Double Sided Machining Using Optical Scanning?

I’m new to CNC so this might not be the best way to solve this problem, all comments and suggestions are welcome.

After watching the news about the Glowforge laser cutter and the build in camera I was thinking something similar should be possible to make for CNCs. Why is there so little data going back from the machine to the software? Everything is so one directional… The CAM controls the machine. What if the machine could tell something about the real world to the CAM software. Wouldn’t that make it easier?

I’m thinking having a 3D scanner/kinetic/camera/range finder mounted on each side of the machine directed towards your platform. The idea is that it should precise enough find your workpiece.

First step would be to 3D scan it. It should show up in the CAM software.

Second, in CAM you specify how you want to cut it and you cut the first side on the machine.

Third you flip the workpiece and rescan it.

In CAM you click that you have flipped it and the CAM software calulate what has been cut and what is left to cut.

You cut the final side.

As I see it the main problem would be to find a good enough scanner hardware that is still not super expensive, but with new cheaper tools like MS kinetic maybe it it possible.

Secondly, it would also need quite a lot of software, but nothing impossible.

I’m new to CNC so this might not be the best way to solve this problem, all comments and suggestions are welcome.

Welcome!

After watching the news about the Glowforge laser cutter and the build in camera I was thinking something similar should be possible to make for CNCs. Why is there so little data going back from the machine to the software? Everything is so one directional…

Because it’s expensive and adds little to nothing to industrial or volume production.

The CAM controls the machine. What if the machine could tell something about the real world to the CAM software. Wouldn’t that make it easier?

In a production environment that is a totally waste to have in every machine. Maybe one machine (or a tiny hand full of machines) would need it. Scanning is rare compared to production.

There are also ethical and legal issues about copying. Many companies would sue.

There are scanner adapters for a CNC machine to be able to scan an object. It is often used to reconstruct and ancient parts (no longer in production and have no owner) that needs replacement (Google J Leno making parts for 100+ year old cars) or for recovering a part you’ve lost the CAD for.

I’ve never used a scanner adapter but I know people who have. The scanner adapters work really real well. The data needs to “messaged” as it is noisy and full of aliases.

A scan produces a “point cloud” which is nowhere near accurate enough for many applications. For instance, a scan of a thread doesn’t produce a… thread… a mathematically precision formula that can be checked. At best we can try to match the point cloud to threads but we might mistake and cause something to be wrong.

Notwithstanding, scanning can be used for many objects. Frankly, many things are “good enough” for low or medium res. Even at high res, many objects produced could not be certified easily (e.g. the requirements are very strict for safety purposes).

I’m thinking having a 3D scanner/kinetic/camera/range finder mounted on each side of the machine directed towards your platform. The idea is that it should precise enough find your work piece.

First step would be to 3D scan it. It should show up in the CAM software.

This does work… but showing up in CAM would be a disaster. Very often the data must be smoothed, cleaned, adjusted and tweaked. It needs to go to for specialized CAD processing before it’s ready for CAM. Go here and read about what it is really like.

http://www.nextengine.com

Second, in CAM you specify how you want to cut it and you cut the first side on the machine.

Third you flip the workpiece and rescan it.

In CAM you click that you have flipped it and the CAM software calculate what has been cut and what is left to cut.

You cut the final side.

I’ve done this using the equipment I mentioned above. It’s a lot of work, even for a super high quality scanner.

You’re idea is sound. It is doable… after a fashion… :sweat:

As I see it the main problem would be to find a good enough scanner hardware that is still not super expensive, but with new cheaper tools like MS kinetic maybe it it possible.

Not even close. It needs to be multi-spectral and super high resolution. It is not a simple problem to get models that are good down to 0.001".

Secondly, it would also need quite a lot of software, but nothing impossible.

Very true. Follow the link I pointed to. Once scanned and cleaned, the data can be turned into CAD data. Then CAM can machine it.

