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] Big Hero 6 Hiro's Fighting Robot - Megabot (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] Big Hero 6 Hiro's Fighting Robot - Megabot (Work In Progress)

Post by Pauven » Wed Apr 14, 2021 5:03 pm

Now that I've been building my movie prop collection for a few years, it's hard to know where to start sharing. Probably the best starting point is with my currently printing project, Hiro's fighting robot from Big Hero 6, which I believe is called the Megabot.

image.png
image.png (1.33 MiB) Viewed 5565 times

I was browsing Thingiverse for ideas a week ago, and when I saw the Megabot I knew this would make an ideal project, so I promptly downloaded it: https://www.thingiverse.com/thing:1358483

image.png
image.png (590.94 KiB) Viewed 5565 times

Magnets
To finish the build, the design required 16 disc magnets of the size 9mm x 5mm, which I didn't have on hand, so I went in a different direction. I've been learning Fusion 360 lately, so I decided this would make an excellent project to model. I took some measurements from the Thingiverse model, but otherwise created everything from scratch. This allowed me to make some very important changes.

First, since I had some 5mm x 1mm neodymium disc magnets handy, I thought I would design for those. I designed a few parts and printed test versions out, and quickly found that these magnets were not strong enough. I even tried doubling them up, using two magnets together to make a 5mm x 2mm equivalent, and while this was marginally strong enough to hold the model together, I still felt it wasn't really strong enough to play with it.

I knew I needed to get some larger magnets, but the 9mm x 5mm were just hard enough to find that I decided to do some research and see if I could locate a better option. I turned to K&J Magnetics online calculator to test how much holding power different magnets would provide. While it's simple to assume that larger magnets would always be more powerful, because these were being fit in domes, the larger the magnet diameter, the further away two magnets would rest, reducing the pull force.

Ultimately, I ended up choosing 3/8" x 1/4" size neodymium magnets from Apex Magnets, a company that I've used before and which has a nice selection. This is equivalent to a 9.5mm x 6.4mm magnet, so even larger than the original Thingiverse design, and noticeably more powerful.

image.png
image.png (72.64 KiB) Viewed 5565 times

Even though these are rated for 9.2 lbs of pull force per magnet, the K&J calculator determined that I would get about 2.1 lbs of pull force in the Megabot, as the distance between magnets will be about 3.3mm. This will be about 9x stronger than the miniscule 0.22 lbs I could get from doubling up the 5mm x 1mm magnets. I also got some 6mm x 3mm neodymium magnets for the tips, more on that later. I just ordered all of these magnets a couple days ago, so I'm waiting on my shipment to arrive, and I'll be updating this post as the build progresses.

Full warning, I ended up buying about $60 worth of magnets for this project - but this isn't so bad because I'm making two Megabots, and I will have extra magnets left over. I'll do a final tally at the end, but for now I'm estimating that cost per bot is about $30 with plastic + magnets, plus more for paint. Cost is not really a factor in my DIY projects, and some of my movie prop projects are rather expensive, but I do find it interesting to track.


Design Changes
Asthetically, I felt that the Thingiverse model was really good - it is a very simple shape after all. But as long as I was designing from scratch, I felt there were a few areas I could greatly improve. First, the Thingiverse model was intended (but not required) to print with a double-sided, two-faced head, so you could spin the head around and change it from happy-mode to destroy-mode, but I decided instead to print two Megabots and give them each a single face only, one happy and the other killer. I also wanted to recreate a couple main poses for each mode - but the Thingiverse design had magnets in static positions, so the model would only ever pose in a single orientation.

image.png
image.png (616.9 KiB) Viewed 5565 times

Above you can see that the middle sphere, which has four magnets, can only ever hold the arms straight out to the sides. The head and bottom sphere does have two magnets closer together for the ears and legs, which is good for the most common poses, but not all.

