topic Help with latch design in Fusion Support Forum

Help with latch design

urischulgasser — Wed, 01 Apr 2026 01:31:57 GMT

I've attached a screenshot from Fusion. Assume that the red tubing and yellow tubing slide linearly towards and away from each other. Each is glued to its 3d printed 'wrap' mechanism as shown. The blue latch lever is actually three of them, 120 degrees around the revolve.

The notch on the end of the lever latch is for a rubber band. As the blue section approaches the red, It is supposed to self latch and it does.

When I actually print this And try it, It only takes about 10 pounds of force pulling AWAY from each other to lift all three levers and release the latching! I need it to technically resist any reasonable amount of force, but at least 25 pounds.

The issue is not any of the plastic parts breaking but the fact that I've somehow designed this incorrectly... Once latched, The lever should not be swinging outward with a simple linear force!

Re: Help with latch design

TheCADWhisperer — Wed, 01 Apr 2026 02:47:41 GMT

@urischulgasser

Can you File>Export your *.f3d file to your local drive and then Attach it here to a Reply?

I am going to first suggest you try adding a cylindrical undercut in the blue latch component.

My guess is that there is a bit of a radius in that corner.

The edge of the pink component pushing against that fillet is enough to start the lever moving up. Once it starts moving up it is pushing against an inclined plane which gives the same mechanical advantage of you trying to push up a heavy load up the plane vs trying to lift the load.

(view in My Videos)

Re: Help with latch design

urischulgasser — Wed, 01 Apr 2026 03:35:03 GMT

Already thought of the undercut or pin/notch type solution a little. I was actually hoping it was more of relocating the pivot point or something similar. (The fillets are .25mm So I don't think that's the issue...)

...The file that this group of components resides in is quite large and somewhat proprietary under development; I actually took out a few unnecessary components from the screenshot before I posted it. When I try copying it I'm told that there's a component pinned, and when I take off the pinning it still won't let me copy it. If I find a way to copy it and then remove some of the proprietary components, and then save that final copy to my local disk, I can post it as requested.

Alternatively I'm looking for a generic way to bring those two 'tubing handles' linearly together and lock them in position, with a manual unlock in any similar fashion BUT... Both components must be able to be 3D printed without any supports! "Challenge, set forth!"

Re: Help with latch design

urischulgasser — Wed, 01 Apr 2026 03:43:16 GMT

I just noticed that your picture was actually a video and I am extremely grateful you went to so much trouble to answer in detail!

I've attached a modification image to point out that since I print that red piece which is actually cylindrical, on the wide base vertically, I can easily add a complete rectangular notch to it in the correct location to engage or even all the way around, and the blue part is printed flat in the orientation shown for strength as well, so I can additionally add a square pin to that!

Presumably that will take care of the problem entirely. Obviously I would have to allow for a little over travel when locking for the height/thickness of those pins...

I am currently looking into one other option which is a type of twist lock with three rotational engaging... Not sure what to call them. Since you can technically twist the tubing clockwise that would force you to lock it manually as well as unlock it but it would suffice in a pinch; printing a test set now. I would MUCH rather have it auto engage to lock and have to be manually unlocked...

Re: Help with latch design

urischulgasser — Wed, 01 Apr 2026 03:46:25 GMT

Lastly, Would something like this do the trick??

Since the screenshot did not show the dimensions, I should point out that that flat part of the latch is only about 5mm long so if this would work maybe it's a better solution...