Where are the constraints / mates ?

[Dan_Margulius]

Hello

Maybe i am missing something, but it is impossible to built mates in Fusion.

How can we work in an assembly without mates???

Moving parts and Joints it's not it. 

Thanks

Dan

[TrippyLighting]

The reason I linked to some training materials was not to address the specific need of that V-Groove/Wheel joint. It was to provide you with a bigger insight as to how the joint system in Fusion 360 works.

I don't think we need to have geometric mates implemented, but we do need 3-4 additional joint types in Fusion 360 that would address such needs with a single joint, which is what Fusion 360 aims for, not with a number of different mates, which is what Fusion 360 tries to avoid.

[neil]

Thanks Peter, and I meant no disrespect.  Yes, there's certainly a workflow re-think that folks like me must overcome to become proficient; the videos help quite a lot.

I'll check in periodically to see if Fusion gets better at the specific kinds of designs I do.

Also FYI I'm already a user for its CAM abilities -- working quite well for me in that regard.  It makes my CNC milling machine "sing"!

Best

Neil

[TrippyLighting]

To be honest, at this point in time I would also not use Fusion 360 for machine design if I'd do that on a day to day basis and in a professional environment.

The joint systems works fine but there are many other stumbling blocks to overcome. I have to admit that in the area of machine design I have higher end expectations on a CAD system than many people. I've worked for two large factory automation system integrators and in these environments the functionality of a software goes beyond it's pure CAD abilities.

The ability to integrate it into a PLM, MIS and ERP systems is essential in a business organization that builds custom machinery on a day-to-day basis. The BOM functionality in Fusion 360 I'd describe as best as rudimentary. For a machine design with potentially thousands of components it's not suitable.

• No configurations (A feature that is an absolute MUST for machine design!)
• No library functionality for components or features.
• Strong limitations in working with linked/referenced components, e.g. lack of in-context design.
• No referencing/linking across projects.
• Performance problems with timeline enabled designs (which is what most machine designers use nowadays maybe just out of habit)
• Drawings not fully up to standards and not performant enough.

On the other hand, the CAM functionality seems to be very strong in Fusion 360

[neil]

Excellent reply, and yes those features are important to me as well.  Particularly (on the cad side) the ability to do in-context design.  And boy would I miss configurations!

For hobby or infrequent use F360 is a great deal.

Best

[TrippyLighting]

---

@Anonymous wrote:

For hobby or infrequent use F360 is a great deal.

---

As it pertains to machine design, agreed!

I know people that use Fusion 360 all day every day to successfully design and manufacture products. A friend of mine runs his three companies entirely on Fusion 360 on macOS. That's also where I personally see Fusion 360's strength.

[Anonymous]

straight trash. people should know what flush and mate mean.

[etfrench]

Is flush like in toilet? Can mate be discussed on public forums?

[Anonymous]

haha you are funny

[Anonymous]

While I admit I do have a lot of learning still to do on how to adjust my work flow to the decisions made by F360 designers...the idea that flush, mate, and insert are difficult to understand is almost laughable. And I'm not trying to be rude. But any engineer should certainly be able to grasp these easy concepts. Now my time with Inventor is orders of magnitude larger than F360 (15+ years to about 5 hours), but it was very quick to see that Fusion 360 pushes for a top down type of designing, which can be great, and in other instances is very limited. Using the same part or sub assembly across multiple larger assemblies is done every day by manufacturers who are not making one off designs never to be repeated. I know I will get better with F360 over time, but choosing to remove the ability to fix an object to another in 3 very simple steps incredibly unfortunate. Why both methods weren't maintained, I have no idea.

In working with Autodesk on improvement projects in the past on previous versions of Inventor, they were some of the first to admit, they are not mechanical engineers. They are not making weldments, machining parts, or anything that I do constantly in my every day work flow of designing, refining, and validating new parts and assemblies. It seems F360 was done in the exact same way. I am very glad to have a home use license for Inventor but I know I may not always have access to that so learning something as powerful as F360 that is free is a useful tool. So I'll go look up some more tutorials and videos before posting here with questions...often they get answered, though a good portion of the reading I've done proves that there are those that seem to take joy at pointing out how ignorant and how much better they are than others who are simply just trying to ask a question to preform and action that they think should be very easy since every solid modeling program they've ever used works differently than F360.

