Decorum for the Forum:
  • Be nice. If you want to be mean, try Reddit.
  • No Piracy. If you want to be a thief, there are dark places on the internet dedicated to that.
  • No Cracking. Discussions on AnyDVD, DeUHD, DVDFab, UHDKeys and similar tools are not permitted here.
  • No Spamming. If you want to make a buck, work smarter... somewhere else.
  • No Adult Content. Half the internet is dedicated to adult content. This half isn't.

Privacy Policy: Click Here to Review (updated September 30, 2020)

[DIY] Jack Sparrow's Compass (Work in Progress)

A place to share your purchased and home-made movie props and replicas. Full-scale, life-size replicas only - figurines and scale models not welcome here.
Post Reply
User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

[DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Sat May 15, 2021 7:59 pm

Jack Sparrow's Compass
Recently I was browsing Thingiverse for movie props and came across Jack Sparrow's compass from Pirates of the Caribbean. In case you don't recall, this is the compass that points towards what the owner most desires, and not true north.

image.png
image.png (507 KiB) Viewed 12517 times

While I've long wanted to recreate the treasure coins from The Curse of the Black Pearl, I'm holding off until I get a resin printer someday, as those have much higher quality for finely detailed objects. My current melted plastic squeezer... er, I mean plastic filament 3D printer... would leave a lot of layer lines that would be challenging to sand out without destroying the intricate details on the pirate coins. But this compass looks like a prime candidate for 3D printing in plastic, with nice large flat sections and almost no intricate details.

While checking out the compass model on Thingiverse, and comparing it to screenshots from the film, I quickly realized that the model was not accurate. For example, here's the main compass body - notice in the details that it has deeply inset panels on a thick solid frame, the triangles in the corners extend to the edges of the octogon, and the hinge extends the full width of the edge:

image.png
image.png (461.38 KiB) Viewed 12517 times

But in the movie scene, the panels are fairly flush with the frame, the frame is segmented as if it was assembled from smaller segments, the triangles are smaller and don't reach the corners (and seem thinner too), and there is a noticeably smaller hinge:

image.png
image.png (1.68 MiB) Viewed 12517 times

I also had to consider how to actually print this model. I really hate using supports, as they mar the finish and require a lot of post processing to clean them up, and with the way this model was designed, there wasn't a good angle to minimize supports.

So I decided that if I was going to bother with this prop, I was going to have to recreate the design from scratch, so I could achieve the level of realism I wanted, while making design changes to make printing easier without excessive support structures.

NOTE: After much research, I see that the compass design is not consistent even in the Disney movies. Some have the larger triangles. The face decal (called a compass rose or windrose, or "Rose of the Winds") has two different designs. This is quite common in movies, where props change from movie to movie, or even in the same movie (Indiana Jones and the Raiders of the Lost Ark have two medallion designs in the same movie, made by two different prop houses, and used for different scenes). Of the various designs, I prefer the look of the smaller triangle design shown above, so that's my target.

Also, before starting down this path, I like to check and see if an official prop already exists for purchase. I found a lot of hand-made examples that didn't even pass the glance test (one glance and you knew it was fake), and even the good ones still looked like painted plastic. The more I thought about creating my own, the more I preferred this path, as it allowed me to integrate some alternate materials like real wood. Perhaps I could even make this the nicest replica of them all.


Beginning the Design
The first step was to get the dimensions right. I started with the screenshots showing the compass being held in-hand, and positioned my hand in a similar pose, and measured between my thumb and fingers. Obviously hand sizes differ, though with Johnny Depp being about the same height as I am (he's 5' 10", I'm a smidge less than 6'), I figured my hand was similar in size to his and would be good to measure. I determined that the compass size is likely between 80mm and 90mm, probably close to 84mm by the looks of it. Also, the creator of the Thingiverse model had graciously provided the Fusion 360 model for download, so I examined that one and found it was sized to 82mm.

Next I needed to determine the size of the fileted corners. I did this by measuring some screenshots and plugging the values into a spreadsheet to come up with a general ratio of side-length to corner-length. I found that the corners were approximately 55% as long as the sides. I then created another formula to plug in different lengths and see what came out. I put my engineer's hat on for this task - and by this I mean I went at this as if I was designing a prop from scratch with no specs, and used nice round numbers. For example, I figured it was more likely that the side was a rounded 47mm long, and not a fractional 46.8423mm.

By plugging in nice round numbers into my spreadsheet, I found that that if you divided a square into a tic-tac-toe grid, with the central section exactly 47mm wide, and the corner sections exactly 18mm wide, that you ended up with an overall size of 83mm. And the 18mm wide corners create a diagonal that is 25.5mm long, approximately 54% of the 47mm length. Obviously these numbers are still just guesses, but if you change them by even 1 mm in any one direction, you get a significantly larger/smaller model or throw off the ratio of corner length to side length, so I went forward with the dimensions I had calculated.

