Alignment/joint challenge

[ipmcc]

OK, let's say I've got a component like this: (F3D Attached)

Then I add McMaster part number: 91290A115. When it first inserts, it does so at the origin. Like so.

 So I move it out of the way and end up with something like this:

OK. Now I want to place it so that it's passing, centered horizontally, through the slot in the tower, with the threads lying on the bottom, and with the head up against the outside edge of the Tower. I want something that might look like this (but this was done by hand, not with help from Fusion). Top view:

Side view:

OK, so obviously I can get to this situation with manual moves, but if I do it by hand, the alignment won't be perfect. This will make as-built joints (based on the non-exact, by-hand placement) problematic. Eventually, the goal is to have the screw move up and down within the slot in the tower.

I have tried numerous incantations of the Align tool, as well as Joints and Constraints. I can follow-up and add a bunch of videos if that would help show what I've already tried, but suffice it to say, nothing I've come across so far has even come close to doing what I want.

I want to:

* Align the inside face of the cap screw to the outside face of the tower.

* Align the outside of the threads on the cap screw to the crossbar at the bottom of the tower. 

* Align the long-way center of the cap screw to the horizontal center of crossbar at the bottom of the tower. 

* Once that's all done:

* Specify (preferably as-built) joints that cause/allow the screw to slide up and down in the slot, while preserving its X & Y alignment to the tower.

I would really like to know how to achieve this with the Align command first. I have participated in a number of other threads where people say, faced with similar situations, "Don't use Move or Align, just use Joints!" or "Just use Constraints!"

Firstly, in this case, I have not been able to figure out any set of Joints or Constraints that even comes close to doing what I want, but more importantly, I believe that I should be able to statically position (perfectly) these components first, and add As-Built Joints/Constraints/Rigid Groups/etc later. I have been told multiple times that using Joints or Constraints is, for some unstated reason, superior to Align/Moving parts into place before creating Joints (or, more specifically, As-Built Joints), but I've never read a good explanation for why. Nor, in this case, do I see any way to achieve what I want with Joints or Constraints.

Fundamentally, as I understand it, Fusion should be able to arbitrarily, statically position components within 3D space, so even if Joints/Constraints are somehow objectively better (which I'm still trying to convince myself of, but don't quite buy yet), I should still be able to align these parts exactly using the Align command, and then add As-Built Joints after the fact. While at this point I'll take any solution, I am less interested in esoteric combinations of Joints/Constraints, and more interested in understanding the complexities of the Align command, and more generally how components are positioned in 3D space.

'Align & Distribute' have been fundamental base-level functions of every 2D and 3D package (for example Keynote/PowerPoint in the 2D realm and  FormZ in the 3D space) I've ever used. Joints/Constraints/Align/Move, and my inability to figure out how to achieve similar results, has been one of the biggest bugaboos for me in ~a year using Fusion. 

I would appreciate any insight you may have on the more off-the-beaten-path uses of the Align command. I've used it a lot for aligning round (lower-case-f) fasteners and the circular holes they go into, but I still have never figured out how to do a relatively simple "align left/right/top/bottom edges" or "align horizontal/vertical centers" -- functions that have existed in many other packages, and are conceptually speaking, downright trivial (including in TinkerCad, also from Autodesk).

- Ian

[jeff_strater]

leaving the basic positioning question for later (or maybe someone else will get there first.  The answer is probably to use a Joint Origin), but I want to comment on this:

"I have been told multiple times that using Joints or Constraints is, for some unstated reason, superior to Align/Moving parts into place before creating Joints (or, more specifically, As-Built Joints), but I've never read a good explanation for why."

I can answer that very simply. Joints (or constraints) are better because they directly reference component geometry.  So, if that geometry changes, the Joint will still solve correctly.  Align (or Move) of components, followed by an As-Built Joint (or Rigid Group - they are the same thing) is not parametric to the referenced geometry.  So, if the geometry changes (cylinders get bigger/longer, etc), neither an Align nor the subsequent As-Built Joint will update correctly.  I will record a quick video to illustrate.

