While loading a 3D model into a CNC program and letting a machine go to town on a piece of stock is awesome, there’s a lot to be said about the artistry, craftsmanship and tactile feedback of carving a project by hand. [Amit Zoran] and [Joe Paradiso] created a nice bridge between these two approaches with their hand-held, but still digitally controlled milling device they call The Free D.
The Free D looks like your run-of-the-mill handheld Dremel tool with an engraving attachment and a few extra servos attached for good measure. These extra parts serve a purpose: the tool actually keeps track of its own orientation in 3D space. With the help of a few magnets underneath the work piece, the Free D sends its orientation back to a computer running a CNC program. When the computer detects the engraving attachment is getting too close to the desired shape, the Free D automatically retracts its own tool head.
Given the insanity or expense in building our buying a mill with six degrees of freedom, the Free D looks like it could be a useful tool in a few model maker’s toolboxes. Check out the demo video of the Free D after the break.
34 thoughts on “Handheld CNC Fabrication”
That’s great – was thinking about possibility of a similar setup a while ago, but hadn’t fully sussed the possible ways of constraining the operator’s movements – (was thinking along lines of a robot arm with motors that would give force feedback) – this is a much neater approach.
Next step – increase resolution? A final finishing pass with a finer tip?
Next step: Mill solid stainless steel.
Yup, just you; a carbide bit; and a location-aware plunge router. Have fun with that, you’d better have fore-arms like Popeye… ;-)
On the up-side, now that I think of it, this would be one hell of a way to upstage the guy at the county fair who carves wooden bears, out of stumps, with a chainsaw…
wow, i am much more interested in the magnetic source tracking system than application of it…
Yeah, I would like to know how they did that. The video made it seem like the only hardware required was a pc and the tool. I wonder if they attach a magnet to the work?
if you look at the diagram, they say they have a magnet under the workpiece.
So now an ‘artist’ can say his milled sculptures are ‘handmade’?
Why not? holding the cutting tool hear isn’t all that different than wielding a chisel or knife Some carvers us lay out lines to guide them or gauges. when the need more precision for some reason
Surely this tool doesn’t have to be use from the beginning of the project. Hopefully it’s smart enough to allow the user to remove most of the waste wood using other tools first My guess is plenty of carvers use rotary tools at some point in their work already. As I see it all this really does it let those without the artistic touch make facsimiles of other artwork. Also might speed up commercial for mass production processes where some handiwork is need to create forms for casting jobs
Very cool! Our company (Blue Belt Technologies) has built a very similar device for use in orthopedic surgery. Please come check out our webpage – It is always gratifying when someone sees the promise of our technology.
Blue Belt Technologies
Without coming through the USPTO, I suspect you are not the only company who “owns” this technology.
Corrected link: http://www.bluebelttech.com
Sorry about that!
Interesting concept, hardware, and software. I can envision a lot of potential applications for it. I wonder, for example, if a tool like this, with appropriate bits, might even be useful in a surgical theater. In that case, your “CAD model” would actually be CT or MRI imagery.
Somehow, though, I have difficulty accepting the demo video’s assertion that this is a tool for producing “hand-crafted” items.
Is this a “hand tool?” Yes. But the “tool” already knows what the desired outcome is, and is actively engaged in achieving that outcome, irrespective of the operator. The “craftsman,” on the other hand, seems to be contributing little more to the process than acting as an active tool stand. My point is that, because something was produced using a hand tool, it doesn’t necessarily follow that the item was “hand crafted.”
Take the argument to its extreme: If I bolt a pistol grip with a trigger switch to the side of a 5-axis CNC, have I not produced a “hand tool?” I would not regard the parts produced with this tool to be “hand crafted.”
Variability is a characteristic of “hand crafted” items. However, I don’t think it is the essential characteristic. Simply introducing randomness or slop to a pre-programmed machined does not result in “hand crafted” parts, either.
My comments are philosophical ones, primarily. At the end of the day, I really like this tool and I’m envious that I don’t have one in my shop!
Simply introducing randomness or slop to a pre-programmed machined does not result in “hand crafted” parts, either.
Or does it?
I think that perhaps this is more of a ‘middle ground’. Personally, I don’t think you can call hand crafted having just used it to mill a basic shape as in the video, BUT, should time be taken after this to smooth & apply detail which is not computer-aided then I am happy to say that hand-craftmanship has taken place. Anyone else care to chime in?
There’s no black and white when it comes to hand-crafted in my opinion. If you manually carve a figure with a chisel, was the chisel not responsible for the forming of the wood? Just different levels of craftsmanship.
Your question doesn’t make sense, unless of course, you take the position that the inability of a human to produce numerous, perfectly-identical objects is a THE defining characteristic of “craft.”
My position was that non-uniformity is a secondary, or even incidental attribute. “Craft” comes from a person’s talents and developed skills, colored by their perspective and experiences.