In the design, I placed fairly large, exaggerated gaps between the various metal frame segments and wood panels. Below you can see my design in progress on the right, with the original Thingiverse design on the left. I made the gaps between the wood panels and metal sections larger than the gaps between the metal to metal joints, as I figured that more closely resembled real-world manufacturing techniques.

I also again used nice round'ish numbers for sizing the metal frame. The circle I sized to 4mm wide, the edges to 3.5mm wide, and the corner triangles to 3mm wide. I think I was right to guess that an engineer designed this in CAD before making the model, and likely used similar if not identical numbers in their design. It's the same thing I do when designing parts, as I prefer seeing round numbers in a design versus long fractional numbers. Ditto with the circle diameter, which I sized to 70mm, a nice round number that also cut into the triangles quite similarly to the actual movie prop.

image.png
image.png (780.56 KiB) Viewed 12517 times

I then printed out a quick sample to see how well the gaps looked in plastic. I put a little paint on the metal frame portions to give it some pop. I like to do partial sample prints to test out design choices, you can save a lot of wasted plastic and time versus printing a full model. This sample only used $0.12 worth of plastic and about an hour to print. Had I printed the full part, it would have taken closer to 6 hours and about a dollar worth of filament.

image.png
image.png (1.36 MiB) Viewed 12517 times

And for easier side-by-side comparison, the movie screenshot:

image.png
image.png (1.68 MiB) Viewed 12517 times

I felt I was on the right path, and that the result was not bad for an hour's work. Holding the part in my hand, recreating the grasp shown in the movie, the overall size seems really good. Happy with the new design, it was time to think about the overall design plan, which would incorporate real metal and wood.


Next Steps
I'm very early in the design process, and can't get too much further until I get some materials ordered and delivered. I also need to watch some of the movies to find more compass scenes for design reference.

I find that plastic painted to look like metal still looks too much like plastic. Instead, I am thinking of using some metalized tape to cover the frame segments. I would really like to use a brass metalized tape, but I can't really find such a product. It seems I can either use aluminum/silver metal tape, or copper tape. I'm worried that the copper may look too orange - some of the film shots look more silver than brass, and definitely not coppery. So I'm leaning towards using a silver aluminum tape, and weathering it with paint to give it an antique look, possibly with brass tones.

Once I get some metalized tape, I need to apply some to my sample prints and see if the large gaps are in fact too small - I fear that the tape will be so thick that it makes the gaps fill up. I plan to adjust the gaps larger if needed to account for the tape's thickness.

Then I will finish a preliminary model design that has separate, non-printed panels for the wood sections - and here's what's really cool: I plan to have the wood panel sections laser cut in Walnut veneer by Ponoko. Since the real-wood veneers can have a significant variance in thickness from sheet to sheet, I then have to wait for the laser cut walnut panels to be delivered before making final design tweaks to the model. I will adjust the depth of the panel cutouts to match the real veneer thickness, and also adjust the wood-metal gap to get the right inset effect to match the real movie prop. I figure it's easier, quicker and cheaper to adjust my printed model to fit the actual cut veneer sections than the other way around.

I'm super excited about making this a true multi-media build, with real wood panels and metal foil hiding nearly all the plastic. Additionally, I will have a small bit of real leather cut to go inside the lid, which is a detail you can see in some movie prop versions. Once this is done, I think the only visible plastic portion of the compass case will be the dome on top.

Stay tuned for updates...
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Sun May 16, 2021 2:19 pm

For those that know me, it should come as no surprise that I'm a huge fan of Adam Savage, of Mythbuster's fame. Adam has a YouTube channel on which he shares awesome tips, cool props, and "One Day Builds" that rarely take just one day.

Adam used to work at ILM creating props for some of your favorite movies, and he has shared a lot of his industry know-how in his many videos. For example, he explained about greebles and nurnies, which are small components, often from plastic model kits, used to add detail to a custom model. A good example is the outside surface of the massive starships that seem to have impossible levels of detail - those are just random bits of plastic parts from off the shelf model kits repurposed for the greater good.

Another tip I learned from Adam, especially when he is talking about movie prop replicas, as that prop designers don't fabricate what they can simply buy. The most obvious example of this is something as basic as a screw or bolt. Knobs are another good example. Recently Adam mentioned that he had found the original source of a knob used on a prop - Chewbacca's bowcaster if I recall correctly.

