Wierd makerware issue

[Rob Griesbeck]

Hey all,

I've noticed that the new version of makerware seems to deal with intersecting meshes differently than the last version, let me try to explain:

We're making small mechs in one of my classes, we use 3d studio max and we used to be able to just "jam" one mesh into another, attach the 2 and it used to just print that as 1 continuous piece. But now I've noticed that when we do this, it seems to not print the intersecting parts, leaving us with 2 pretty much separate parts instead of 1 combined piece. Anyone noticed this?

I'm going to try some other ways, boolean, welding, not overlapping, etc. But I was just checking. As young 3d modelers, it was mych easier for the kids to just overlap 2 pieces and it would print 1 piece.

Thanks

[Tadeusz Piszel]

Hi,

I am completely new to 3d printing (1 week since my 2x arrived) but I have run into similar problem. I have designed small object using 2 separate intersecting objects. To my surprise my model was printed as 2 separate items that were barely bound together.

I had to go back to Solidworks and use Combine feature to convert my two objects into a single solid body. This solved the problem completely. 

I have no idea how things used to be in the past but I can confirm that it is likely a common issue with current makerware version.

Regards,

Tad

[Jetguy]

I think your logic is flawed here?

You were exploiting a bug in that the Makware was incorrectly merging the 2 meshes and now it properly enforces this.

You were teaching bad habits in a 3D modeling application IMO.

I'm not sure anything about that it OK?

 

Yes, it requires you to properly teach students how to create a model. Isn't that the goal? 3D printing is just an output, if they are learning CAD for a future career then just overlapping parts would only get them into trouble later.

[Franc Falco]

Thats explains so much!
I couldnt work out why when i first got my MB2 a year ago i could send any lashed together model to the printer and it would print ok - most times.
But recently (not sure which Makerware version fixed that ability) as you say jetguy, we cant do that anymore!
But do agree that correct optimised modelling is the way to teach, but if i do ocassionally feel a bit lazy, Netfabb cloud saves the day!

[Clinton Hoines]

Cura will also stitch multiple mesh models while slicing if you have a model that was created that way, found often on thingiverse.
It has 4 different methods of doing so and they work really well in the right combination.

[BTHOON]

I'm not exactly with you here.   I'm not sure I know a use case where I'd want to have the models NOT merge if I were to put them on top of each other.

I'm also not at all sure the previous functionality was unintentional.  Look at the "helper disks" they had (have?) in makerware.  The idea was you'd drag them to the corners and they'd stay down.   I tried that the other day, and where the disks were intersecting with the model, there was nothing printed.  That certainly doesn't match up with the previous included tutorials that MBI had put out on how to use the disks. 

Just saying that given most 3D apps poor ability to do a boolean merge, it is a loss in my experience to not have this work the same way any more.

[BTHOON]

Just to double check I'm not crazy, I did a fast google search and found this:
http://www.makerbot.com/blog/2013/04/19/keep-corners-flat-with-makerwares-helper-discs/
see step 7.

It's clear that MBI intended the described use at least at one point. 

[billyd]

Regardless of what MBI intended before or now I agree with Jetguy. Proper 3D modeling is not just jamming pieces close together and printing. Use a Boolean operation, it takes a few seconds and it's cleaner.

 

On Saturday, October 19, 2013 2:06:35 AM UTC-4, Rob Griesbeck wrote:

[Eighty]

I completely agree. But the glaring exception is that helper disks have been rendered useless. So it appears, yet again, that someone didn't think things through all the way before clicking "publish".

[Rob Griesbeck]

Let me be a bit more specific, as I feel I need to defend me teaching. With my senior students, they all know proper modelling practices, booleans, welding vertices, etc. (http://www.therpf.com/f9/my-latest-hellboy-prop-completely-3d-printed-samaritan-196171/) But I teach younger kids sometimes, and a 10 year old is lost with the "proper" techniques. "Jamming" a box into a cylinder is about their level, for the amount of time I have them for. My issue is just the change, I was just inquiring if anyone else had noticed the change,it appears some have.

[Jeffrey Birt]

I have complained to MB about helper discs being useless now. While one can be snooty and say that 'proper 3D modeling dictates that you do a Boolean and join the objects' that does nothing to address the real issue. Helper discs are handy to prevent warping and they can't be used now. I have tried creating a helper disc with one quadrant missing and overlapping it less than the shell thickness that this seemed OK.

 

I agree with an earlier comment that I cannot think of a case where I would purposefully overlap two objects in MakerWare and not want them to be attached. Remember MakerWare is NOT a 3D modeling tool, it is a 3D printing utility and thus should be designed to make the process of actually printing 3D objects easier, not designed to enforce 3D modeling constructs that don't really apply.

[BTHOON]

Agreed.

For my part, I do not have Solidworks, and find that few programs will do a GOOD boolean join and leave the model properly manifold.   Tinkercad will do it, but will decimate your model.   Netfabb Pro and the home version will do it, but not the free version.   Sketchup?  NO.

