So, after another day of earsplitting whine I finally have a reasonable good (as in not total trash) core of a scale model of the blades that I intend to use.
That leaves lots to be done still, but there is definite progress. The main problems solved today were a glitch at the end of the curve generated (it dropped off too fast, connecting the top and bottom half in the wrong spot), a better way of 'zeroing' the Z axis of the router and a dramatic improvement in the speed with which the machine operates.
It now first makes a 'roughout', and then finishes off with another pass to produce a reasonably smooth blade.
I still seem to have a little bit of a problem in the Z direction, the blade is a little thicker (3 mm, 1/8") than I intended. That doesn't seem like a lot to sand off, but the whole point of this exercise is to make blades fully automatic, and with totally free geometry (within the limits of what the machine can make). Different sizes, twists, tapers, profiles, all controlleable through parameters in a custom piece of software.
Another thing that I'll need to add is a final pass to cut the blade loose from the remainder of the blank. The foam is very hard to cut nicely, it will tear easily. The mill does a much better job, so all I should add is a run around the outline of the blade at 5mm above zero at the last pass.
I'm beginning to get a pretty good idea what I want for a jig in order to produce blades. It will be a steel 1.5" square tubing frame, with an insert that can slide across the long axis. The frame will support the foam being cut lengthwise, the slide will help support the free end (possibly more than one slide). A zero preset will be integrated on one end (outside of the area where the foam is situated) to make sure that both 'halves' match seamlessly. The whole thing will bolt onto the existing table and have a set of levelling screws to make sure the bit runs true with the top of the foam. Pins in the edges of the tubing will hold the foam in place, four sharp stell pins will stand upright on the slide and the left hand side to hold the foam and to help with the registering. I may add a vacuum chamber underneath the foam (with supports) to help hold it flat, I've noticed a tendency of the foam to sag when it's being cut.
As you can see from the images with this much wider blank the vacuum cleaner is of almost no use in collecting the dust, it is simply too far away from the cutter bit most of the time. I'll have to somehow extend the hose and fit it right next to the router bit.
The machine is now so reliable that I just plunk in a blank, start the router (and the program!), then go have a cup of tea. That way I don't need to be in the noise and the dust longer than strictly neccesary. By the change in tone of the router I know when it's done, come back, flip the blank over, hit a key and take off for another cuppa. A little while later the blade is done. The machine travels now at 3"/second (75 mm), and could go much faster still judging by the size of the dust particles. Another possible speed increase could come from having a more curved path, or from running the bit lengthwise instead of crosswise (that way you don't need to ramp up and down all the time).
I had to touch up the edges a bit with sandpaper because of the thickness error, hopefully the next run that won't be neccesary.
Anybody interested in a discount on some styrofoam insulation ? I have *lots* of it... bring some glue !
here are the pictures of todays work:
Top view of the machine just before cutting:
The start of the first pass:
The roughing out phase:
The final pass of the back of the blade:
The finished core, you can clearly see the airfoil shape (naca 4412), the twist and the taper, the surface is pretty smooth (highest features < .5 mm), but could be smoother still. I intend to fiberglass over the finished core and I don't think that will be a problem without any further surface treatment. The bit of text that is still visible on the foam is proof positive that the 'z' was too high during this pass, otherwise it would have just removed that.