The idea is sound, the technology just isn’t there. The legal and ethical issues of copying is non-trivial.

mark

Industrial machines use servo’s not steppers which have feedback so at least the machine knows where the motors are. That way if you miss a step (the motors slip) then the machine can compensate, if stepper driven machines like the nomad miss steps then the rest of the job will be wrong so it’s usually a case of scrapping the job.

If you are interested in scanning you might want to check out the Atlas 3D

The key points:

  1. scanning produces point-clouds, which get sampled/processed into meshes
  2. meshes are noisy representations of a surface, from which the computer doesn’t know how to “automagically” make a clean mathematical definition of design intent, even with very expensive software. Even with very expensive software you’re adjusting things by hand and with expert knowledge to get them right.
  3. noisy surfaces SUCK for doing CNC machining because of the physics issues of cutting things with a spinning tool (instead of burning them with light). namely the torque/moment forces involved, inertia, the stiffness of materials used in construction of the machine, the physical resonance/vibration of the machine, and so on.
  4. the scan data frequently is very “off” of the target surface. Like as much as ~5% off in commodity scanners (speaking from experience, not advertised specs)
  5. Scanning would be extremely difficult in a machining environment because unlike lasers which vaporize the removed material, which is then removed by a vacuum system, CNC machines throw the chips everywhere. It is very dusty in a CNC machine, and/or slippery/oily from lubricants/coolants. This makes any non-contact metrology hard.

And now more of the novella answer extrapolating on why things are never as simple as you’d like:

As someone who has done quite a bit with scanning and reconstruction of CAD data from scans, and used to sell the $20k+ scanners (Creaform & Konica Minolta, if you want to know brands) let me tell you this process is NOT easy, nor automatic (not even remotely) even with moderately sophisticated software. Even with very expensive software (RapidForm/Geomagic packages) you’re almost ALWAYS left editing things by hand to get it to what you need.

The problem with scanning is that it can’t easily infer design intent, and without knowing the design intent you’re working with a noisy observation of a thing instead of a clean definition of a thing

In most CAD packages (NURBS/“Solids” modelers) surfaces are defined mathematically and as simply as possible, and not as meshes (how the scan data comes in, if not as a point-cloud). The typical workflow is to “re-fit” your best-guess/best-case definitions of the surfaces over the scan, and then look at your variance between your scanned data and your CAD that you’ve drawn from it, and see if it’s “good enough” for your application. Often, your scan data is crap and very noisy, so you have to make sure your re-model from the mesh is even close to where the actual geometry on the source part was.

As @mbellon said it’s hard to get a part accurate to manufacturing tolerances for what the end customer/user needs. I’d argue it’s hard to even get 0.01" precision on most raw scan data, let alone 0.001". Error adds up quickly in scans, so the larger your part the more it’s going to be off—usually error is quoted for scanning processes as some amount of error per amount scanned. So “0.1mm per 100mm” or something like that.

Now, the reasons you need a nice clean, mathematical definition of your surfaces are many, but in particular you’ll discover as you do more CNC just how much motion planning goes into driving the CNC machine around, and how complicated the calculations for the tool-paths can be. This is the “CAM” part of the problem, and it’s very non-trivial already to get the right parameters for high-quality clean cuts on even fairly simply cutting paths that are traveling in clean lines and arcs—to say nothing of complicated jerky movements to follow the surface of a scanned mesh. In driving any physical machine you want to have as smooth a movement as possible, but this is especially true in a CNC machine.

This is because, unlike a laser, you’ve got to deal with a lot of inertia and forces that change rapidly and in a dynamic way as the tool engages more or less material, as it wears, and as the motors are working with and against the machine’s own inertia and stiffness characteristics. With a laser, it’s looking at a 2d picture and driving a known mass (gantry with mirrors & lenses) around and turning the laser on and off—there’s no sudden spike in torque as a tool first starts to cut into fresh material, or sudden surge of too much inertia when it breaks out of the material.

