Our Collective 5-Foot Millennium Falcon Build

Nuthin' that 9.9 million data points can't pick up.

The fun part will be resizing it for your printer so it doesn't take a week to print.

Whoever tweaks it, please repost your tweaked file back into the source file so that not everyone has to replicate the work.

And please show pictures of how your print came out -- these will not be as perfect as resin-cast or originals, obviously, but I'd like to see how "close" we can get.
 
Clarification: by "resizing" (in the post above) I mean resizing the data file (i.e., simplifying it, reducing the massive size of data points in it), and NOT resizing the physical measurements of the scanned object -- these will all be done at 100%, or 1:1 scale.
 
Congrats and thanks Studio KB for your troubles and frustrations getting things set up for the 3D scans. It sounds like it was quite a process climbing that mountain but we will all be rewarded with a great view when we get there hahaha!! Ten million data points for a single greebie - the subassemblies?? (shudder)
 
I was wondering how low on data points one can go before detail loss is noticeable. This could and would help greatly in reducing file sizes.
 
I was wondering how low on data points one can go before detail loss is noticeable. This could and would help greatly in reducing file sizes.
That's an interesting thought for the end user to perhaps figure out and I imagine it has alot to do with the user's software and 3D printing set up. I can't speak for Studio Kitbash but I think he is intending to provide the files at a very high fidelity so they will be larger files initially.
 
It's a great question, and I don't know the answer yet. I'm hoping to find the perfect balance between "highest print quality scan" and "lowest possible number of data points" -- anyone with 3D scanning experience is welcome to chime in, as the learning curve is fairly steep and moreover, the high-data-point scans are extremely time-consuming.
 
I can't say that I have experience printing from scans, only parts modelled from scans.
However from my limited scanning experience I would suggest that it is extremely variable, and would very much depend on both how detailed the subject is, and its size. For a simple and small shape you will obviously need fewer mesh vertices. However the more vertices the crisper the edge detail.

From what I have seen, a scan result is quite different from a CAD modelled mesh geometry, as the scan result mesh is typically uniformly created during the solve, where as a modelled part often optimises the mesh to be more dense around detailed features so as to optimise the file size.

Hence a much larger raw scan file size is needed to duplicate the result of a modelled CAD mesh.

I would however suggest that scanning, merging and solving should be much faster than manually modelling in CAD for most parts. Especially for such a complicated shapes as the Entex rotary box. And could be considered more accurate depending on the resolution.

I think with experience you will soon get a feel for how detailed your scans need to be. And possibly a good work flow to speed it up.

Really looking forward to seeing more results as they appear!
 
It begins...

5-foot Millennium Falcon beginning layout.jpg
 
You have to work out your own salvation with fear and trembling....

So for each subassembly, there's always the question of the measurements, and you can either a.) trust someone you ask, b.) go with somebody's blueprints, or c.) figure out yourself. I typically choose d.) all of the above, in comparison to the neighboring greeblies, in order to see what looks perfect.

So on this turret ring, there are two schools of thought:
A.) 12" base, with 11.5" upper ring
B.) 11.5" base, with 11.0" upper ring
C.) Maruska's drawings, which favor B.
D.) My non-computerized, non-algorithmic interpretation of the Japanese Star Wars Chronicles book pictures of the 5-footer Falcon, using their rulers in the picture as the ruler to apply to the rest of picture, actually gets you closer to somewhere between 10-3/4" and 10-7/8" for the upper ring. IF (big if) this was a one-piece disk that was trimmed down on a circular lathe, then it would make sense to have the bottom be a clean 11.5 and the top be whatever was left after all the angular trimming. That's the way I'm currently leaning in my interpretation, and without a lathe, I am trying to basically have "all options" on the table so that when I get to placing this on, I can make a definitive judgment of which one looks/fits best in relation to the side docking tunnels, the front jawbox, and the rear engine deck.

So long story short, what you're seeing on the picture is a combination of a.) Glowforge lasercut rings, b.) Shapeways versions of the ring turret, and c.) a big fat 12" ring that is now categorically "off the table" unless and until the overall geometry proves otherwise.

Part of how I judge/assess/discern this is to try and get inside the head of Joe Johnston, the original designer, and the other builders at ILM. Johnston had a wonderfully "asymmetrical symmetry" pattern for most of his designs, and on the Falcon this is worthy of its own separate (coming soon) post. But one thing it meant was that the front mandibles had an outer dimension (from outside left to outside right) of 11.5", which would have been visually "balanced" by an 11.5 bottom ring, tapered down to whatever was left after the lathe did it's work. But if I'm wrong on the lathe approach to the turret ring, then it's entirely possible that it was a 12" ring at bottom that was tapered down to 11.5" on top in order to achieve this same symmetry between front and center of the ship. This is how Lee Malone built his, and it looks amazing. Sean Sides did 11.5 on bottom and 11.0" on top, and his also looks amazing. So again, it's a parts-to-the whole and a "work out your own salvation with fear and trembling" approach that MUST be taken because in the end, only you have to live with the end result of what you've decided.

