What kind of engineer
 

can build me what is essentially a laser guided CNC machine?

If something is mounted to a table, I need a laser guided arm to "read" the object and then modify it (sand) to a certain spec. Even though hundreds of the "same" piece will be mounted, one by one, each one is unique.

Who would create this kind of thing?

I'd ask a cue factory who engineered their cnc work. McDermott cues has many different cues with varying inlay patterns. Here's a factory video on how they make cues. It only touches on the CNC work, but should give you an idea...

https://www.youtube.com/watch?v=Kj-jkVcYPpw

Try looking on Pinterest. ;)

Shaun, it sounds like you would plan to procure a commercially available laser-guided CNC milling machine and are trying to figure out who you would hire to "program" it for your use case right? What I am disambiguating here is that you are not talking about finding someone who will build you a custom CNC mill.

Grant, I think using and off the shelf robotic arm would be cheaper and more effective than a CNC machine actually. It would be best to replicate a human arm and hand, I think. Talking with an inventor/engineer would help finalize the design.

The key element to this device is the laser tells the arm where it is and how it should be sanded or otherwise acted upon.

CNC machines have Renishaw probes that can probe the surfaces in every plane and then determine the amount of material to remove.

3 axis or 5 axis??

That is great news, thanks. My sense is 3 axis would be fine.

A laser would be helpful in that it (just thinking outloud) may be more precise. For example, say you have a piece of aluminum and it has a pit here and there. The pit may .2mm in diameter and .5mm deep. The sanding head will need to sand either all of the material away to remove the pit or create a blend to a non-pitted surrounding surface.

What you are describing is a full scan to map the surface of the object that would then be acted on intelligently by an algorithm to drive selected remedial corrections which "should" give a corrected result. However, constantly modifying the "map" real time would be intensive. Resolution of the scan is a determinant factor.

It's a bit different than programming to take off material overall to a certain depth.

How does it know the pit is not supposed to be there?

That question is specifically meant to get you thinking.

You could try Haas but the way their F1 team is doing that may not be the best idea. ;)

Look up Coordinate measuring machine, so using one of these to scan and a reg CNC mill to finish, the software to get the program to talk would be another thing.

I have never tried sanding using a CNC mill though.

I used some small HAAS toolroom mills TM2, they were good machines.

Very good points here.

What might work as a v1.0 is the operator can determine a depth of material removal overall. That in itself would be a big help. For example, if there are 5 pits and a deep scratch, the deepest pit is 1 mm and the scratch is .05mm deep with .2mm deep pits here and there, having the entire piece sanded down .5mm would be very helpful.

In my thinking, the laser is used for:

Determining the entire scope and shape of the piece

Determining depth of flaws to be removed by sanding

That latter may be a fixed amount (operator controlled) based on the initial entire piece scan.

Perhaps I am not understanding the issue, but if you have the part geometries predefined, it doesn't really matter how near or how far the unmodified part is to the finished geometry. A CNC machine will cut it to the geometry that is predefined via the CAD/CAM program. The measurement probe, be it small point contact or laser scan, could provide verification of the finished item.

A similar methodology is used in the manufacture of orthopedic prostheses (e.g. knee and him implants) which are somewhat customized to patient-specific dimensions.

If I am misunderstanding the question, please let me know. I do have close to 40 years of experience in the design and manufacturing of prototype precision parts for aerospace and biomedical applications, much of this via CNC and CAD/CAM, and perhaps I can provide some other suggestions.

You know that thinking is not my strong suit. :)

I think that would be operator set. Point the laser at a known perfect surface and that is the standard. Any surface farther away would be defined as a pit or scratch.

Will add this to my research, thanks.

I'm probably not being clear in how I define the original parts to be sanded. They are parts that are 50 years old. They were originally manufactured to be identical but were created and modified by hand, slightly, when new, to all fit together. And they have been in use for 50 years subject to all kinds of stresses which means that two of the same pieces can be pretty far off in terms of overall dimensions.

So any machine acting on them needs to know the exact shape and dimensions of that piece vs. another "identical" one next to it.

Does that make sense?

How big are the parts? Small, like Porsche letter emblems?
Is this a one time deal?

long and thin and some are bent and creased in complex shapes. This a many time deal. I wouldn't even blink if I put $30K into the project. I hope that doesn't sound like a naively small amount. I have no idea what something like this would actually cost.

I was thinking if the parts were small (like porsche lettering or smaller), and it was a one time deal, befriending a dentist and using their digital scanners, and then find a machine that could read the files and mill it out. I think the files are stl.

I don't know how these scanners would read longer objects, but I don't really know.

Older equipment can be found cheaper, probably not what you want, but, more just thinking out loud. Good luck.

edit- now youve got me googling the accuracy of the dental scanners- Values ranged from 16.3 [2.8] µm (CO) up to 89.8 [26.1] µm (OC4) for in vitro trueness, and from 10.6 [3.8] µm (CO) up to 58.6 [38.4] µm (iT) for in vitro precision for the complete-arch methods. They seem to handle small stuff really well.

edit two- that might be a good thing to research on the milling end also- what it the ability of different machines to cut to the specific tolerances you need.

This is a tougher thing to automate than it probably seems. It is multidicipline for mechanics, machine learning, and metrology

There are a lot of ways for a machine to physically measure a part, Photogrammetry, Laser scanning, direct probe. Hopefully the basis for the sanding or machining of the parts already exists for adaptation or it could get VERY expensive to develop.

I don't have any great useful advice and I don't understand the requirements well obviously but my experience is that this sort of thing is very expensive to do well, but the only way it can be economically feasible is if existing machine platforms can be adapted.

Another thought is that if you have a desired final shape a machine can be used with open loop tool passes programed to the right geometry, maybe on the low material condition side of the tolerance. That way you can knock down high spots to an acceptable geometry.

If a machine needs to recognize qualitatively unsightly defects and make decisions on how to blend them into the rest of the part, it could be quite an ambitious project.