So that means you need clean, smoothly defined surfaces as much as possible for CNC machining, to make smooth tool-paths. You’ll discover as you go to cut lots of small details that require constant path-direction changes that the machine has to go very slowly, because it can’t build up any momentum. It’s constantly having to slow back down to go a different way—which means detailed parts or parts with noisy tool-paths take longer, and they also require greater stiffness in the machine and have a higher likelihood of errors in the final cut part because of the flex in the structure of the machine at each change in tool-engagement force.

Now, with all that said…

I think what you’re really asking for is “could the machine see my stock that I’ve flipped over, and automatically register the machining origin in order to not have misalignment between the ‘top’ and ‘bottom’ sides?”

And that’s a much easier thing to do, but typically it’s done with a contact probe in fancy expensive machines, or the stock is re-zeroed by hand on simpler setups. If you were to do it optically it’d still require some fancy computing and precise optics because the materials you typically machine are shiny, and reflections are very bad for scanning, generally speaking. This is also assuming that you’re going to thoroughly clean out your machine before running such an operation, and make sure there isn’t dust or coolant on the lenses or the stock.

There are people in this forum who have implemented their own cameras for registration on their machines, where they’re using it as an aid to aligning the origin of their CAM tool-paths to the stock corner or other reference point—but it’s still not automagic.

Hopefully that at least provokes some interest in the topics involved to learn more, and gives you a new respect for the technology we already have :wink:

1 Like

As someone who has done quite a bit with scanning and reconstruction of CAD data from scans, and used to sell the $20k+ scanners (Creaform & Konica Minolta, if you want to know brands) let me tell you this process is NOT easy, nor automatic (not even remotely) even with moderately sophisticated software. Even with very expensive software (RapidForm/Geomagic packages) you’re almost ALWAYS left editing things by hand to get it to what you need.

Thanks for jumping in! You’ve had much the same experience as I had. Yeah, it can be done but boy it is a lot of work…

I’d argue it’s hard to even get 0.01" precision on most raw scan data, let alone 0.001".

That’s why I emphasized needing multi-spectral (and multiple perspective) scans. Without stereo views and multiple wavelength scanning even begin to to reconstruct an object that well.

Now, the reasons you need a nice clean, mathematical definition of your surfaces are many, but in particular you’ll discover as you do more CNC just how much motion planning goes into driving the CNC machine around, and how complicated the calculations for the tool-paths can be…

I do 4 and 5 axis machining where the kinematics (how things need to move) are critical. Not only is the part in constant motion (movement and rotation) but so are the fixtures. Know precisely where things are becomes ultra critical. A collision - on a high end machine - doesn’t just ruin a tool or a part, it destroys a US$30K spindle and fixtures worth thousands of US$.

We’re talking about moving at speeds like 1200 IPM. Fast enough that if you stand near the machine - at least the first few times - you will jump when the machine moves towards you (even in an enclosure)! Boy, one better know where everything is and simulate like crazy before they try to actual machine!

In High Speed Machining (HSM) how the machine moves, even how it accelerates and decelerates, turns a corner and so forth is calculated into the machining process. The feeds and speeds is an enlightened SUGGESTION to the CAM software, the AI like software adjusts things constantly.

It’s been said that the an apprentice machines, a journeyman machines to machine, and a master machines to machine to machine. Once one begins to use and/or make jigs (machine to machine) - complex or otherwise - one becomes very aware of needing to know EXACTLY where everything is… often to levels well below 0.1".

While it’s fine a for a cleaned up scan to make a duplicate - common for valuable museum pieces (e.g. even on the scale of King Tutankhamen’s tomb and Chauvet/Lascaux caves) - (even cleaned up) scans just don’t cut it for precision parts.

I think what you’re really asking for is “could the machine see my stock that I’ve flipped over, and automatically register the machining origin in order to not have misalignment between the ‘top’ and ‘bottom’ sides?”

Optical registration in CNC is VERY DOABLE. The super high end CNC systems - 0.00001"- all use optical registration (e.g. LIDAR) and measurement. They sit on multi-ton, specially designed tables and float the entire CNC machine on air to minimize vibration. Needless to say, they are way outside any of our price ranges.