image.png
image.png (690.63 KiB) Viewed 5565 times

For the happy pose, I wanted the arms to angle slightly upward, like the screenshot above. I also wanted the ability to slightly position the arms angled forward, like this pose:

image.png
image.png (667.83 KiB) Viewed 5565 times

For the destroy pose, I needed to be able to rotate the head upwards, the arms to extreme angles, and the legs further apart:

image.png
image.png (509.86 KiB) Viewed 5565 times

My solution was to create slots for the magnets to slide around inside the model. Here's the middle sphere with slots to allow the arms to position up/down:

image.png
image.png (725.61 KiB) Viewed 5565 times

The bottom and head spheres are even more complex, allowing the ears/legs to position up/down, as well as rotate back/forth on the torso. Remember that this is just half a sphere, so with two halves joined it will rotate nearly 180 degrees.

image.png
image.png (835.22 KiB) Viewed 5565 times

I also did something similar for the ear/arm/leg bases, creating a slot for the magnet to move in one direction only, which should allow for the final adjustability to recreate these poses:

image.png
image.png (534.58 KiB) Viewed 5565 times

There was one more change I wanted to make related to magnets. Megabot can extend like a snake, connecting arm-tip to arm-tip, like this:

image.png
image.png (927.68 KiB) Viewed 5565 times

The Thingiverse model didn't have magnets here at all, so this is where I designed in the 6mm x 3mm magnets. These smaller magnets won't provide as much pull force, only about 1/2 pound, but that should be strong enough to hold the snake pose together.

image.png
image.png (216.75 KiB) Viewed 5565 times

In all of these changes, I've done a lot of trial and error printing to discover how close I can place the magnet to the surface before experiencing printing issues. Too close and it doesn't print well, or is too weak to keep a magnet from breaking the plastic shell.

One last change I made is to how the face is shaped. In this shot, you can see that Megabot has an internal sphere that rotates, and there is a slight gap between the inner face sphere and the outer head sphere. Plus the head sphere's opening is chamfered, not flat cut.

image.png
image.png (1009.46 KiB) Viewed 5565 times

The Thingiverse design didn't incorporate either of these elements:

image.png
image.png (736.24 KiB) Viewed 5565 times

While it would have cool to make the face rotate, this would have required going with much smaller magnets, so I gave up on that idea pretty quickly. Though it may be hard to tell in the screenshot below, I added a small gap between the head and face spheres, plus the chamfer on the head opening:

image.png
image.png (701.72 KiB) Viewed 5565 times

The printed effect is subtle and hard to see in photos, especially in black, but in person it does give a distinct sphere inside a sphere effect, especially when compared against the original Thingiverse design (please ignore the teeth marks, my dog found it really enjoyable to play with, haha):

image.png
image.png (1.74 MiB) Viewed 5565 times


Printing Megabot
For those curious, I have a Prusa i3 Mk2.5S 3D printer. I bought it several years ago as a Mk2, and have applied several upgrades over the years to bring it up to Mk2.5s spec. While personally I'm really happy with it, I won't go so far as to provide any recommendations - this is the only printer I've ever used, so my perspective is greatly limited. I've had just enough frustrations with it at times to seriously consider getting a Prusa i3 Mk3S, but honestly I'm holding out for what comes next, possibly a Mk4, as I'm not sure there's enough improvements over my Mk2.5S to justify the expense to get a Mk3S. Actually, I'm more likely to get a Prusa Mini, much cheaper, and dedicate a detail 0.2mm nozzle on it, and use it only for small, highly detailed parts. While I can swap nozzles on my current printer, most of the time I find that to be too much hassle, and incorrect tightening of nozzles has led to many lost hours spent troubleshooting, so these days I just make do with a 0.4mm nozzle all the time. I also print in PLA which is one of the easier plastic formulations for printing.

For slicing (converting the 3D model into print-head move commands), I use PrusaSlicer, as I've found that I get some of my best print quality on my Prusa printer by using the software they provide. For this project, I ended up learning a lot of advance techniques in the slicer to improve my printing.

First of all, my personal preference is to have the printer print in high-detail mode to minimize my post processing later. I'm not in any rush. For the many spherical shapes in Megabot, this meant I needed 0.1mm layers, or finer. But thinner layers greatly extends printing time, and is wasteful in areas that don't need detail. I discovered that PrusaSlicer has a detail modifier, where you can easily raise or lower detail on a layer by layer basis. So for these prints, I start with a base 0.20mm layer height, and gradually drop to a 0.07mm layer height towards to peak of the spheres:


image.png
image.png (582.68 KiB) Viewed 5565 times


By gradually increasing detail only in areas where it provides a benefit, I have been able to greatly reduce printing time without sacrificing quality - in fact these spheres look better and print quicker than simply printing the full sphere at 0.10mm.