It's tips like these that really altered my frame of mind when doing prop replication, and I'm applying this way of thinking to the Compass of Destiny (which, to the best that I can tell, is the true name of Jack Sparrow's compass).

At first glance, Jack's compass looks to be a completely hand-crafted, one of a kind piece with no off the shelf greeble. But on closer inspection, a few details pop out. The hinge, the hasp, and the hook (otherwise known as the three-H's). If you were a prop designer, working on a tight timeline, these are not parts you're going to fabricate from scratch. No way. Rather, you'd just find something in a store. With any luck, I could find the exact same parts...


The Hinge
I started with the hinge. Based on the measurements I calculated earlier, I determined that this hinge is likely somewhere in the 25mm-27mm range, and since this prop was likely made in America I figured that meant a 1" hinge.

image.png
image.png (531.7 KiB) Viewed 12514 times

There seem to be precious few scenes showing the back side of the compass, though I did find one in Dead Men Tell No Tales. While this might the the result of multiple props being used with different features, I find it interesting that the hinge brackets are not visible on the back, nor the inside - somehow they are hidden inside. Most likely this design choice is actually the hinge being part of the metal frame. This does pose a problem for me if I want to use a metal hinge - perhaps I would be better off including this as part of the 3D print, though I have serious concerns about painting it to look like metal, and this appendage would definitely complicate printing. Perhaps I can create a slot in the model, and slide the hinge inside?

image.png
image.png (703 KiB) Viewed 12514 times

So I'm starting to think that this hinge might not have been store bought. You could easily create a hinge by cutting 4 segments of brass tube and inserting a pin, and that could be soldered easily enough to the metal frame. This could also explain how the hinge barrel protrudes from both the back and the inside, a feature you don't see on typical butt style hinges which have one side flat.

While I could do something similar, I don't have an actual metal frame onto which to solder the hinge, so this idea doesn't appeal to me. Instead, I found these antique butterfly style hinges on Amazon that I could easily bend and cut into shape, leaving tabs to insert into the model to be glued in place.

image.png
image.png (272.49 KiB) Viewed 12514 times

They're a tad long at 29mm, but I'm not sure if the extra 3mm will be noticeable or not. I plan to cut it down to just have small tabs, like this:

image.png
image.png (73.92 KiB) Viewed 12514 times

The Hasp
The Hasp took a lot of searching, using various terms like clasp, latch, and hasp. From the picture, I only knew the basic shape of the latching mechanism, that it appeared to be pretty basic and cheap, and that it was approximately 1 inch wide, similar to the hinge.

I started looking on Amazon, but found nothing I liked there. I expanded my search to Ebay, AliExpress, and Etsy, and still nothing that looked like the latch used on the compass of destiny. During all of my searching, one thought kept nagging me - all of these online sources were quite different 20 years ago when that first prop was likely fabricated.

image.png
image.png (552.31 KiB) Viewed 12514 times

So I had to break my habits and think like a wood worker making something like a jewelry box. Immediately I thought of Rockler, and realized their latch options might be pretty much the same all these years later.

Sure enough, not only did Rockler have a good option, I believe I have found the actual hinge used in the real prop!

image.png
image.png (139.96 KiB) Viewed 12514 times

The only detail that seems truly wrong is the little curl at the bottom of the clasp, though that can easily be cut off. The scroll detail looks slightly different too, perhaps that's changed over the years, though unless I can find an in-focus picture of the scrolling on the actual prop, the Rockler design is good enough for me.

image.png
image.png (184.74 KiB) Viewed 12514 times

Actually, from a slightly different angle that is a little more in focus, I have a idea that they either simply put the latch on upside down, so we see the undecorated flat side, or they ground the details off altogether to make it look more basic. I've gone ahead and ordered it, fingers crossed...


The Hook
Jack always has the compass on his belt, with a leather lanyard attached to a eye bolt. This hook is another feature that is rarely revealed on screen, though this actual movie prop made of foam is perhaps one of the best examples I've found:

image.png
image.png (583.69 KiB) Viewed 12514 times

I didn't know what this part is called, so searching for it was quite difficult. First I searched for brass eye bolts, but those are mostly industrial and don't swivel (an important feature visible in the movies). Next I searched for brass lamp finials, and while some of these looked similar they also didn't swivel. But this led me to brass loop hangers (also non-swiveling) which in turn led me to cabinet pull rings (bingo!!!) and drop ring hangers (double-bingo).

