A 3d Scanning Project

There's a plaque I want to produce copies of it.
I put together a mobile scanning setup:

Laptop
3d Scanner (s)
AC Powerbank
3d scanning spray
Nylon bristle brushes

The object I was looking to scan is a 13hr drive away. I rented a place nearby to serve as basecamp.

Here is the subject:


It measures ~20" x ~13"
I have both a Creality Lizard and Revopoint MINI 3d scanner. The lizard is reliable and has good tracking but loses some detail, especially on the smaller text. The MINI can capture far more detail but has more tracking issues, it has a much smaller FOV. I brought both & attempted to use both. The subject required application of 3d scanning spray to make the scanners happy.
The lizard got me a reasonable scan.
The MINI fought me and could not keep tracking. The scans I have may be usable, but there's currently no way to throw out mis-tracked frames like there is with the creality software. (i did not even need this feature with the Lizard scans) It did grab significantly more detail in the areas it kept tracking, which is frustrating. The lizard was able to give me a usable mesh with little to no editing.

The mobile setup as far as support equipment goes performed nominally as expected based on testing. Basecamp nearby was excellent, as it gave me several opportunities to travel to the object, scan, and return to process & assess the captured data

Very frustrating to travel that distance and have the REVO software not perform in the field. Practice objects of the same size & shape gave me good results, but in possibly better scanning conditions

Here is the captured mesh as seen in JMstudio:


I tried 3d printing some versions of the data:
Resinprint on Elegoo mars2pro:


FDM print on ender3 w/ .6mm nozzle


Resin print came out awesome, but it's small. FDM print shows a lot of layer lines, I could tune to eliminate this but details are still muted

I am able to load raw mesh data into Vectric Aspire for 3d carving via CNC router. It does not need to be manifold mesh. This subject is a perfect use case scenario because it's a flat bas-relief carving that is already appropriately drafted. I scaled it roughly 50% because of 1.) time constraints 2.)wood stock size for pattern on-hand 3.)volume of metal required, crucible & raw stock vol are small

Vectric's default settings simplify imported meshes, which I did not initially know. The first cuts of the finish pass showed a lot of artifacting on the edges and a rough suface. I figured out how to import it at much higher resolution and re-cut the finish pass again (another ~14hrs) Finishing cutter has a 0.5mm (20 thou) radius tip cutter so it can do fairly fine detail but takes foreverrrrrrrr (pass stepover is 4 thou)

Initial results:


Sanded and sealed:


I then sand-cast the wooden pattern in Silicon Bronze. Sand is 130mesh greensand. Raw casting:


Initial cleanup:


Chemical darkening applied, highlights buffed, laquer topcoat applied:


Final thoughts:
3d scanners are fucking awesome, but also frustrating AF. This is a as far as I am concerned a near perfect use-case for this technology, capturing a mesh of a flat, pre-drafted object and being able to digitally reproduce it with no editing or reverse-engineering of the mesh. Most of the other objects I have 3d scanned are mechanical type parts to be modified, which requires reverse engineering. The preliminary results for this project did not require any editing of the mesh, except to trim off the excess around the edges.

The project will still require some digital editing before I am wholly pleased with the result, but for now this is very acceptable

-Rocketman

 

Nicely done!

 

The CNC router is a 24"x 24" unit made by Geometric Robotics
https://www.ebay.com/itm/133697164060
I purchased it in 2012 and have made some upgrades, it's not the most rigid machine but it's acceptable. Significantly cheaper now compared to when I purchased it. Someone else here has the same model as well

Thank you for the kind words, I have hundreds of hours accumulated into this project and it's hard to sum it up into one forum post

 

Very nice work!

Given the huge amount of work involved in making this piece, do you think the wood substrate you used as a pattern material is ideal? It is beautiful in its own right. But, it is wood with all of wood’s foibles. Something like Renshape 4000 is supposed to be absolutely uniform and stable. I’ve not used it and I also understand it is very costly. You have the CNC file, so I guess cutting a new one would be straightforward.

Denis

 

Yes - the wood is not ideal. I have looked into HDU (sign foam), renshape, machinable wax, etc. The wood for this scale piece is fine for now, I would eventually like to make a full-sized reproduction and that will probably need one of these options. Right now I am in the "rough draft" and "damn I spent a lot of money already" phase of this project, I'm not sure where I am headed with it. FAFO, I suppose?

The current angle I am looking at is to resin print a similar scale piece as I have above, but it will have to be printed in at least 4 sections and joined as my print volume is small. It will certainly give better resolution and detail compared to the CNC, but how will it look once it comes out of the sand? (should I even be sand casting this?? the investment casting voices are getting louder) The CNC has a way of naturally smoothing noise from the scan data. Either way I am wanting to cope the back side of the pattern in the 3d model (it is flat-back currently) to save on resin and metal. My 3d modelling/editing skills are not great so that is my roadblock at the moment, as well as a lack of time, next month + is very busy for me.

But regardless of which way I make the pattern, I will likely use Freeman Repro to make a negative>positive to generate a nice durable pattern that should be stable over time. The stuff is very good in that regard.

-Rocketman