invocation.
<peepsalot>
JordanBrown[m]: well, i hope you haven't run it since late october ;)
<JordanBrown[m]>
I ran it a few days ago.
<peepsalot>
oh, right i guess as long as you're not using mesa for graphics, that maybe its fine.
<JordanBrown[m]>
Looks like Nov 1st.
<peepsalot>
oh, i'm guessing now that the individual install is mainly for purpose of timestamping? probably not that important to keep that level of granularity, unless msys2 package installs were timing out or something
<peepsalot>
JordanBrown[m]: the time save would be mostly for shortening CI build time for every PR update.
<JordanBrown[m]>
Sure. Like I said, for *my* use case...
<JordanBrown[m]>
I don't know enough about how the CI builds work to have a real opinion there.
<JordanBrown[m]>
It seems ... sad ... that they rebuild the platform from zero on every build.
<peepsalot>
understood. and my use case is that I'm impatient to wait for that check mark (or X) after each push :)
<JordanBrown[m]>
I can't see any reason not to change it over to doing them all at once.
<peepsalot>
looks like it did save a few minutes for that step
<JordanBrown[m]>
Excellent.
epony has quit [Quit: QUIT]
Trieste has quit [Ping timeout: 260 seconds]
Trieste has joined #openscad
<teepee>
peepsalot: IIRC it did fell over when doing system update + package installs in one call
marcus has quit [Remote host closed the connection]
marcus has joined #openscad
<peepsalot>
i think the initial commit mentioned AppVeyor which I'm guessing is different from Github Actions environment, so probably no longer an issue?
milza1 has joined #openscad
milza has quit [Ping timeout: 260 seconds]
milza1 is now known as milza
little_blossom has quit [Ping timeout: 260 seconds]