But most of these were too large, and had different stem designs or ornate loops that didn't match the compass' drop ring. I've searched for way too long to come up with nothing. I did end up purchasing some small drop pull rings from Etsy, which are drop shipping from China so they're a good month away if not longer:

image.png
image.png (171.17 KiB) Viewed 12514 times

I might take a stab at 3D printing the stem and using the brass ring from the ordered part. If I had a metal lathe (one day I hope...) I would just fabricate my own in a few minutes. Perhaps that's what I'll ultimately do - use the ordered drop ring as a stand in until I get a lathe to make my own. Then again, perhaps I'll be happy with what I found on Etsy and leave it at that...
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Mon May 17, 2021 4:18 pm

Work on the Compass of Destiny continues, and I'm beginning to think I need to expand my material options. For example, can you even tell what this item is supposed to be?

image.png
image.png (733.6 KiB) Viewed 12510 times

If you squint your eyes, you might be able to tell it is part of a compass plate, with numbers running around the edges. This is how it would print in plastic if I used a 0.2mm nozzle on my printer. I normally run a 0.4mm nozzle, and running a 0.2mm nozzle can be quite challenging (clogs are a real problem with that tiny needle sized opening). And that screenshot above represents a best case result... it never looks that good in melted plastic.

Here's a screenshot of the original prop, with the brass compass ring in the middle:

image.png
image.png (979.82 KiB) Viewed 12510 times

And here's my extremely accurate Fusion 360 model:

image.png
image.png (504.07 KiB) Viewed 12510 times

Since this part is only 24mm across (a smidge under an inch), these details are just too small to render in plastic. So how else could it be made?

One way would be to do CNC milling of a piece of metal. This is probably how the prop department created the movie version. But I don't have the necessary CNC machine to do this.

Another way it to farm out the printing. Shapeways can make a metal version of this part by using lost wax casting. The print the part in extreme detail using wax, then melt it out and cast your chosen metal, like brass.

So is the resolution of Shapeways' wax printer high enough? Sure looks like it:

image.png
image.png (277.95 KiB) Viewed 12510 times

The wax ring above was printed by Shapeways, and the T in Technology is approximately 1mm tall. This is the same font size I used in the Sun Dial's compass above.

So having that part fabricated and cast in brass would have to be prohibitively expensive, right? Surprisingly, no. I just uploaded the model, and it would only cost $17.50 for the flat dial portion. Obviously the price will go up to include the arrow and the three support legs, but the cost should still be resonable. Considering that buying a cheap, printed plastic Jack Sparrow's compass "kit" (you have to assemble and paint) cost about $150, I find this to be a relative bargain and with amazing quality to boot!

So I've pretty much decided this is the way to go for the brass sun dial components. The only remaining question is do I have Shapeways print everything as one component (support arms, dial and arrow), or separate components that I later assemble. I'm actually thinking separate, as the overall quality may be higher due to more favorable printing angles, and I can use real world assembly methods like soldering since this will be real brass.

I've always wanted to try out Shapeways to see what a professional shop can deliver, but I've never had a good enough reason. Unexpectedly, the Compass of Destiny is changing that.

One thing I'll have to contend with is shrinkage (like a frightened turtle, hehe). Apparently the casting process can result in a 2.5% shrink in part dimensions. Since I need this part to connect to the compass box, if the shrinkage is too much it won't fit. Though I'm thinking one simple way around this is to design the arms extra long, and they can be filed down to fit. Alternatively, I can wait for the brass sun dial to come in, and then resize the printed plastic compass box to fit.

At this point, I'm convinced I will have the best Jack Sparrow compass in the world, probably better than even the real movie prop.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Tue May 18, 2021 3:38 pm

I finished designing the parts for Shapeway's brass casting. In addition to the sun dial support arms and the windrose hemisphere, I went ahead and modeled the drop ring to make it a near exact match for the original part.

image.png
image.png (825.32 KiB) Viewed 12502 times

Here's another angle of the four parts (the 3-armed support brace is a separate part from the sun dial that rests on it, for improving print quality):

image.png
image.png (911.87 KiB) Viewed 12502 times

I am a bit frustrated with Shapeways regarding their lost wax metal casting guidance. They indicate that parts can shrink 0.25mm + 2.5% of the part size. Which for a 100mm part would be 2.75mm of potential shrinkage. But then they give an example formula using 0.25%, i.e. 0.25mm + 0.25% = 0.5mm, which is a much smaller shrink. I asked their support team for clarification, and they simply said the info on the website was correct (completely ignoring the mathematical discrepancy I pointed out), so I'm really at a loss for how to design for the shrinkage. Most of the numbers mention the 2.5% shrinkage factor, so I added that much extra to the 3 support arms, to make sure they are long enough to connect to the box. If they end up too long, I will file them down to fit.

I also adjusted the compass font size a half size bigger, to give it a better chance of printing clearly, even though it is now slightly out of proportion to the original prop. You'd never notice this minor tweak without both the original and my replica in hand.


