And some more info:
ATC
My ATC rack is pneumatically actuated and the tool holders hang over the machine bed only when the rack is extended. I didn’t want the cylinders protruding out the back of the frame, so mounted them at an angle. (Under the table, collinear with the rails could have also worked)
The rack is mounted to the linear bearings with rubber isolators, so the whole rack has some compliance to it and thus I don’t have to be super precise about alignment and calibration of pocket locations. Not pictured in this particular shot, but I did wire up leaf switches to each fork to sense tool presence in each pocket.
Tool Setter
You can see it in the video up top, but it’s just a basic ebay-grade tool setter. I put an optical beam-break sensor above it for a little more peace of mind; it lets me rapid down and then slow down for the last bit of the approach to the setter without having to preconfigure estimated tool lengths in the tool table. The downside is that it’s rigidly attached so it eats up a chunk of the working area. It’s on my to-do list to mitigate that limitation, just deciding between several options.
Dust Shoe
System Design
This is the part that I’m most proud of and took the most effort for both design & fab. My CNC router is in my metal shop in a packed garage bay, nowhere near my wood shop, and I wanted neither the bulk nor the added equipment to go with a traditional 4"+ dust collection system for it. It also needed to be ATC-compatible. So, I elected to go with the approach popularized on higher-end handheld power tools with high collection efficiency (e.g. Festool) and go with a smaller-diameter system with high static pressure. The shoe design is inspired by the DATRON CleanCut, rides very close to the surface with a short foam skirt (rather than a large surface area of fairly leaky bristles), and concentrates the flow closely around the cutting tool. The bellows are 3D-printed from TPU/TPE; normally I run another matching bellows on the spindle itself so the spindle and shoe are essentially sealed together when the spindle is lowered, but on this foam workpiece the suction was too strong and it was lifting the foam off the bed.
I run it with either a powerful shop vac and a 1- or 2-stage dust separator, or direct on a dust extractor (Fein or Festool), depending on the job.
The hinge is lift-off and the ball linkages have quick-release sleeves, so removing the dust shoe entirely takes about 10 seconds. It also means that if the shoe crashes into an obstruction when lowering, it will just pop off the hinges - which isn’t completely failsafe, but more so than if it were to just get stuck and then be forced into the spindle (potentially the cutting tool) once the gantry moved and cleared the obstruction.
Actuation
As demonstrated, the dust shoe swings out from under the spindle and retracts vertically during tool changes. I wanted to keep the actuation mechanism compact, so both cylinders are mounted vertically and the shoe is hinged (to swing in and out of spindle alignment) and driven by a bellcrank mechanism. The bellcrank has a ball linkage to the hinge, and it’s an over-center mechanism which mechanically locks in place when the shoe is under the spindle, so I’m not relying on the relatively weak mechanical advantage of the relatively weak cylinder to keep it aligned.
Efficiency
The collection efficiency has been excellent. This particular case was a pretty good stress test since the foam “chips” are much bulkier than sawdust and the high MRR means the sheer volume of debris was near worst-case. The amount of foam dust you see on the machine spindle/bed in the last video is what I ended up with after cutting 15+ such foam panels and collecting 12+ gallons of foam chips in my dust separator.
Control
Not really the focus of this post, but I’m running the machine with LinuxCNC and a Mesa FPGA board. My configs are on github for reference: GitHub - Funkenjaeger/fj-lcnc-cfg: My personal configuration files for LinuxCNC · GitHub
Overall machine design
Not a lot to say here, it’s generally pretty similar to the Avid 4x2 except my Y rails protrude out the front so the spindle can reach a few inches out past the front of the bed. That way I can mount a workpiece vertically and machine features on the end, such as fancy dovetails on a board. I built a quick-release vertical clamping jig for the front for just such a use case, though that’s also not the focus of this post.
love it!