Printable View
Hope my screenshot clearly explains my query:
Attachment 16199
Q1: Is it therefore down entirely to the user to decide on that option?
Q2: I'd also appreciate feedback on my code please.
Code:$fn=100;
translate ([0,0,6])
rotate([180,0,0]){
linear_extrude(6){
circle(d=74);
}
linear_extrude(3){
circle(d=87);
}
translate([0,0,-10])
linear_extrude(12){
circle(d=6);
}
}
translate([0,0,13])
linear_extrude(3){
circle(d=12);
}
I learned elsewhere that
A1: Cura was correct to add support (although it was far more substantial than i'd expected).
A2: My code was OK but 'cylinder' is better than using Linear_Extrude on a circle.
Yeah, I was going to say the same thing about cylinder vs. linear extrude applied to a circle. I also typically use a $fn for each instance of a circle, sphere or cylinder so that I can adjust the number of facets based on the diameter. For example, having 100 facets on the narrow stem is a bit of overkill.
So, something like this...
I could offer more style comments like defining the height and diameters as parameters instead of hardcoding them, but that's another level of preference.Code:cylinder ( d=74, h=3, $fn=100 );
translate ([ 0, 0, 3])
cylinder ( d=87, h=3, $fn=100 );
translate ([ 0, 0, 6 ])
cylinder ( d=6, h=9, $fn=20 );
translate ([ 0, 0, 13 ])
cylinder ( d=12, h=3, $fn=36 );
}
I view the need for supports driven by a number of factors. Print temp, application of print cooling airflow, extent of an overhang, are some factors that come into mind that can make a difference. In general, I'd say that yes, that lip is going to require supports in order to print. I'm not familiar with supports in Cura, but in Simplify3D there are a lot of tweaks that can be made regarding the type of support and the nature of the interface with the desired print. It can take some experimentation to get that working right.
Thanks a bunch Kevin, great post! Particularly for that advice about facets. I actually tried to get a deliberately low facet value for the little handle but gave up and settled for that universal 100. All because I couldn't find the syntax to confine it to a single object. (Cheat Sheet doesn't show me how/where to use it.)
I intend to start using parameters. Didn't on that first design due more to impatience than ignorance ;-)
Simplify 3D looks good, but at $149 (about £108) I'll settle with Cura for now. Barely used the features of that yet!
Support! Did you see the discussion I had at
https://www.reddit.com/r/openscad/co...upport_option/
That taught me a lot, and with subsequent reading I now appreciate that even apparently trival designs are not so trivial ;-)
BTW, did you leave an unwanted curly bracket at the end of your code? Works fine without it. And it also triggered me to learn that I can do a 'collective' translate() by using curly brackets.
Finally, an associated query (which I've also added to that reddit thread): do you know how to ensure there's not even a tiny gap between a model's surface and the base?
You're correct that the cheat sheet doesn't show all of the parameter options. The cheat sheet, however, is a great tool to have available online since each of the table entries are hyperlinks for more detail. Clicking on cylinder for example brings you to the detail page for that object, including a list of all the acceptable parameters.Quote:
Originally Posted by terrypin
Yeah, Simplify3D is pricey. I bought it years ago when it was $99 USD and I had grown frustrated with everything else, including Cura. Actually, support capability was a key reason to why I went with Simplify3D. I would assume other options like Cura have improved since then, and I'd likely have a harder time spending the $149 today.Quote:
Simplify 3D looks good, but at $149 (about £108) I'll settle with Cura for now. Barely used the features of that yet!
No, I hadn't seen that. I didn't even know there was an openscad reddit page. Joined. Just what I need - another reddit time suck.Quote:
Support! Did you see the discussion I had at
https://www.reddit.com/r/openscad/co...upport_option/
That taught me a lot, and with subsequent reading I now appreciate that even apparently trival designs are not so trivial ;-)
My bad. My test file had a combination of your script and my script so I could overlap the results, and that evidently led to an extra bracket being caught up in the copy and paste.Quote:
BTW, did you leave an unwanted curly bracket at the end of your code? Works fine without it. And it also triggered me to learn that I can do a 'collective' translate() by using curly brackets.
The curly brackets are a powerful tool.
I'm not sure what you're asking. If you're asking about how to make sure the bottom of your lid will end up right at z=0 on your print bed, there's no issue. Openscad won't leave a gap under that first cylinder, and even if it did, most (if not all) slicers will automatically place an imported design so that the very bottom of it sits at z=0.Quote:
Finally, an associated query (which I've also added to that reddit thread): do you know how to ensure there's not even a tiny gap between a model's surface and the base?