Here is the video.  Apologize for all the fumbling around trying to find commands - I am not used the new UI yet.  Two components.  In the first sequence, I connected them using a Joint, selecting the center of two faces.  Then, I go back and edit the block component two ways - making the block longer by editing the sketch, and then made it higher by editing the extrude length.  You can see that the cylinder responds as expected, and is correctly positioned.  Then, I undo past the joint (to reset), and do the same thing using Align and As-Built Joint.  The components, initially, are positioned the same as the Joint.  However, when I edit the block, the cylinder does not move in response.  That is because the Align is not associative to the geometry.  Hopefully that is clear.

[TheCADWhisperer]

@ipmcc 

Sketch2 in your Tower Component not fully defined.

Edit: Oops, I made a mistake but you should be able to now figure out the solution from my example.

Post back if you can't figure out how to change the constraint to remove the interference.

Note that the head of the fastener against the tower must be two different positions when it gets to the end of the slot where the cylinder material is not missing.

[jeff_strater]

and, here is how I would use Joints to position your two components.  Joint Origins allow detailed level of control that allows me to define a location (which is also associative to the selected geometry) to position the components as I need to.  The Joint command, then, allows you to choose the axis along which motion happens (in this case, a slider)

[ipmcc]

To start, I understand that fully-constrained sketches ought to be the first, most-fervent pursuit of every Fusion designer (until you start using splines, but I digress...). In this particular case, the unconstrained-ness of `Sketch2` appears, to me anyway, to have absolutely nothing to do with my inability to align the components. If you see some way in which I'm wrong about that, I'd love to know specifically in what way I am wrong. And yes, you posted the file `Assembly Constraints.f3d` that appears to have achieved the stated ends using Component Constraints. 

@jeff_strater has also given me a concrete example of how Joints are better than Move/Align. I'll need some time to digest what he showed in that video, and decide if that's how I want to approach future challenges like this. That said, I initially asked:

I would really like to know how to achieve this with the Align command first.

I really like to learn domains from 'first principles', and to understand those things from the basics, up the stack. I genuinely thank you for your time and effort, and for showing me another way to achieve my goal, but my original question remains: How can I achieve this desired alignment using the Align command? I'm still very much in learning mode. I know from previous threads in which we've interacted, that Component Constraints and Joints are your preferred way to encode the physical relationship between two components in a Fusion assembly. But for me? I'm still trying to figure out how to precisely locate un-Jointed/Constrained components in 3D space without using Joints and Component Constraints. If this seems silly to you, then I will fully understand if you disengage.

Again, I very much appreciate your effort -- especially putting together a version of my file that shows how Component Constraints work -- but I am clearly far below your level of expertise, and I'm still trying to figure out the basics like how to align two (lower-case b) bodies in 3D space, without using higher-order features like Joints or Component Constraints. I'm just trying to figure out how to make, for example, the left extents (or centers, etc) of two bodies line up. As far as I've been able to tell, Component Constraints were released in the July 2025 version of Fusion. I have trouble believing there was no way to accurately align multiple Components to one another before that.

Regards,

- Ian

[jeff_strater]

"but my original question remains: How can I achieve this desired alignment using the Align command?"

I don't believe it can.  Why is this important?  If a) joints are better because they are associative and b) there is a way to do this with joints, why does doing this with Align matter?

"I really like to learn domains from 'first principles', and to understand those things from the basics, up the stack"

Perhaps you are assuming that Joints, somehow, are built on top of Align.  That is not the case.  In fact, the opposite is really more accurate.  The original vision for what became Align was:  "Joints without the Joint".  That is:  Joints serve two purposes:  Aligning two components, and defining the relative motion between them.  At first, Align was just the Joint dialog, but with a checkbox that said "don't create the Joint" (or something, it was a long time ago).  But, in the interest of making Align simpler, a lot of Joint functionality was dropped.

[davebYYPCU]

I have trouble believing there was no way to accurately align multiple Components to one another before that.