In no case should you let someone else "do the work for you" and "trust their judgment" simply because they've done it first. You have to be happy with the ship you've built, and you have to own your own work. Or at least, that's my philosophy, one reason I'm still not happy with the Y-Wing armature, and still working with a colleague in CAD to perfect that substructure.
 
Couldn't agree more. You are definitely going about it the right way.

I see it as like a detective trying to solve a mystery or puzzle, by pulling all the clues together and using all the techniques at their disposal.
I think to me thats what makes it most appealing. One big giant puzzle...

Well done!
 
Okay, so here are the PRECISE cut lines you want for the Airfix Hurricane:

Airfix Hurricane Cut Lines 1.jpg

As you can see, the two horizontal cuts (1 on top, 2 on bottom) follow natural lines on the model itself, and are fairly straightforward. But the four angled vertical cuts are trickier, and unless you have 3 Hurricanes to experiment and fail on, you're going to want to get it right the first time.

Airfix Hurricane Cut Lines 2.jpg

So the bottom most vertical cut lines simply extend the line implicit in the landing gear wheel well, and using a ruler to guide your knife or scriber, you can do this pretty easily. When you get to the big wing cut, however, you're going from outside light emplacement corner to the third (3rd) scribed panel line of rivets where the wing begins on both outer edges. See next pic for close up of this particular "spot" that you'll want to mark to get your angle just right.

Airfix Hurricane Cut Lines 3.jpg

This is the right side, up close. From the far outer edge (of where the aileron would be), count inwards on the vertical rivet lines: 1, 2, 3. Mark the 3rd one in as the final arrival point of your scriber, and you'll have the perfect angle for the rear deck engine bottom portion subsection.
 
Awesome work Studio Kitbash! This part is good value. Heavily used on this ship and other SW models.

Hammer3246, Casting would make good sense. Or 3D print.
And casting the 8rad would be accurate given how the ILM team seem to have cast most of these for this model and others.

Make that 14-15x by the way, in 16x locations. ;)

From what I can tell there are:
8x small rear pieces on the upper and lower aft engine deck,
2x full cast pieces on the upper and lower inside jaw boxes,
2x full pieces on the inside mandibles,
1x forward piece on the port outside mandible,
1x full long rear piece in the upper starboard Hetzer pit,
1x full cast piece in the upper port 8Rad / Kettenkrad pit,
1x full cast piece hidden under the lower starboard hull damage. - thank you t2sides for the excellent reference! (without your thread I had no idea.)

Looking forward to the day when I can compare this scan and see how close I got with my .stl derived from simple measurements!
 
Last edited:
Bjorn -- Good call! I thought the inside mandible pieces were not "full pieces" but were just sections abutted into the 57 Bel Air chassis. And I forgot about the hidden piece underneath the lower starboard hull damage, so thanks for the reminder.

Hammer3246 -- "Cast your own" is what I've been doing, and I can't tell at this point whether casting is "cheaper" than printing overall, in terms of cost, compared to casting. But in terms of acquire the kit, make the modifications, make the mold, pour the mold, set the mold, mix the resin, pour the resin, let the resin cure, then pull the casting from the mold, then pull 16 more "perfect" castings from this same mold, which is hard to do because there's a failure rate with castings, and because the mold breaks down fairly quickly at this level of use (esp. with high-detail, complex molds), I think the "print out on 3D printer" option may simply be "easier." I'm of a mixed opinion, because castings are still to my eye much more accurate and detailed/faithful replicas of original pieces, and until I learn CAD software to clean up/perfect every scan, or until the software gets so good that it can make perfectly accurate watertight meshes of these complex things, then the old 1970's casting route is still my preferred mode of building.
 
1594695760830.png

Schematic for building using the girder-riser system to make your domes. Girder bottom length is 17.75" and end height is 3.75" (But girder bottom length should be an even 18")

Big patty is 46" even. Circle for turret ring is 11.5"
Center mounting hole is 2.25"

Remaining measurements in the following image:
1594696437781.png
 
Last edited:
View attachment 1324991
Schematic for building using the girder-riser system to make your domes. Girder bottom length is 17.75" and end height is 3.75" (But girder bottom length should be an even 18")

Big patty is 46" even. Circle for turret ring is 11.5"
Center mounting hole is 2.25"

Remaining measurements in the following image:
View attachment 1324993
Amazing squematics thank you for sharing!!!
 
Last edited:
View attachment 1324991
Schematic for building using the girder-riser system to make your domes. Girder bottom length is 17.75" and end height is 3.75" (But girder bottom length should be an even 18")

Big patty is 46" even. Circle for turret ring is 11.5"
Center mounting hole is 2.25"

Remaining measurements in the following image:
View attachment 1324993
The cockpit tube Dia. is 5.657? What are the dims you have regarding this part? Thanks again for the drawings :)
 
Last edited:

Your message may be considered spam for the following reasons:

If you wish to reply despite these issues, check the box below before replying.
Be aware that malicious compliance may result in more severe penalties.
Back
Top