For my part, I like to leave my model alone in instances like this and not add the ears (so I don't have to version my models as a reference copy and one specifically with the ears for printing).  I like to add them at slicing time, and the change to makerware was nasty in my instance.

Oh well - we'll see what happens in then next version!

[Jetguy]

Ok guys, I get it you love to be lazy. The problem is, look at this from the "everyone else" perspective.
MakerBot has gone to great lengths to improve Makerware and dualstrusion, along with the new dissolvable support material.
The holy grail of 3D printing is making mechanical objects in a single print such as gears than would be on top of and extremely close to each other.
If your method of just being lazy and placing parts on top of each other assumes you meant to merge them, then the above scenario is very hard to detect when you don't want to do that function.

So the right answer is just do it right and the software works for everyone correctly. Now, I get the helper discs problem needs modified and is a broken feature simply because it was incorrectly implemented. However, given the new advanced dual features, I think it's a step backwards to please you folks who want automatic merges (which is really just dodgy anyway). I support a feature in the software that would detect overlapped parts and then ask you if you want to merge them. This requires additional code that would need to be developed and supports both camps. But it's also turning a printing software into a CAD software.

[Big-E]

True, I and many other people used to exploit the feature in makerware by which you could merge multiple models; this was a very useful feature.

I feel I should remind folks that, when slicing with Skeinforge, that this wouldn't work. However there are a few workarounds I've learned over the past year that I can share:

Firstly, don't jam objects inside one another. depending on the slicer you use, the space where objects overlap will be printed as a void; which is useful for printing cavities (like a hollow mold for casting). instead, design objects so they can be butted up against one another. If you overlap the perimeters of an object (think .1mm overlap) they will print as a single, solid object. Makerware allows for precise positioning of youe objects, so it's easy to align models this way to get a single, solid print. any slicer will work using this method.

as for the helper disks, they have a flaw, in that they are round. No flat sides make a disk very difficult to attach to an object without major overlap. like a tire, a round object has a very small contact patch, which in general makes them unuseable. There is a way to correct this; Make the disk very thin by reducing it's z-height to .1mm in thickness. Then, you can overlap as much as you want, as it's thickness will not intrude into the object, as it's no thicker than the perimeter. it provides a single layer, which will prevent curling. you can adjust the thickness to correlate with your selected layer height.

Of course, I've taken to using "helper squares instead. they work just as well, can be re-scaled easily, and can be butted up against an object perfectly. using two small squares on either side of a corner acts to "tape down" corners, they don't need to be large, and corners tend to be where an object curls up from the plate to begin with. simply model a simple cube, add it to the plate in makerware, and adjust it's scale so it's a layer thick in z-height, and adjust the x-and y to your needs. after positioning it, duplicate it, and move the duplicate to where you need it. repeat that last pair of steps until you have secured your object for printing. They come off easily once the print is removed.

Hope this helps.

[Jay Roth]

I disagree that this new behavior is correct.  Also, it no longer conforms to industry standard.  It is common practice for intersecting objects to be merged together as one at print time, and virtually every 3D printer system out there, regardless of manufacturer, follows this convention.

For those who argue against this, that is fine for you, however, there should be an option for either result, as there are good arguments for both sides.  That said, we are dealing with virtual data right up till printing.  That fact gives a certain amount of leeway into how we handle said data.  That is a benefit, though not one that has a real world analogy.  And that's OK, not wrong. We should benefit from the fact that we have this ability to manipulate this data in cool ways that are only allowed digitally, as opposed to constraining everything ONLY to the way the physical world actually works. That thinking is too restrictive, ignoring the benefits that the technology brings to the argument.

So, please make the behavior a user choice.  That would satisfy all, would it not?

Jay

[BTHOON]

Hey, lets not push on the LAZY ok?  Or more to the point, how about this: Tell me which free application YOU use that will merge models without completely destroying their printability.  It takes me 40 minutes in sketchup PRO to do this even with their solid tools.  It inevitably makes the models not manifold and requires a lot of cleanup.  If you're saying its lazy to not want to have to do version control on models for just wanting to throw some helper disks on there, I think you're being silly.  I didn't demand the feature in Makerware, but I did use it.  That clearly defined and documented FEATURE being removed was a pain, since I haven't found any mention of the removal in any of their release notes.  It is, as said in the initial post, an issue. 

[Steve Johnstone]

+1

[Steve Johnstone]

Again +1

[drobbins]

"Tell me which free application YOU use that will merge models without completely destroying their printability"

Blender

http://blender.org

On Saturday, October 19, 2013 2:06:35 AM UTC-4, Rob Griesbeck wrote:

[Joe Larson (aka Cymon)]

As a blender affectatio and apologist I have to admit that Blender's Boolean function sucks. It sucks less than it used to, but when meshes get complex it still sucks. I recommend taking your mesh and running it through netfabb, which often does a sort of Boolean but almost guarantees a printable mesh.

And I, too, have noticed Makerware's oddities about intersecting planes, most recently ruining my metachess set print. It also doesn't like flipped planes. It would be nice if it could detect some of these options and at leas advise you clean up your mesh instead of messing up your print. Then again I guess it's my fault for not turning on preview. On the 3rd hand it would be nice if preview could be turned on by default.