[Anonymous]

I am coming from Inventor, and joints are really melting my brain. Maybe I am missing something, or I haven't played with joints long enough to get used to them yet, but I have a feeling that this will never fully make sense to me after coming from the Inventor world.

Here is a very simple problem that I cannot solve using joints (I am saying that I MYSELF cannot figure it out, not that there is no way): I want the face of a part to be tangent to a cylindrical surface. The way I've tried to solve this is by creating a construction plane that is tangent to the cylinder (easy enough, and familiar to an Inventor user). Then, I try to create a planar joint between the part face and the tangent construction plane. No dice! Very very very frustrating! What I am trying to do makes perfect sense, even within the wacky world of Fusion 360 joints, yet the software won't let me do it.

I think I've figured out why the joint that I want isn't allowed (but I still haven't figured out how to solve the original problem that I described): You have to select a two POINTS to form a joint, and there are no (selectable) points in a construction plane, so there is no way to form a joint between my tangent plane, and part face.

As you said, joints do not lend themselves very well to the task of taking two parts, and saying "this is how they are supposed to be positioned in relation to one another". I do not see how you can have a productive CAD system where this basic task is very awkward at best.

Why can't y'all just have both: joints AND "traditional" constraints?? Obviously, the underlying technology of joints relies on constraints (a joint is really just a combination of constraints), so it shouldn't be super crazy to add constraints (as a feature of the software).

I'm sure joints work great in some situations. I can definitely imagine they require less steps to get certain tasks done, but it has been a huge struggle for me to work with them, and it sounds like you are already aware that there are important tasks where they make work very cumbersome.

Rigid groups/joints do not address the positioning problem at all. Rigid joints are only useful if you have a good way to control the position of parts (and sub-assemblies) in the first place. Move and align are ok in limited situations, but they are not good for proper position control, because (as you say), move is a bit "free form". Align is kind of interesting, but it has the same problem: it does not actually establish any relationship between parts, which is fine in simple cases where a single align (move) is sufficient to move a part where you want it, but if you try to do anything even slightly fancy involving a second move, then whatever align operation you performed before could very easily get lost (since, again, align does not establish a persistent relationship between parts). A second move is not at all uncommon, since (for example) aligning a face with a construction plane is not necessarily sufficient to fully specify a part's position. Sure, I can measure the distance between my face and the tangent construction plane to make sure that my part face is still coplanar after another align, but even in the optimistic case where the tool reports a distance of 0.00, how do I know that the distance is exactly 0 and not just some small amount that cannot be resolved by the couple of displayed decimal places?

This type of uncertainty does not exist in a traditional constraint system. Yes, traditional constraint systems are tedious, because of the (slightly large) number of steps you need to perform to create all the constraints that you want, but you always know what you are getting, unlike move/align.

The main situation where I can see rigid joint/group being useful is if you design a part directly abutting parts that it is going to be rigid joint-ed to later (which you describe as top->down design). In that case, the part is designed in exactly the place where it is supposed to be, and does not need to be moved into place at all before being rigidly joint-ed, avoiding the whole loosey goosey-ness of move/align.

Maybe all I am looking for is to make planar joint sane by removing the requirement that two points be specified, and instead only require that two planes be specified. Really, mate and flush constraints are equivalent to planar joint (other than this weird two point requirement).

[laughingcreek]

post a simple example model, I'm sure everyone here will be happy to brain storm way's of doing this with you.

[macleesj2]

i realize 5 years is a long time in the computer time but are there updated links to replace these?

[todd_sandercock]

Joints just don't work the way you want them to. After 5 years they are still painful and the solution is still mates. Stay with inventor/solidworks for real top-down modelled assemblies and use fusion for what it's great for which is simple, fast models and machining.