And that’s a much easier thing to do, but typically it’s done with a contact probe in fancy expensive machines,

Super high end machines like a Datron - with insanely small repeatability - use contact probes for edge finders. Yes, even with all the available technology.

Edge finding (I would have used a mallet to seat the stock; YMMV):

Z0 determination:

mark

Thank you all for the info. It sure is a lot of expertise on this forum and I appreciate the feedback. Would it be easier if the solution only focused on getting it correct in 2D, (x, y)? Maybe coloring the waste board in some high contrast color could make it possible for a camera to find the edges…

The edge finders on the video seem to only look for a few locations. Maybe you don’t need a big point-cloud to find out how the stock has been moved. Maybe a handfull of points is enough. So maybe it should not be a 3D scanner but more something closer to a laser range finder (or a row of range finders on each side).

I simply use two dowel pins in the middle line of the plate to secure the stock. This gives you flip line. In order to use it zero X and Y in the center using Quick Position feature of the CM.
In this situation human is the weakest part (drill the holes properly, setup CAM software)

If you need full howto let me know

1 Like

Thank you all for the info. It sure is a lot of expertise on this forum and I appreciate the feedback. Would it be easier if the solution only focused on getting it correct in 2D, (x, y)? Maybe coloring the waste board in some high contrast color could make it possible for a camera to find the edges…

Frankly, this is so much harder than getting good with methods that have served well for 200 years.

Two sided machining requires stock of known thickness and parallelism. Without this, the assumptions of the flipping math will not lead to properly lined up faces.

I do it this way:

A) Place a spoiler and machine it flat.

B) Place stock and machine it flat.

C) Flip the stock and machine it flat, also to the necessary thickness.

D) While performing C I cut out the stock to the dimensions I require.

Now I have stock (“a plate”) with known thickness and face parallelism - and the right size with nice edges.

The edge finders on the video seem to only look for a few locations.

Since one has to machine their stock to be parallel and of known thickness, it’s almost trivial to also machine it to have squared sides. Sampling one point is all one needs if stock has been machined… :wink:

Maybe a hand full of points is enough. So maybe it should not be a 3D scanner but more something closer to a laser range finder (or a row of range finders on each side).

You only need two points - form a line that can be used to determine the rotation (flip). Pins, screws, bolts, guides. I’ve seem all of them work quite well, even for parts that had to be accurate to 0.001".

There are variations - opposite corners, edge flip, four corners (which is real two pins with paranoia) - but which method to use depends on how you CAM software handles the flip.

Two sided machining is often done today with 4 and 5 axis machines. The machine itself provides the flip line.

Regardless of what machine one was, with some thought and some practice, their work flows can become highly efficient. We don’t need fancy tools to be surprisingly effective.

In this situation human is the weakest part (drill the holes properly, setup CAM software)

When I machine the stock for two sided, I machine the holes for the flip line! No human issues for the hole drilling! CAD/CAM screw up? Yeah, still human… :blush:

Here is a good example of the two pin method using the machine itself to create the two holes that define the flip line.

See the first two holes drilled? They are the flip line. When the flip is done, screw/pin those and the piece stays precisely aligned.

Here is another that uses the opposite corner method:

And this is WAY COOL - use the stock itself for the rotation:

mark

I do machine flip holes by Nomad as well. It works:)

I did tried “soap dish contour” method. Its too much waste material and that bring us to the size of the Flip Jig (almost:) )

I did tried “soap dish contour” method. Its too much waste material and that bring us to the size of the Flip Jig (almost:) )

Agreed. It’s just that it is such a cool illustration of thinking outside the box for two sided machining. :wink:

The keys are properly machined stock (parallel, known thickness) and a flip line that is aligned with the machine and also symmetrically with the stock.

mark

You are right, I should learn the standard way before trying to invent something new. I like the “soap dish contour”, maybe that can be a good learning project for me to try out.