Shapeways Quote
With all my parts designed, it was time to see how much money Shapeways wanted to turn my CAD files into physical brass. $16.26 for the Sun Dial, $12.50 for the Support Arms, another $12.50 for the Windrose Hemisphere, and a pricey $27.50 for the Drop Ring.

image.png
image.png (170.22 KiB) Viewed 12502 times

Interestingly, if I got rid of the circular ring and only had the ring base printed, the cost dropped to $13.50, so that 1" brass ring adds an extra $14, which is too much for a generic part. Since I've already ordered a drop ring with a 1" ring, I can just have the base printed and borrow the brass ring from the one I ordered.

image.png
image.png (22.8 KiB) Viewed 12502 times

Shapeways adds a processing charge and shipping on top, so the final brass bill came to $68, which climbed to $72 with tax. While not cheap, I'm still coming out significantly less expensive than cheap looking replicas and kits I've found online.

Shapeways turnaround time (without paying $$$ extra for faster processing and shipping) is about a month, so it's going to be a while before I can share the results and wrap this project up.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Wed May 19, 2021 7:47 pm

Metalized Tape Research
My metalized tape samples arrived yesterday, so I have been able to do some testing with them. I got some copper tape (good for making electrical slug barriers in your garden), aluminum tape (often used in HVAC duct work), and some decorative washi tape, which is essentially fancy masking tape.

image.png
image.png (1001.91 KiB) Viewed 12497 times

I think it is pretty amazing how just a bit of metalized tape can give the look of real metal, vastly superior to paint in my opinion. Its an effect that is hard to capture in photos, it's much more convincing in person.

My primary objective at this stage was to determine if the panel gaps were large enough to remain visible after applying the metalized tape - I was worried my 0.3mm gaps would disappear. Somehow I guessed right with my original gap, and as you can see the panel gaps are quite visible while remaining realistically small.

My next goal was to pick a color. I knew going in that the copper tape would be too orange, and sure enough it met expectations, though it really looks amazing, like a freshly minted penny. I'm sure I will use this for other projects in the future.

I had very low hopes for the washi tape, as it is designed to be low adhesive and removable. As you can see in the picture above, it is already delaminating. I will say it worked better than I expected, and the light brass color is fantastic, almost exactly what I was looking for.

But after examining way too many images of the compass of destiny, I've come to the conclusion that the inlaid trim is silver in color, not brass, and it get's brownish tones from dirt and weathering effects. I plan to recreate this effect from various Rub'nBuff colors and paint weathering effects. So the aluminized tape wins.

I also learned that these tapes are thin enough that they reveal the uneven surface below. So I will need to do some sanding prep before applying the tape for best results.


More CAD Modelling
Now that I knew that the original panels gaps would work, the next step was to panelize all sides of the box. I also increased the wood panel depth to 3.2mm, since these will not be printed but rather cut from walnut by Ponoko. And since my hinges arrived, I took some measurements and added a slot in the design for it.

At this point, the lower half of the compass box design is essentially complete. Here's a render showing the full assembly, fully colorized to illustrate the final appearance before I apply the weathering effects:

image.png
image.png (950.15 KiB) Viewed 12497 times

Here's a render showing just the 3.2mm thick wood panels that I will have cut by Ponoko.

image.png
image.png (1.28 MiB) Viewed 12497 times

Here's a render showing the core box that will be 3D printed in black PLA and the exposed panels covered in aluminized tape. Ignore that the render makes it looks like it was milled out of aluminum, I was too lazy to change all the hidden sections to look like black plastic. The wood panels will be glued into the cutouts. Also notice that there are both cutouts and support shelfs for the sun dial's support arms.

image.png
image.png (944.79 KiB) Viewed 12497 times

I expect this main frame to be a difficult model to 3D print, as it has a ton of overhangs. I plan to use plenty of supports, and some judicious sanding during post processing to remove the support scars. Luckily I won't have to paint it, since the aluminized tape will cover almost all exposed surfaces except for the inner cylinder where the compass plate resides.

Below is the Windrose spinner assembly, which is composed of two printed halfs (bottom and top) that screw together through a 608 ball bearing - that's right, this thing is a fidget spinner! The secret feature here is that you will be able to spin the compass with your fingers by rotating the bottom circle wood panel, so this compass truly will point at what you want most (to point at...).

image.png
image.png (629.2 KiB) Viewed 12497 times

And one last glamour shot of the brass parts that Shapeways is casting:

image.png
image.png (809.35 KiB) Viewed 12497 times

Next Steps - Time to Top It Off
Obviously all the progress so far has only been on the bottom half of the compass. Now that it is complete, I will design the top half. I expect that it might see even more variance compared with competing designs, as the measurements I've taken from still frames has revealed that the lid is taller that those other designs, and that the dome has a different arc. I know that these are just minor details, but in the quest for realism (ironically, of a not real object) it's the little details that make the difference.