One of the important elements that a craftsman brings to the table is the ability to deviate, at any point, from the plan. The carvings in Mt Rushmore, for example, could not have been done by a programmed machine. Why? Fissures, weaknesses, and flaws in the rock caused the artist to have to constantly reinterpret and re-envision the finished product throughout the project.
A machine, in contrast, would have ground through the material without regard to the personality of the material it was working. The result would have been a perfect copy of the CAD model, that would have crumbled to pieces shortly thereafter.
I’m inclined to agree with Cold_Turkey’s assessment. Necks and body components for “handmade” guitars are routinely shaped by CNC wood machinery. However, all of the fine fitting and woodwork is done by hand. The only difference between these guitars and ones that are carved/cut/shaped completely by human hands is that they are less expensive and are of a higher quality.
An interesting discussion, I must say.
There is a great deal of difference between a passive tool like a chisel and the tool we’re talking about here.
The chisel does not make independent decisions about where to cut. The tool in this Hackaday post does. In fact, if the latter tool doesn’t like where you’ve positioned it, it has the authority and power to reposition it’s own tip.
The difference between this and the chisel is not a matter of degree. It’s a matter of apples versus cinder blocks.
I see your point and agree there is a world of difference, but this tool still requires the operator to appreciate the desired final form and manipulate the tool around the stock accordingly.
While it can adjust its position, its ability to do so has its limits as witnessed by the imperfections in the finished piece. At the end of the day the results achieved with any tool are always going to be influenced by the tool and the operator.
Oh this is very nice.
Now how about doing this in reverse, a additive process, like a handheld reprap until it matches a virtual object?
If your use of the term “matches” is extremely loose, then perhaps one could accomplish something approximating a 3d object with a reprap “glue gun”.
Something tells me though that the final results are not likely to be pretty.
I think too much of additive style 3D printing relies on the position of adjacent voxels in order to produce decent accuracy in the reproduction. If it was being done free-form, you wouldn’t be able to control the order in which droplets were deposited.
Just one question, why bother with the retractable bit, why not just kill the power to the bit?
Was thinking the exact same thing.
good point – maybes a bidirectional PWM motor controller running the spindle in a tight PID loop off of the spatial position data – initial ‘rough cuts’ run the bit at aggressive high speeds, then progressively slower/finer cutting as you reach the detail – bit comes to a dead halt when it hits the defined spatial boundary
Loop control might also benefit from being ‘aware’ of the stock material – e.g. expanded polystyrene vs wood/steel etc
As per the ol’ :
‘different strokes for different folks’
Controling the speed may not yield the desired results. I know that Dremil bits are “designed” to run within certain speed ranges. They are meant to run at high speeds with little pressure. Slowing them down appreciably leads to burning the material and dulling the bit faster. All tools have an upper and lower acceptable speed, for any given material. Operating outside of those limits tends to drasticlly reduce the life of the tool, and often doesn’t do the material being worked much good either.
Sabre toothed armadillo?
In general, I appreciate the concept, but I think the big deal here may be the tracking technology, rather than the fact of the device being handheld. My thinking is that the whole point–besides being the only real point of reference for the modeling procedure–is to compensate for the inherent instability of the operator’s arm; so what if one used a robotic arm that could perhaps be mounted in such a way as to access the whole of the model (directly above, for instance).
I believe the main reason it isn’t common for hobbyists to use robotic arms for rapid prototyping is the instability (lack of rigidness) in affordable models, but what if even a budget model could become a versatile investment for a maker? What if makers were commonly investing in, upgrading, modifying, and developing robotic arms as rapid-prototyping platforms with plenty of potential for re-purposing (in other aspects of small-scale manufacturing and engineering).
I think with this sort of tracking, a bot of most any quality could become an incredibly accurate CNC. I saw that they were in fact applying this for an additive process (probably extrusion) in a similar device, so my idea would apply to additive processes like FDM as well as SLS and other powder printing applications.
Finally, I wonder if the magnetic tracking could be used on a small delta robot module, then mounted to the end of a simple robotic arm for even greater compensation.
Hmm, I would have put the magnet on your material, so you can pick it up and move it around.
here’s the tracking device they used http://www.polhemus.com/?page=motion_fastrak
any clues how it works? – from web page: ‘AC tracking is also more accurate and faster, it’s clearly the superior choice. In fact, Polhemus originated pulsed DC technology, but chose not to pursue it because of performance shortcomings’
So, hows it work?? Not a Hall Effect array? No, can’t be… ‘pulsed DC technology’ / ‘pulsed AC’ – wots all that mean?? Sensor is a multi-angle RF antenna?? – I don’t get it – pls elucidate ! : )
& can’t see any prices there (but normally ‘contact sales team’ = LOT$…)
Pity they showed that African guy, because that sort of brings home that this would be used for stealing their business, when instead they should simply hire those Africans to sculpt stuff.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)