Custom machine for running router profiles on lead paint window sashes

[SeamusHannan]

If this is in the wrong sub-forum I apologise, I'm new here.

A friend of a friend of mine is a window restoration guy. He takes old windows and cleans the lead paint off them and redoes the woodwork, etc.

He wants to build a CNC machine that can relatively-automatically clear the paint off window sashes (specifically the 'inside' portion, the profiled wood that holds the glass pane in place against the frame and glazing on the other side) -- hopefully that makes sense.

The basic workflow, idealized, is:
1. Place window sash on CNC bed and clamp in place.
2. Probe (or some other method) to find interior corner locations for each pane.
3. Generate toolpaths for each square, rhombus, rectangle that is sensed.
4. Run a profiled router bit along each sash. (+Z at some generalized value to avoid hitting the glass)
5. Efficiently and effectively collect all dust generated.
6. Profit

Now, I'm not nearly experienced enough with software development or DIY CNC hardware to do this on my own. I know a little bit about CNC hardware and the basic CAD-CAM-CNC workflow. I use Fusion 360 and a Laguna Smartshop M and that's the extent of my experience.

From what I've researched, I could achieve his goal by using a closed loop CNC & controller with a touch probe, finding location data in the machine coordinates, converting that data to something Fusion can read, creating toolpaths, sending those toolpaths back to the CNC router, and running the program.

Ideally we can do something more streamlined where the location data is automatically converted to toolpaths at some variable -Z from the plane of the points found that keeps the bit away from the glass. This requires direct integration between the touch probe and the software.

From a hardware perspective, I can see this machine existing. 3-axis closed-loop CNC router, 4x4 bed with T-slots and clamps, touch probe, computer controller with custom software, profiled router bit. Ideally the touch probe and the router bit don't have to switch out on the spindle, and could exist near each other on the gantry.

From a software perspective I'm completely unaware of how to make this happen.

Any input would be greatly appreciated.

[awerby]

Sorry, but this sounds like an uneconomical thing to do, even if a machine like this existed. Trying to remove all the lead paint without taking out the glass just doesn't seem possible, and there's no way to "efficiently and effectively" get rid of all that toxic dust, which would coat the machine and the inside of any enclosure that contained it. The profiles of each old window mullion are going to be different, and the router bits to create them would have to be custom-ground. And being covered with old lead paint is not the only problem that old wooden windows tend to have. They're often cracked, rotten, and disjointed as well.

If your friend wanted any of item #6 on your list, he'd be better off concentrating on #2 - capturing the dimensions of the existing windows. That might be done with a machine, although it's hard to see it being much better or faster than an experienced person with a measuring tool. Making new windows from scratch that conform to the dimensions of the old ones, and fit in the same spaces, would not only be cheaper and easier, but the end product would be far superior.

[SeamusHannan]

I agree with you but he's still paying me to really push into the idea and see what is possible, so...

here I go, lol.

[SeamusHannan]

The profiles of each old window mullion are going to be different, and the router bits to create them would have to be custom-ground.

Most of the windows he has are pretty standard profiles and he's got a collection of bits to match them all.

Being covered with old lead paint is not the only problem that old wooden windows tend to have. They're often cracked, rotten, and disjointed as well... If your friend wanted any of item #6 on your list, he'd be better off concentrating on #2 - capturing the dimensions of the existing windows. That might be done with a machine, although it's hard to see it being much better or faster than an experienced person with a measuring tool.

This is an aspect I'm working on. A laser scan with good resolution fed into software designed to recognize the XY and ZX/ZY plane edges and create vectors / toolpaths along them to post to g-code could be a fairly streamlined process.

Making new windows from scratch that conform to the dimensions of the old ones, and fit in the same spaces, would not only be cheaper and easier, but the end product would be far superior.

Completely agree here. Just smash it all to dust and call it toxic waste I guess lol.