Also, since I needed so many copies of the same parts (6 arms, 3 bottoms, 2 middles, 1 head) and was printing out two complete Megabots, I found it a huge time savings to place multiple part instances on the print bed and print them all together. The problem with this approach is that you get stringing - spiderwebs of melted plastic connecting one part to another, due to the print head jumping between parts on each layer, that increases manual post processing to remove them. The best prints come from printing one part at a time, ensuring that the print head stays within the part's perimeter, eliminating stringing.

While PrusaSlicer has an option to print each part sequentially, it has very restrictive default settings for part spacing and height, and all of Megabot's parts were too big. With very careful measuring and planning, I determined that I could print 6 parts if I spaced them out far enough, with the print order left to right, front to back. The parts in each row also have to be staggered towards the back. Failure to space everything perfectly will cause the extruder assembly or X-axis slide rod to collide with a previously printed part, which would likely ruin the current part and possibly even damage the printer. Once I knew I had safely spaced all parts out far enough, I overrode PrusaSlicer's restrictions to force it to accept my layout as safe.

image.png
image.png (517.06 KiB) Viewed 5565 times

In the layout above, there is only 5-10mm of open space between the extruder and each previously printed part, but that's more than enough to successfully print all 6 arms together, one-by-one. Here's a pic I took of the print in process, and you can see just how close the extruder assembly and chrome x-axis support rod comes to the previously completed arm:

image.png
image.png (1.11 MiB) Viewed 5565 times

By the way, at 0.20mm detail level, these 6 arms take about 10 hours to print. Each entire Megabot takes about 27 hours to print at these high-levels of detail, so it's taking over two full days of round-the-clock printing to get my two Megabots, plus several more days of development and test prints to hone the final design. Fingers crossed the magnets fit okay when they arrive...

image.png
image.png (1.21 MiB) Viewed 5560 times


For those who enjoy knolling shots, here's all 39 parts (minus the 22 magnets):

image.png
image.png (1.27 MiB) Viewed 5565 times

Plus a gratuitous 2nd knolling shot with all the parts flipped so you can see the internal passageways:

image.png
image.png (934.86 KiB) Viewed 5565 times


And even though I can't perform final glue-up until the magnets arrive, here's a test fit of all the parts. Overall size is almost 12 inches tall, about 29 centimeters:

image.png
image.png (770.43 KiB) Viewed 5565 times


Future Design Changes
With the original Thingiverse design, the tightly fit magnets also provided an alignment function. With my design that allows most of the magnets to slide in longer slots, that alignment benefit is lost. With careful gluing, this may not be a problem, but I do have concerns. One easy fix would be to change the center joining pin for the 3 body spheres from cylindrical to cubic or some other poly shape, which would hold the two halves in proper alignment during glue-up. Unfortunately, this design change would mean I would have to reprint the entire body again. I may go ahead and glue one together to see if this is a problem or not before making this change.

If I do make this change, I may also revise the head slightly again, enlarging the gap I placed between the face sphere and the head sphere. While I like the current design, it's still a feature that is hard to notice on black, and might even get clogged in painting. But then again, in the movie you can only see the gap when the face rotates, as otherwise the gap disappears. Perhaps my barely-there gap is the best compromise.

There's also the possibility that the magnets won't fit or slide correctly, something I can't test until they arrive. But if there's any problem there that requires design tweaks, I'll likely make the above tweaks too.



Post Processing
At this point, I'm still a little undecided on how I want to post process the printed parts. Typically, I want to hide that my props are 3D printed, but this time I actually feel that seeing the layer lines is true to spirit of what Megabot would likely be in real-life - a 3D printed robot. That said, in the movie Megabot appears to be weathered gunmetal:

image.png
image.png (1.26 MiB) Viewed 5565 times

I've also found that with the plastic filament I'm currently using, even light sanding leaves white artifacts, so if I do any sort of surface smoothing I'll probably have to paint to make it look okay again. I know some of the close-up photos above make it look like painting is an obvious choice, but in person the black filament isn't bad at all.