Todd

[Anonymous]

I am starting to get used to them. I think my hang up now is "How does Fusion know what orientation I want?" Seems to me, when you specify a joint, you need to specify a ray on each of two parts. A joint puts these rays together. E.g. a rigid joint makes the rays collinear and their origins coincident. But that seems to be underconstrained: what's not specified is the "clocking" of the two rays. That seems to be left up to the Fusion 360 gods (or there is still something left for me to grasp in my journey to understand joints in Fusion 360). Often, that's ok. E.g. if you want to rigidly join bolt to its hole, you do not care about this last degree of freedom (unless you are **** retentive, most people probably leave that last degree of freedom unlocked down in traditional constraint systems). I guess for other joint types (e.g. revolute), that last degree of freedom is meant to be not constrained, so this "underconstrained" problem only exists for rigid joints.

One of the keys that helped me wrap my head around joints is the notion of the joint "origin", which often coincides with geometry of the face to be joined (e.g. the center of a circular arc, or the center of the face). For cases when none of those points is what you want, you can create your own joint origin. E.g. you can draw a point in a sketch. There is also the "between two planes" option, but frankly, I do not understand how that works anymore XD.

Reading the documentation really helped (sometimes, the manual is comprehensible!). The problem with tutorials that you find in YouTube is that everybody seems to gloss over this joint origin concept, which is explained in the manual. The pages in the manual that really helped me were

https://help.autodesk.com/view/fusion360/ENU/?guid=GUID-2C7DE501-13D6-416A-A203-F070D11FA889

and the subsequent page explains placing your own join origin:

https://help.autodesk.com/view/fusion360/ENU/?guid=GUID-D1A77DB8-8E42-4BEA-82DA-F51B39635DC0

[Hfrossard]

For years now I wanted to migrate from Inventor to Fusion. But these kind of basic problems are extremelly stressful.
Constraints are needed for a lot of things that joints do not work for.
How on earth assembling a rigid metallic structure using "joints" make sense? 

Maybe Autodesk should just focus on one software: Inventor or Fusion. 
Wanting to embrace the World might not be the best thing.

[TrippyLighting]

---

@Hfrossard wrote:

...
How on earth assembling a rigid metallic structure using "joints" make sense? 
...

---

Can you elaborate on that?

What difficulties are you encountering in joining rigid metallic structures?

[neil]

It's been a while since I looked at Fusion.  In fact I posted a challenge video two years ago to see whether Fusion could conveniently, and in a proper parametric manner, allow for a wheel in a track.  See: https://www.youtube.com/watch?v=AKgo81sLXv4

Read through the responses, it's NOT possible, other than with laborious band-aid approaches.  I gave up and went back to Solidworks, where I'm very happy.  FYI it's also easy in Inventor,  a product Autodesk is trying to kill.

Mates -- they're exactly what's needed for certain jobs.

[TrippyLighting]

---

@neil wrote:

... Inventor,  a product Autodesk is trying to kill.

---

That is probably the reason why they've added a lot of things to Inventor that were first available in Fusion 360, like T-Splines, Joints (yep, Fusion 360 style joints). 

[neil]

Not to get into a contest about this, but didn't they just cancel Inventor LT?  I see the migration of those Fusion features into Inventor a not-so-subtle way of paving the way for Inventor users to "see the light" and switch over to the obviously more worthy Fusion.

I'm NOT bashing Fusion, it's amazing in its own right.  I just bridle against corporate decisionmakers who KNOW what's best for us.  You like what you like, I get that.  But let's not put on blinders.  

I throw down the gauntlet to you as well.  Really solve the V-Wheel-In-Track problem (as seen in the video) with Fusion in less than about a dozen steps, keeping it parametric, understandable by mere mortals, exact, flexible and easily modified.  No fudging, and absolutely no numbers punched in.  Then I'll reconsider my stance.  You should read some of the kludges that "experts" put forward in the comments to my video.  Not one of them came up with a sensible replacement.

BTW this specific task is trivially easy with Inventor and Solidworks (with constraints or mates), takes very few steps, is easy to understand and instantly accommodates changes in wheel diameter.  Autodesk could EASILY add an equivalent Joint in Fusion to make this problem go away, but my protestations fall on deaf ears. 

Prove me wrong, please!  And do it in a video for the world to see.

Neil

[davebYYPCU]

Just as easy in Fusion. Wheel Component needs a sketch point.

Difficult is segmented and 3d tracks.

Might help