I also need to get the wood panel sizes from the top before I can place my order with Ponoko to have the walnut veneer laser cut. Hopefully I can get my order in this week.

I'm not decided on when to print the main frame, as the walnut veneer isn't exactly 3.2mm, but in the range of 2.9 to 3.5 mm. On one hand, if the veneer ends up on the thin side, I can adjust the model before printing to try to make them flush with the frame. On the other hand, imperfection is probably a good attribute that will help the model look more realistic and less mass produced, so perhaps I go ahead and print it now and just roll with the result.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Thu May 20, 2021 8:06 pm

Designing the Lid
The shape of the compass' main body dictates most of the size of the lid. The horizontal dimensions are the same, and by counting pixels in reference screenshots, I'm able to calculate the height of the box lid and dome. On the top, I again counted pixels to figure out the spacing of the trim rings and the width of the dome.

While the end result looks extremely similar at first glance, comparing my design (on the right) to the Thingiverse model (on the left), you notice some important differences. The trim rings are spaced differently, and the dome width is dramatically wider.

image.png
image.png (1.06 MiB) Viewed 12487 times

These differences extend to the horizontal profile too, where you can see that both the main lid body and dome are slightly taller in my recreation:

image.png
image.png (392.89 KiB) Viewed 12487 times

One element I've been wondering about for a while now is how much clearance the sun dial will have inside the lid's dome. I had a feeling it would be tight. Since both elements rest on the same plane, for a quick check I simply loaded them both into my slicer and peeled away the top layers.

image.png
image.png (1.31 MiB) Viewed 12487 times

The dome wall is 3mm thick in my design, so I would estimate there's 7-8 mm of clearance between the sun dial's tip and the dome's wall. That's more than plenty, I was expecting a much tighter fit.

Another difference in my design is that I cut the bottom edge of the dome vertically, to allow for the laser cut veneer to simply drop in. This is especially important as the dome is so much wider it nearly reaches the inner trim ring, so every millimeter in wood width matters.

image.png
image.png (1.67 MiB) Viewed 12487 times

And here's a gratuitous render revealing the differences between the Thingiverse model which didn't come with the wood panel parts in the design, my and ground up redesign based upon pixel-by-pixel analysis of the compass of destiny in various screen grabs.

image.png
image.png (908.96 KiB) Viewed 12487 times

The biggest design change is the sun dial, which I believe was designed for easier printing in the Thingiverse model, but which didn't try to capture the design details of the original prop. For the average home printer, I think the Thingiverse design was a good compromise, as the level of detail required to print my version isn't accessible at the home level. Thank goodness for affordable commercial services like Shapeways, I'm truly excited to see the final result in a few weeks.

The next biggest design change is simply segmenting the inlaid metal panels to recreate that real-world effect of metal joints. Ironically, even some of the movie props lack this detail - instead they seem to just have flat metal tape on flat wood, though the nicer movie props have this feature.

Another significant change was replacing the printed hinge with a real-world hinge and relocating it to the correct orientation. Ironically, the designer of the Thingiverse model likely could have designed a printable hinge of the correct design more easily than the path I took, and it would have made printing even easier.

The final big change was adjusting all of the dimensions, from box heights to the length of the corners, trim widths and dome shape, to fully capture the movie prop's presence.

Perhaps you'll look at the picture above and think it was all a waste of time, as similar as they are to each other. But for me I know I'll always have a special connection to this prop, knowing it is the most accurate and finely detailed replica in the world.

image.png
image.png (1.4 MiB) Viewed 12487 times

Now that I have the panels for the top lid designed, I can place an order with Ponoko to have them laser cut from walnut veneer.... erhhh, well... almost but not quite yet.

One feature I've been dreading is the inside of the box lid. Check it out:

image.png
image.png (594.75 KiB) Viewed 12487 times

Talk about a multi-material masterpiece. The reflections on the bottom edge confused me at first into thinking the entire bottom edge was metal trim. But it turns out only the outside millimeter or so is metal trim, with the rest being a wooden rim. Above that is the leather detail, for which I've already ordered some material to hand cut. Then there's another circle of wood with a very deep bevel. And finally the internal dome which is painted dark blue with gold dots for stars. If I had proper woodworking tools like a lathe, some of this would be quite easy to mill from a block of wood. But I'm not so fortunate.