Really. Who said that?  
We have always had parametric alignment. (It’s called Joints in Fusion)

Trying to use the wrong tool for the job, is bound to produce frustration.

Might help…..

[ipmcc]

I don't believe it can. Why is this important? If a) joints are better because they are associative and b) there is a way to do this with joints, why does doing this with Align matter?

Fair question. As I think about that question, I find myself thinking about all the physical modeling I've done. In every case I can think of, in trying to design a physical object, developing a static design has always come first. Whether that's using foam-core board, or a plastic model, or even a just a paper/cardboard mockup, getting a static model to look right comes first, and conceptually 'joints' come after that. If I can't mock-up an assembly in one, static configuration, I certainly can't mock-up an assembly with movable joints. Maybe this is my own, personal, mental limitation, but every other 3D package I've used has given me tools that let me mock things up statically before getting more sophisticated.

The other reason I ask is that there seems to be a stunningly huge amount of UI associated with the Align command, and, as best I can tell, no comprehensive documentation explaining what that UI does. If I have a cube face, and I want to align something to it, when I hover over those edges/face, I'll have access to (it seems) no fewer than 7 UI 'pips': The four corners, the four edge midpoints, and the center (centroid?) of the face. What do those pips do? How do I use them to achieve what I want? As I've said, I have successfully used the Align UI hundreds of times to guide round fasteners into round holes, but not much more than that, so what are all those pips for? They seem like they ought to do something, but as I've tried various combinations of pips, other than concentric circles, nothing ever seems to do what I think it's going to. It's been truly baffling. It seems like an immensely powerful feature, but one that just flummoxes me at every turn.  

At the risk of hijacking my own thread, another thing I've noticed is that sometimes components seem to revert to their initial position if you delete joints, ground or unground them, etc. I would really prefer to work in a mode where things stay where you put/move them, which is perhaps why I'm so hell-bent on positioning things correctly to start. I don't know that I have any deeper philosophical answer to your question, but one thing that sticks in my craw about all this is that the most basic of the Autodesk products (TinkerCAD) supports these basic, static positioning operations, so it's a puzzle for me why I can't achieve a similar result in Fusion. In previous threads that have touched on this, people have essentially told me 'you're thinking about it all wrong!', but what could be more basic in a 3D package than being able to reliably position items in 3D space?

Thanks for all the input. I'll try doing everything with Joints and see what happens. I still don't see how Component Constraints would be necessary. They may be a nice shortcut, but until I understand 'the hard way', I'm always skeptical of some new 'easy way'.

Regards,

-Ian

[jeff_strater]

"The other reason I ask is that there seems to be a stunningly huge amount of UI associated with the Align command, and, as best I can tell, no comprehensive documentation explaining what that UI does. If I have a cube face, and I want to align something to it, when I hover over those edges/face, I'll have access to (it seems) no fewer than 7 UI 'pips': The four corners, the four edge midpoints, and the center (centroid?) of the face. What do those pips do? How do I use them to achieve what I want? As I've said, I have successfully used the Align UI hundreds of times to guide round fasteners into round holes, but not much more than that, so what are all those pips for? They seem like they ought to do something, but as I've tried various combinations of pips, other than concentric circles, nothing ever seems to do what I think it's going to. It's been truly baffling. It seems like an immensely powerful feature, but one that just flummoxes me at every turn.  "

Those are "snap points".  The same snap points appear in Align and Joint.  They are intended to be "interesting" points on a face.  Things like vertices, edge midpoints, face midpoint, hole or arc centers.  If you pick one of them, you are saying, either to Joint or Align - this is the geometry I want to use to align the components.  If you pick two hole centers, they will be aligned.  In Joint, those snap points are associative.

I agree that Align, in particular, is under-documented.  I would love to take that task on at some point, but it's not my job, and I don't have time to do that.

"At the risk of hijacking my own thread, another thing I've noticed is that sometimes components seem to revert to their initial position if you delete joints, ground or unground them, etc."

