So I've been thinking (again, dangerous as you all know) about technology. Apple some time ago started adding lidar scanners to its iPads and iPhones. And to that effect, I had thought of using the technology for work when we have to measure out houses for ductless systems. The technology is proving so far pretty spot on for dimensions of rooms and whatnot (even though there is still a learning curve there), but I do know that it can be used to scan objects. Now, a method of doing so would certainly need devising and testing, and then you need to be able to clean up scans on the other side... But what's to say that this technology can't be used to aid in replication of rare or nearly impossible to find components? I'm not sure how well it would do to scan a worn out mechanical component, but there are plenty of other things you could easily scan for replication. The main restriction would be hardware and technical capabilities of whoever implements it, but personally? I'm thinking it's entirely doable. Just a matter of figuring out the learning curve... Thoughts anyone?

Not sure Lidar is what you need to scan a part for creating a replica, but there are programs out there specifically made for that purpose. Scanning a worn part than manipulating the data to produce an unworn version - why not? Given the options of materials available for 3DP right now, I think creating a lost wax type pattern to be packed in green sand and cast in molten appropriate metal makes sense to me.

In regards to scanning... you get what you pay for. Would a point cloud scan from my phone work? for some applications yes. For many? No. (been there done that) At the moment I have a scanner for our lab spec'd and quoted and waiting for the funding to be in place. At $67,000.00 its not cheap! Even at that price the accuracy is +/- .0009" For most applications we encounter it will work just fine but again.... to produce a precision part from even a scan of that quality takes some work in a 3D modeling program such as Solidworks or Inventor - its not scan and forget and those programs are pricey.

3D scanning is a wonderful tool for Quality control and reverse engineering (when applied to the correct application) but it still requires a considerable degree of work to transfer the raw, scanned model into a finished 3D model free of blemishes etc. and again meets design intent and required tolerances. I view a 3D scan as I would a rough casting. i.e. there is still a lot of work to be done.

If your paying someone to do that work than its going to be costly. Like any other service, before you jump do your research. There are many people out there offering these types of services. However, the quality can be patchy to say the least. They have to have a solid understanding of your expectations and likewise, you of their expectations and capabilities. Unless you can do the work yourself its not simply a matter of having a part scanned and simply printed. They have to be capable of determining the design intent (i.e. how the part interfaces with other components) determining the type and class fit and calculate the appropriate tolerances etc. and again, provide proper shop drawing complete with all GD&T. If they cannot or unwilling to meet these basic requirements than you need to look elsewhere.

None of this is intended to discourage but to help to focus expectations. Especially when we are frequently told that 3D printing and scanning is the "be all to end all.

As for 3D printing... wonderful technology for printing foundry patterns - Lost PLA or traditional. But again, if you start with a scan it has to be manipulated in a 3D modeling program to adjust for the shrink rate, machining allowance, draft and of course any blemishes or damage.

Attached is a sample rendering from a project we recently worked on for a museum. They need a set of castings for the skis for a Lombard tractor. We reverse engineered the components referencing surviving originals. We used Solidworks to model the components and develop the shop drawings. Next step, the 3D part files will be sent to the foundry where they will be used to generate 3D printed sand molds which saves a ton of money over traditional patterns and core boxes. Albeit, the 3D printed molds are a one time use.

Yeah, what Terry said. High definition 3D scanning can be a wonderful tool. However, the costs can be painful especially for a one-off or limited run of parts.

Case in point: A group of railfans/model railroaders were bemoaning the lack of FM H20-44 models available. This conversation had been beaten to death over the years so I contacted TruePoint Laser Scanning in 2018 for a ballpark estimate to scan in high definition the one in Galveston. It’s not far from them so there’d be no travel costs and it was similar to something they’d been involved with before. The ballpark was $4,000 and that would get you a thumb drive with an AutoCAD 3D mesh file. Now you have to find a someone to import that file into 3D AutoCAD, clean it up and work it into a useable format in the proper scale for 3D printing. That you’d pay hourly for.

Can you do it? Yeah, I think so. But be prepared for sticker shock even for smaller parts.

Its like any other technology be it 3D printing, CNC etc. its all about using the correct tool for the job.

Here is another example: We were tasked with replicating a radiator cap. All we had to work from were dimensions pulled from the radiator fill neck and one vintage photograph. With the filler neck serving as the known dimension we were able to import the photo into Autocad and scale it to the correct size and use it to pull further dimensions. This gets a bit tricky because you can end up with some very odd dimensions so we had to rationalize. For instance if the dimension pulled off the photograph was say 1.512" we fixed it at 1.5". Our rational being that they originally used common precision - 1/2, 1/8, 1/16 and so forth. Features such as for example a bored hole that would require a running fit for a shaft we would determine a nominal size and than calculate the fit and tolerance (Machinery's Handbook is your friend!) to maintain design intent.

Next, using Solidworks we created the 3D model and 3D printed a mock-up of the cap and locking knob to test the fit etc. Our original plan was to have it cast so it was a fairly easy step to modify the Solidworks models (adding draft, machining allowance and scale for shrinkage) to creat and 3D print the foundry patterns. However, time was working against us so we explored 3D metal printing. Again, I treat a 3D metal printed part the same as a rough casting. You just cannot use it right out of the box - it has to be machined as required. Unfortunately, the size of the part and the aspect ratio made it a very poor candidate for 3D metal printing.

Pressed for time - we ended up CNC machining it from solid. If we had been charging the client that radiator cap would have cost over $2,500.00 between cost of material (lots of waste) engineering and computer time as well as CAM and machine time. Our first choice - having it cast using the 3D printed patterns would have been a fraction of that with less than an hour of machine time to mill the seat and bore and tap a hole.

Again, its using the correct technology for the job. Now if you want to explore some of the technology... AutoDesk offers a stripped down version of Fusion 360 for free to students and home hobbyist. We actually use the full version for the CAM package. 3D printers are fairly cheap. We have 3 Creality CR10 printers in the lab and they go none stop.

Yes, one should use the right technology for the job, but in reading up on the Apple LIDAR capability it is quite impressive and will only get better. With the right app, it could probably be used quite successfully in scanning of parts. For the moment, I see it more as replacing the tape measure in many cases which could be quite useful in some of our more challenging restoration projects, replicating complex wooden components as one example. Reminds me of the oscillating saw that in many ways has replaced the hammer and chisel. I suppose its time for me to get a new iPhone.

So through work, I did get one of the iPads with the equipment. Thank god for rewards programs that we tend to totally forget about until we're told we need to use the points or loose them. I'm finding that with measurements, they're generally accurate when you're square on a target. In fact, normal four sided dimensions it can pick up on automatically. Perfect example was the window on my office door. Picked up the edges of the frame itself and pulled dimensions from that, and it was spot on. Some of the scanning apps as well do show promise, but it comes down to learning how it works and methodology. I've grown up around technology, and I'm finding this to have a rather unique learning curve considering that it's so new. Even if it's not there just yet? I could see this stuff being there within a decade to be completely viable as an option.

For many parts a surface plate and height gauge along with solid works creating a model goes fairly quick. The parts we are dealing with are usually in even fractions variations are usually in the shrinkage of a casting or built in draft.