Instead my plan is to have two pieces of 3.2mm thick walnut veneer cut, one for the outside rim, and the second for the beveled circle. Of course, Ponoko can only cut straight down with their lasers, so I will have to hand sand that bevel. I will probably order some spares for that part, as I'm sure I will fail the first few times...

I've modelled these features and I think they turned out surprisingly well:

image.png
image.png (1.25 MiB) Viewed 12487 times

They are comprised of the two extra wood pieces to have cut:

image.png
image.png (1.42 MiB) Viewed 12487 times

And the small piece of leather that I will hand cut:

image.png
image.png (774.73 KiB) Viewed 12487 times

Unfortunately, this makes for a very challenging design for a 3D printer. Notice that the bottom edge is only 0.5mm wide, which is the only contact with the print bed, and there are multiple shelfs plus the dome that will need a ton of support (ditto on the wood panel cutouts on the side needing support too).

image.png
image.png (1.2 MiB) Viewed 12487 times

Fortunately, all of these plastic surfaces except for the dome will be covered by other materials (aluminized tape, wood veneer, and leather), so the typical scarring from the extensive supports shown below won't affect the end result.

image.png
image.png (1.21 MiB) Viewed 12487 times

I'm also noticing that the slicer doesn't want to output some of the finer details, like these trim corners:

image.png
image.png (8.15 KiB) Viewed 12487 times

They disappear in the sliced output:

image.png
image.png (199.68 KiB) Viewed 12487 times

The effect is worse on the extremely thin lower sections. I may have to come up with some very creative solutions to get these printed.


Next Steps
The lid is 98% complete, I just need to design in the hinge slot. Then I can safely assemble my Ponoko order for the laser cut walnut veneer.

Then I need to figure out the printing challenges I outlined above. The answer is probably dividing the model into even smaller components (separate trim pieces) that print separately, then get glued together. Otherwise, the model is not printable as-is, as least not with my printer. Dividing the model would also resolve the issue of trying to print a large model standing on a very thin 0.5mm wide outline, which in my experience is not a recipe for success.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Sat May 22, 2021 10:22 am

Solving Printing Problems with a Redesign
In my earlier test print, which I used to validate trim panel gaps and test metailized tapes, I didn't have any vertical walls to print. Because of this, I didn't realize until I had pretty much finished the model that those features don't print well vertically. You can see that in the slicing picture above (previous post) that the sharp points of the vertical trim sections were missing in the slicer's output. No amount of tweaking my settings would make them re-appear. The issue is that these features are just too small for a relatively large 0.4mm nozzle to print vertically.

And there doesn't seem to be a way to print a 6-sided cube without 67% of the faces being oriented vertically.

While brainstorming possible solutions, I realized that I could peel off each face and print them separately, and glue them all together onto a central core.

image.png
image.png (686.67 KiB) Viewed 12474 times

This would give multiple benefits. Not only would I be able to print each face in a horizontal orientation, improving printability of fine details, but I would be able to print them face down.

image.png
image.png (726.17 KiB) Viewed 12474 times

On a smooth print surface, the first layer gets squished extremely flat onto the surface to force it to stick. This has a side benefit of making the first layer extremely smooth, without any sanding, like glass. This becomes a great feature when you consider it is these surfaces I'm covering with the aluminized tape. In my earlier tests, I discovered that surface imperfections show through the tape quite easily. So the best way to prevent them is to print those surfaces against the print bed to make them glass smooth.

I added the Lego like registering pins to ensure alignment during glue up, I don't expect it will snap together like a Lego.

I included the box core with the bottom face, and took this opportunity to split out the inner 608 bearing support structure, which would have required a ton of printing support to print elevated off the print bed. By splitting it out, not only will it print quicker, but with better quality, and is a simple drop in solution during glue-up.

image.png
image.png (678.65 KiB) Viewed 12474 times

The top face is similar to the sides, just without the registering pins. One huge benefit to this redesign is that it allows me to make another critical design change. Notice the open ring below the trim ledge:

image.png
image.png (249.64 KiB) Viewed 12474 times

This is the inner barrel section where the Windrose compass plate resides. I'm fairly certain this area is supposed to be wood, but in my earlier design I had no way to slot wood in that area. Now I can have a circle laser cut of the 3.2mm thick walnut veneer, and it would drop in right before gluing on the top.

The only problem now is that 3.2mm might not be thick enough:

image.png
image.png (764.33 KiB) Viewed 12474 times

You can see in the image above that the compass plate is about 1.5mm to 2mm lower than the shelf where the wood trim ring will go. One way I can solve that is to double up the wood rings to get 6.4mm in height - the model has plenty of vertical space for a second ring. That would allow the wood trim inside to be consistent, furthering the impression that this is a wooden box with metal trim. Ponoko actually offers the walnut veneer in two thicknesses, 3.2mm and 6.4mm, presenting an alternate way to get a thicker ring, though this may be more costly than just doubling the rings - I'll have to investigate my options there.