This is an important concept to understand in Fusion:  The idea of a "capture position pending" state.  There is a lot of history as to why this was chosen, but that's for another post.  The important thing to know is:  Whenever a component is Moved (whether by Move, component drag, Drive Joint, etc) Fusion leaves the component where it is, but this UI appears in the upper right:

This says:  One or more components have moved.  There are two actions here:

1. capture position.  This inserts a "Capture Position" feature into the timeline.  This "records" the positions in the timeline.  Doing that will make the "Position" UI disappear again.  It is important to understand that this captures just the transform of the moved components.  There is no geometry reference captured here.  Capture Position is useful for doing "position-based" modeling.  For instance, creating a cutout for a component in its "closed" position
2. revert.  This puts all moved components back to their unmoved locations.  This is super useful if you are, say, testing the behavior of a mechanism by dragging components around.  Move them for a while, then just Revert back to their former locations.

Some operations, as a side effect, do an implicit Revert.  You have identified a few.  It is unfortunate, IMO, that this does not result in a warning about losing position state, but today, it does not.  Other times, if the operation could be position-dependent, you will see a prompt asking whether you want to capture or revert.

[TheCADWhisperer]

---

@ipmcc wrote:

but what could be more basic in a 3D package than being able to reliably position items in 3D space?

---

First Principles - there is a logic difference between reliably position and parametrically position.

I can create, for example, a 50x100 rectangle without dimensions or geometry constraints like vertical, horizontal, coincident.  But this rectangle has no adaptive intelligence that ties it to any other geometry (the blue lines in your sketch - they could be accidently moved, or the distance might not remain symmetrical about the Origin if underlying geometry edited).  If other geometry is edited - the underdefined geometry is not predictable.

Same with Align vs parametric Joints or Constraints.  

This is what sets apart parametric CAD from toy CAD.

Rather than use Align - I recommend using Move (Component) by eyeball close to intended location and then add appropriate parametric Joints and/or Constraints.

[ipmcc]

Rather than use Align - I recommend using Move (Component) by eyeball close to intended location and then add appropriate parametric Joints and/or Constraints.

Yeah... this is kind of where I've ended up, but I've just been burned so many times doing things "by eyeball" that I really try to be very methodical about things, and I try to only do things that produce reproducible results. For instance, if I have a hex cap screw that I want to 'perfectly' align in place, and I know exactly how 'tall' its cap head is (thanks to the Measure tool), then, if I have no better way, I can align it to its hole 'by surface', and then do a single-axis translation of the height of that cap, and even if the result isn't perfect, at least it's reproducible.

When I simply eyeball things, it's not reproducible. I raised this in another thread, but my real nemesis here is non-constrained point stacks at the origin of a sketch. I have been spending some time exploring the 'many, simpler sketches' philosophy that people pointed me toward. (As an aside, there are about as many things that I like and dislike about that approach, relative to the "one sketch per reference plane" I was using before, but it does seem to reduce the detriment of the "non-constrained point stacks at the origin" problem. I guess it's probably a net win.)

Thanks to all for the input. There is definitely some food for thought here.

Regards,

-Ian

[davebYYPCU]

You do know that the rest of us (who respond) don’t eyeball anything but splines, because there would be a future of frustration in built.

Joints allow selection of the mating faces.  
If you are calculating the thickness of the bolt head, you are doing it wrong.

Play around with all the setting options in the Joint dialogue box.  You would never need Align or Move Tools again.

Might help….

[TheCADWhisperer]

---

@ipmcc wrote:

When I simply eyeball things, it's not reproducible.

---

Just want to make this clear - when I refer to eyeball positioning of components, I mean virtual twin of the process we use to assembly in the real world.

In the real world we move the fastener close to hole and THEN fine position (Joint the threads).

Same with sketching a line, circle or rectangle.  I seldom enter the dimensions as creating.

I sketch the line, circle or rectangle by eyeball and THEN add the dimension(s).

(Maybe that is why we call it "sketching"? I am certainly going to have precise parametric dimensions/locations, but I'm not entering exact coordinates to create geometry like I did in AutoCAD back in the last century.)