[Big-E]

Tinkercad does this as well; I often use it to merge STL's into a single printable object.

As for those complaining about missing features in the latest release of Makerware, just use the previous release. That's what I do. I refuse to use any new build of makerware until Skeinforge is implemented again.

I'm with Jetguy; you folks are lazy. Complaining does little to resolve a problem, it's better to just find a solution to resolve the problem. I, for one, have given several already.

[BTHOON]

Joe,
Contact Netfabb about getting a trial copy of Netfabb Private.   It's SHOCKINGLY more capable than basic, and the boolean joins/subs/diffs it does are AMAZING.  They always result in perfect models.  They also have an extended repair that skins the model and essentially rebuilds it.   I'm trying to steel myself for the $300 price tag I'm going to pay when my trial runs out.

After using it, I'm so loathe to go back to anything else!

[BTHOON]

Tinkercad DOES do the boolean operations, but as I said above, it also decimates your object for you, without asking.  If you're skipping Makerware completely, you're missing out on what is slowly turning into a very capable slicer with features that Skeinforge just doesn't have.  It might be worth a look for you.

You may be with Jetguy, just realize that Makerbot deprecating a feature that had become part of someone's workflow and triggering them to verify their issue with other users does not equal lazy.  If the need is to merge models, and there is a tool that does it easily, why not use it?  If that feature is gone, why not talk about it?  Moreover, while your solutions will indeed work in the case of the helper disks/squares as you recommend, it doesn't answer the initial question of merging models in general (except Tinkercad which again decimates models)

I have access to Netfabb PRO most of the time, but when I don't, I've still yet to find any tool (aside from Makerware back in the older versions) that can merge models without causing issues, so there's SOME reason to be slightly peeved.  That just doesn't mean the same as lazy which is a rather rude thing to say.

[Steve Johnstone]

Nicely put Bthoon.

I'm alway looking for ways to do things better / smarter. I just have two criteria's...

1....The quality must be the same or better
2.... The taken to perform the task must be the same or shorter.

As the saying goes " there are 100 ways to skin a cat" and I would never suggest that my way was is the best or the correct way of doing it. If someone asks, All I can do is let them know what works for me. If they take that and rework it... Hack it, whatever, to do something else that I hadn't thought of, FANTASTIC!

More importantly I also try to do this in a professional, objective and constructive way.

[Rob Griesbeck]

Finally had a second to play around in 3ds max. Max's boolean and pro-boolean work when done as a union. Attach and boolean attach do not work and intersections are not printed. 

I'm glad this has generated some discussion, I like the idea of a pop-up that asks how you would like unions dealt with. For now we'll use boolean unions to deal with the issue.

Grrrr, gotta teach booleans to 10 yr olds =(

[Big-E]

On Monday, October 21, 2013 12:24:11 PM UTC-5, BTHOON wrote:

If you're skipping Makerware completely, you're missing out on what is slowly turning into a very capable slicer with features that Skeinforge just doesn't have.

That's the operative word, though... Slowly. If you need to do something in the present, Makerbot Slicer doesn't cut the mustard. Skeinforge is better, and is consistent, whereas MBSlicer screws up prints so much lately, it's not worth the effort. the past few releases of makerware have yet to resolve slicing issues with MBSlicer.

Seriously, slic3r handles models better, and it'll give a preview of the gcode. That feature's been in there for a while.

Frankly, if you want software that does what you want, you'll probably need to pay money for it. Not trying to be rude, just honest.

I've said it countless times before: When MBSlicer outperforms and produces better results than skeinforge consistently, I'll use it again, but for the past few releases, it's only gotten worse. I used to love miraclegrue, now that it's MBSlicer, it just sucks.

[BTHOON]

Did I become immune to rudeness when I bought Simplify3D Creator?  ;)

I personally use Simplify3D for most slicing, but have found that my results out of Makerware are now close enough to ReplicatorG that most prints are indistinguishable between the two. 

[Da Ansa]

Sorry to bring up an old thread, however I'm having this same issue. I don't know which makerware version you guys were using, but for the current version 2.4.0.14, this feature is seriously flawed. For most of my models, I can just add helper disks just fine and they merge with the model. However, once in a while I have models that do the intersection void thing you guys are talking about. It's a very weird issue I can't find the solution of. I even tried inverting the faces using netfabb and trying it that way to no avail. Even when I have two models that are on the same bed, one of them will always work and one of them will always void. I'm lost. Did you guys find any answers?

I almost feel like this was an intentional feature but they forgot to add the menu setting as to how to flip the operation.

[Jetguy]

Probably should update to 2.4.1 which offers a lot of changes and bug fixes.

 

It's one of those things were it seems to leapfrog problems. They add a feature in one version, it works but has issue X, next version fixes issue X, but now adds Y and so on.

3 or 4 versions later, it mostly works as expected.

 

Being that no active development is happening on Rep-G, it always seems to be the fallback.