If you're asking how to ensure there is no gap between the cylinder stackup in your design, well, that's a different question. For simplicity, I ignored that in the sample code I threw back at you. There's generally not much of a problem with openscad leaving infinitesimally small gaps, but there are ways to ensure it. I find this more of a concern once you start applying a difference to carve away from solids in openscad. This can leave you with an infinitesimally thin slice of material that ghosts in and out as you rotate or move around the model view in openscad. Most people ignore that ghosting effect since it doesn't actually lead to a printing issue, but I find it annoying and strive to eliminate it. It's all about providing some overlap between elements in the design - let me know if you want me to expand on that.
If you're asking about how to make sure the bottom of your lid will end up right at z=0 on your print bed, there's no issue.
Yep, that was it. I had in mind an arithmetic error - typically when using Difference - going unnoticed, placing the model above the base.. But I've now learnt how to select a surface in Cura and elect to have that on the base.
if you wanted to make that completely without supports.
print the shaft and bottom cylinder as one piece.
then have a hole through the second cylinder so that it will fit over the shaft.
And nhave the same size hole partially through the top cylinder.
And just press and glue the three parts together :-)
you'll also find that $fn=50;
is usually round enough, the printer and slicer will round it off even more.
I always used to use fn=100
these days I rarely go higher than 75 and normally nearer 50. And It makes no obvious difference.
uhu all purpose glue is the best thing I've found for pla. Pretty sure whatever solvent it has does a partial dissolve on the pla.
I do know when i've deliberately stuck a couple pieces together and then broken them - they never break at the glue line.
have you bought a set of micro files yet ?
I prefer to print holes tight and then file to the fit I want.
Put it this way - if you make the hole too large, it can't be reduced. But if you make it too small - you can easily enlarge it.
Thanks CA. I managed to cleanly remove the main Cura support around the rim. But not the much smaller support underneath the little pull tag on top, which broke. Luckily enough left to ream out a shallow hole and glue it with superglue.
I realised afterwards, reading up on Support in general, that I should have done it exactly as you described.But I’ve since decided that if I print another for the same purpose I’ll simply enlarge the lower cylinder and hollow it out, effectively turning the ‘rim’ into a ‘lip’, and hence eliminating the need for support. The pull tag could also probably do without support by being a single tapered cylinder.
Or I could just buy a heating pad!
Now you're getting it :-)
If you bear in mkind how a 3d printer works whilst you are designing. You pretty much never need to use supports.
You can build out oir in at up to a 45 degree angle, beyond that it tends to get ugly.
here's an example I knocked up monday:
Now the code isn't pretty - cos I just wanted a box I could glue magnets to and attach to metal things like a magnetic window box.Quote:
$fn=50;
difference(){
union(){
difference(){
cube([100,40,30],true);
translate([0,0,2])cube([96,36,30],true);
} // end diffe
translate([40,18,0]) resize([14,8,14])sphere(d=5);
translate([-40,18,0]) resize([14,8,14])sphere(d=5);
} // end union
translate([40,20,0]) rotate([90,0,0]) cylinder(d=10.4, h=3.25);
translate([-40,20,0]) rotate([90,0,0]) cylinder(d=10.4, h=3.25);
translate([0,30,0]) cube([100,20,30],true);
} // end diff
The magnets are 10mmx3mm
If i'd just popped a cylinder into the sdie of the box. it would have needed supports.
So being a sneaky bugger I used a flattend sphere and popped the magnet socket into that.
Attachment 16210
The result being a box that has two magnets flush with the side that can be printed cleanly at 0.3mm layer height and - on the sapphire pro - 150mms
With no supports needed.
The magnets are a fairly loose fit - as I glued them. I could have made the hole diameter smaller and used a press fit. But I wasn't sure about applying pliers to the sockets so just used glue.
They are great bits of kit - really useful, don't know why i didn't think of it years ago.
But that's the beauty of 3d printing.
You can design something that does the same job as an injection moulded part. But designed specifically to be 3d printed.
Your homework - make the box parametric :-)
lol
Neat! Do you have a photo of it in use?
"Your homework - make the box parametric :-)"
The only parameters I didn't define were those for what I think of as your 'third cube'. That's your homework! ;-)Code:$fn=50;
// Defining the box
ExteriorL=100; // Exterior cube, length
ExteriorW=40; // Exterior cube, width
ExteriorH=30; // Exterior cube, height
InteriorL=96; // Interior cube, length
InteriorW=36; // Interior cube, width
InteriorH=30; // Interior cube, height
// Hole making
SphereDiam=5;
CylDiam=10.4;
CylHeight=3.25;
difference(){
union(){
difference(){
cube([ExteriorL,ExteriorW,ExteriorH],true);
translate([0,0,2])cube([InteriorL,InteriorW,InteriorH],true);
} // end diff
translate([40,18,0]) resize([14,8,14])sphere(d=SphereDiam);
translate([-40,18,0]) resize([14,8,14])sphere(d=SphereDiam);
} // end union
translate([40,20,0]) rotate([90,0,0]) cylinder(d=CylDiam, h=CylHeight);
translate([-40,20,0]) rotate([90,0,0]) cylinder(d=CylDiam, h=CylHeight);
translate([0,30,0]) cube([100,20,30],true);
} // end diff
Attachment 16212
(This web site has possibly the most buggy formatting/rendering I've ever seen!)
not bad.