While redesigning the model with separate parts for each face has been fairly quick (a benefit of Fusion 360, I can simply copy my sketch into a new document and extrude the new part without recreating all the dimensions), I've yet to update the domed lid. I'm printing the main box and sides now so I can do a test fit to ensure this solution works as I expect. It's possible I may have to tweak the solution a bit to make it work in the physical world, and it's probably best to learn these lessons before redesigning the lid.

I've got about 10 hours of printing left on the main box parts. Updates later...
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

Jamie
Posts: 942
Joined: Wed Dec 27, 2017 11:26 pm

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Jamie » Sat May 22, 2021 3:04 pm

Paul,

I am amazed at what you have done!!! You're one talented individual!

Jamie

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Mon May 24, 2021 3:45 pm

:D Thanks Jamie! It's a fun hobby that forces me to improve my skillset at every turn.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

User avatar
Pauven
Posts: 2777
Joined: Tue Dec 26, 2017 10:28 pm
Location: Atlanta, GA, USA
Contact:

Re: [DIY] Jack Sparrow's Compass (Work in Progress)

Post by Pauven » Mon May 24, 2021 4:37 pm

It took much more effort than I ever expected to print the new design, as some of the parts have extremely thin areas that were being held together with the minutest amount of printed plastic. I had to add chamfers and gussets to various parts to make everything thick enough to print, plus carefully orient all of the bridges (long spans of printed plastic that bridge over open areas in the model) to make sure each bridge had enough "land" to hold secure.

Luckily I was able to get the lower box printed without any major design tweaks.

In addition to printing the box components, I also printed the "wood" components, which I will have laser cut out of real walnut veneer. I needed to double-check all the dimensions to ensure that the wood parts would fit as expected. Interestingly enough, I printed these wood parts with actual wood infused PLA, primarily to give some color contrast for my mock-up, but also because I felt the wood component was suiting.

image.png
image.png (1.67 MiB) Viewed 12443 times

All of the sides fit well on the box core, with the alignment pins doing their job, though I needed a rubber band to hold the assembly together. I plan to apply the aluminized tape before glue-up, so this is just a temporary fitment test.

Here's a pic of the bottom. The circle panel spins and is connected through the 608 bearing to the windrose compass plate.

image.png
image.png (1.29 MiB) Viewed 12443 times

I find that I actually like the small print imperfections that lead to inconsistent panel gaps, as it makes this seem like a more realistic object.

For the top, I printed out all of the brass parts in yellow PLA. I also discovered something very helpful by printing two inner wood rings of 3.2mm thickness, instead of a double-thick 6.4mm. The gap between the two rings is where the sundial's support arms need to land, so I used a hobby knife and cut some tiny triangles on the upper ring, and the three arms fit it nicely. This confirms I will be ordering those inner trim rings in the same 3.2mm veneer instead of the thicker 6.4mm option I had been considering.

image.png
image.png (1.3 MiB) Viewed 12443 times

Most of the "wood" panels fit as I hoped, snug or slightly large requiring a little sanding. The 4 tiny triangles on top printed small, and I find the gaps too large for the effect I desire. I will size them up by a half millimeter or so before having them cut. The nice thing about having all these panels cut in real wood is that sanding to size for a nice fit is super easy, so I want to err on slightly oversized.

I'm also really happy with the weight of the box - it feels "chonky", and not like a plastic toy at all. For some of the components, I printed with high infill percentages just to add some extra weight. And it's only going to get heavier, with the brass parts Shapeways is casting, the hinge and hasp, and of course the lid. I'm fairly certain this compass replica will weigh over a pound when done. I personally find that weight is a critical element in a replica's authenticity, though it's hard to say how much a magical compass should weigh, so I'm just going for a heavy as possible without adding any artificial weights.

The only other change I need to make to the lower box half is the catch pin for the hasp. I've yet to figure out exactly what I'm going to do here. In the movie prop, it is only just the pin itself, but the Rockler hasp I received has a large mounting plate with the pin. I don't think I can cut it out, and it really needs to be longer so I can embed it in the model. Had I been smart, I would have modeled a pin as an appendage on one of my other models, and I could have simply cut it off to use for this. Alternatives are to find some brass rod at the hardware store, or simply print it in plastic.

Regardless of what solution I come up with, I need to first complete the box lid and locate the hasp on it, so I can determine the correct spot to locate the catch pin on the lower box. Now that I'm happy with the lower box design, I'll start on the lid rework.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

Post Reply