[CamperBob2]

Also not clear why you couldn't get 90% of the way there by just filling the conventional T-slot more completely. We'll have to wait for some ME types to chime in, I guess, but I don't get this at all.

I could see this solution being better if you only had one fastener along the length of the pieces being fit together, but in many/most cases there will be more than one fastener. Seems like a real-world T-slot system will end up with fewer unwanted degrees of freedom than the ad is telling us to expect.

[johnwalkr]

I'm an ME and will share my thoughts. I think this product does make a lot of sense and I think the claims are true if a bit exaggerated. With T-slots, first of all, relying on friction only is totally valid but it is true that with the standard accessories like brackets, you can overcome the friction with a much smaller load than the structural members can handle. To fix this you end up with more cross-members, corner plates, more brackets, etc. The t-slot assemblies do not have more degrees of freedom, it's just that in some directions of loading, they fail to prevent movement at a low value. Anglelock uses friction too (just like anything assembled with fasteners) but it is more clever about it.

You could make the female threaded parts that go inside t-slot extrusions to fit more precisely, longer, with more fasteners, etc (and this is common). You could also add a protrusion to the accessories to fit inside the slot. But these obviously need a small clearances for all 3 parts to mate. Rotation would technically be prevented eventually, but stiffness at the instant a load is applied would not actually be reliably added. Reducing clearances is a non-starter. With great cost you could do it for a very small assembly, but it wouldn't be possible to go down this path as a general solution. The tolerances don't work at scale (at any cost).

At first glance you could make a very minor modification and add a chamfer to each groove in a T-slot, and a mating convex feature to the accessory parts such as brackets. But if you actually try and make this, You will also run into tolerance issues. You will have to choose to have clearance at the chamfer, making it useless to add stiffness, or clearance at the outside of the extrusion, making it useless to locate parts relative to the extrusion and also taking away the most useful contact area for stiffness.

   ___________
  |___     ___|
   ___\___/___
  |  _\   /_  |
  | |_______| |
  |           |

The anglelock solution is pretty genius. By having the fastener angled by 20 degrees (parallel to the chamfer), you can now have the required clearances for assembly but still locate precisely in 2 planes and apply a clamping force in multiple planes. Both of these features are not found in normal t-slots. There are nicely designed details about this but you can also just think about the angle of the fastener and realize the preload (let's call it P) of the fastener must ultimately make a vertical clamping force to the t-slot of Pcos(20) and a horizontal clamping force to the t-slot of Psin(20). The second video here[2] does a good job of explaining it, especially at 1:45.

I would probably prefer anglelock over t-slot or a weldment for something like a one-off cnc machine frame.

All of that being said, a huge demerit of this vs t-slot is it is not as readily available and you need a quote to order it. With t-slot extrusions you can configure an assembly with Misumi and within a couple of clicks order pre-machined extrusions and all of the related parts, and have it in a couple of days. But, checking the profiles it looks like anglelock profiles are also compatible with standard t-slot stuff. It also looks like the anglelock stuff is compatible with t-slot profiles (without all of the benefits of course)

[1]https://anglelock.com/collections/hardware/m6x10-39mm-bright...

[2]https://anglelock.com/pages/videos

[CamperBob2]

Great info, makes sense. (Mostly... I'm not sure I see how the tighter clearances with AngleLock are somehow immune to the tolerance concerns that exist with the legacy hardware.)

If it's backward-compatible as you note, that's definitely a win.

[johnwalkr]

The video/timestamp in my last comment is probably easier to understand than my explanation. Using z as the long axis of a beam, let's set up a coordinate system. There are 6 possible degrees of freedom: rotation about x, y or z or moving in x, y, or z.

     y  
     | 
     |___ x
     /
    /
   /
  z

Anglelock doesn't have tighter clearances per se. The anglelock is designed so that there is lots of clearance for assembly but when the screw is tightened, the part is not only clamped to the surface of the beam but also wedged into the slot. The only constraint that is not a hard shoulder is in the z movement direction. But it's wedged in the slot and not only held by the friction on the xz surface. So both movement and rotation are very well constrained in x, y and z.

Normal t-slot beams are designed so that when the screw is tightened, the part is clamped to the surface of the beam only. So movement is very well constrained in y only, and rotation is very well constrained about x and z only. The remaining constraints are all limited by friction on the xz surface.

Does this matter for most structures? Probably not. But if you need something very stiff and precise/self-aligning, anglelock would be a nice solution compared both t-slot beams or weldments.

You can add more screws but from experience to make stiff structures from t-slot beams, you have to connect beams together in multiple planes. i.e.: it takes a combination of machining threaded and thru holes in mating beams, corner brackets and/or cross-members plus plates. And you have to be quite deliberate in where/how to connect things together because you also need space to everything else like doors, motors, sensors to the structure. Plus all this stuff is not automatically perfectly aligned due to the clearances involved.

In my last comment I was trying to explain how just reducing clearances won't work. You always need some clearance to assemble and account for tolerances of parts, so even if you add a feature to constrain movement, when you apply load in some/most directions, it's only reacted by friction at first. So it won't turn into a parallelogram, but it also won't add stiffness in the same way. It also won't have the wedge everything together in the same way.

I think anglelock beams/legacy parts are compatible, but not legacy beams/anglelock parts. But this still means all your existing doors, mounts etc work with anglelock beams. In a new design you would still probably still use standard parts except for the structure, and in a pinch you could always add legacy beams to your anglelock structure.