About the only thing I'd have added would be a way to change the number of magnets and automatically move them about depending on number and size.
Or possib;y even change the number based on the size of the box.
But that's really overkill and i rarely bother to do stuff like that, when you can just copy and paste an existing line and change a couple numbers.
;-)
Also if you have the same thickness wall on all 4 sides of a box, it's much easier to have: 'wall thickness' as a variable and saves you defining all the interior measurements of the box seperately.
It also then allows you to use the wall thickness in placing things around the model easily.
Your method allows for different thickness walls, kind of.
And yeah the forum software really does not like a lot of browsers/opearting systems. We've never really tracked down exactly what combinations work best and which ones are a really bad idea.
I think it just doesn't like apple - but that's probably just professional bias, as I hate them lol
But for someone who hadn't even seen openscad a couple weeks ago - damn impressve :-)
Thanks CA.
I'm learning by the hour about each of the three main areas of my workflow: OpenSCAD, Cura and Ender 3 V2. One particular aspect of the former that I'm grappling with is visualising the effects of Difference(), especially when there's a 'hierarchy' of them. With that and Translate() too, I'm too often resorting to trial/error. I'm sure I'm not making full use of the various views in OpenSCAD and Cura.
I was about to give you an example I'm working on right now, but I'll do so later in a separate post with a more relevant subject heading.
But the biggest obstacle is of my own making: impatience to get stuff printed!
don't forget to use 'union'
Basically union glues everything between it's brackets into a single lump that you can then apply any of the other operations to.
And also it's worth knowing that you can apply operations to an entire bracket set.
Took me a while to figure that out.
for example:
makes my favourite torture test model. Screw benchies if you can print the hollow cube - the machine is dialled in !Quote:
difference(){
translate([0,0,10]) cube([20,20,20],true);
translate([0,0,10]) sphere(d=25);
} // end diff
So if we want to do something to the whole model we can simply put the operation in front of the difference and it will effecdt everything between the brackets - much like making it aseperate module does. But without the extra hassle.
So if you've got a section in the middle oif a module you want to change all at once - slap a set of 'union' brackets round it and apply the operations to the union operation :-)Quote:
rotate([45,0,0]) resize([10,40,30]) difference(){
translate([0,0,10]) cube([20,20,20],true);
translate([0,0,10]) sphere(d=25);
} // end diff
This is particularly useful if you use a 'minkowski' command to round off corners.
Because minkowski takes the degree of roundness form a sphere or a cylinder. It also adds all measurements together and changes the size of the model.
So a resize command before a minkowski command will keep the model the right size and apply more or less curvature depending on the diameter of the cylinder or sphere.
minkowski(){
cube([20,20,20]);
cylinder(d=5,h=1);
} // end mink
versus:
resize([20,20,20]) minkowski(){
cube([20,20,20]);
cylinder(d=5, h=1);
} // end mink
I did read the explanation of how minkowski works. And it involved robots driving around rooms on different vectors and colliding with things - yeah....
Minkowski is a fun command to play with - but it does reqquire some serious calculation - what with all those little robots driving around.
I really need to find that 'explanation' again - it was completely surreal.
It's amazing just how complicated you can get with simple things.
It's also why I always mark what operation the second curly bracket applies to. Otherwise it gets really confusing, really quickly.
It's also why i use very short variable names.
Some calculations for parametric models can contain several sets of bracketed calculations in a single slot. If you use long names for the variables, they become almost impossibly long and damn difficult to read.
For example, this line from a script that makes hinged moulds for clay sling glandes: translate([(gd+hod)/2+hod/2,gl/2+st,gd/2])
Is pretty easy to read.
but if the variables read:
gd=glandediameter,
hod=hingeouterdiameter
gl=glandelength
st=shellthickness
It would be so bloody long the whole thing just looks like a 60's stream of consciousness novel.
And that's just one operation in a line of script that has several, equally convoluted operartions and a shape.
There is frequently/occasionally (delete as appropriate) method to my madness.
Thanks CA, good stuff there that I’m looking forward to trying out.
well I may not be much good at maths - but I've got pretty good at working round them :-)
The minkowski thing was incredibly frustrating until i worked out i could just use a resize to keep it the right size. And then all increasing the diameter of the cylinder or sphere did was make the corners more rounded.