Another common tweak I do to make my 3D printed props more realistic is adjust weighting, as often even if printed solid with 100% infill, plastic is just too light to accurately recreate the experience of holding certain objects. I've used plaster of paris, resin, and even lead shot, for various projects to add weight and hardness. In this case, printing with 20% infill the plastic Megabot feels about the right weight, plus the magnets will add a bit more mass. Also, if I add weight, it may make the parts too heavy for the magnets to hold everything together. So for Megabot, I won't be adding any weight. I will say that with 3 perimeters and 20% infill, everything feels really solid, so if I do paint it to look like metal, I think it will be stiff enough to feel believably like metal, though obviously it won't sound like metal. There's only so much you can do.


Cost
A lot of factors go into calculating cost, but one factor I won't include is my labor. I've spent dozens of hours on tweaking the design, but that was enjoyable for me. Most of the time was just waiting for prints to come off the printer, so that's not really a factor either for the home hobbyist, though if you were doing this for a business that would definitely be a consideration.

So I'll just look at plastic, magnets, primer, and paint.

The amount of plastic used depends great on your 3D slicer / print settings. Two perimeter shells or three? 20% infill or 50%? 5 solid bottom layers or 7. There's a ton of settings to tweak that can greatly affect how much plastic is used. For example, in an earlier test, I printed an arm at 5% infill and with only 2 perimeter shells and reduced bottom/top solid layers, and printed it weighs only 10 grams, making it more suitable to the smaller test magnets I had on hand. But this part felt a little flimsy and too light, so in my final design print, I used 3 shells, 20% infill, and default bottom/top layers - this doubled the weight to 20 grams. The part feels much more solid, which I like, but definitely requires the strong magnets. This doubling of weight means a doubling of plastic and a doubling of cost.

I just weighed everything together, and in total it came to 218 grams. I'm using relatively cheap TecBears black PLA filament, which I got for $17 on Amazon. That means my plastic cost is only $3.71, amazing!!! Of course, I'm not factoring in the cost of my 3D printer, but I will say that my printer has more than paid for itself over the years.

Even though I had to buy more magnets than needed for this project, I'll only include the cost of the magnets used. 16 3/8" x 1/4" neodymium disc magnets cost $19.19, and the 6 6mmx3mm neodymium disc magnets cost $1.08, so $20.27 in total magnet costs.

The Megabot model is small enough that two full-size cans of spray paint (one sandible filler/primer, one paint) should be sufficient for both models. Without double-checking prices, let's say that's about $12 in paint. Plus I already have some rub-n-buff paint in various colors from other projects that will come in handy for simulating scratches, but if you need that then tack on a few more bucks. Worst case, let's say $20 for paint ($10 per Megabot), though it will probably be less, and I'll probably have plenty of paint left over for other projects even after painting both models. I also won't bother factoring in a few drops of superglue to assemble the model.

Altogether: $3.71 in plastic, $20.27 in magnets, and $10 in primer/paint, and I can safely estimate this model cost about $34 each.


Conclusion
What, are you crazy?! This isn't concluded! I'll post updates later on assembly, finishing, and glamour shots. Stay tuned!!!
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] Big Hero 6 Hiro's Fighting Robot - Megabot (Work In Progress)

Post by Pauven » Thu Apr 15, 2021 4:27 pm

Super Duper Magnets Have Arrived
The magnets came in the mail today, so I had a bit of fun playing with them.

First of all, those 3/8" x 1/4" neodymium magnets are ridiculously strong!!! Not too strong for use in the Megabot model, they end up spaced far enough apart that I would deem the completed model safe for young child use. But simply trying to get one magnet separated off the stack is incredibly difficult, I've never felt anything like this. And these aren't even the magnets that come with all the major safety warnings, I can only imagine what those are like - must be scary.

With the high power of the magnets, plus the loose fit inside the model (since I gave plenty of extra room to allow to printing tolerances), assembly is pretty much impossible at the moment. I can get two opposite magnets in the body, maybe even three, but with the middle sphere using four magnets, I can never get the 4th magnet in, they all jump together. Combined with the fact that I have to manage the North/South orientation of the magnets, I'm going to need some extra help just to assemble them. I think the best solution is to create an assembly jig. It will be a concave dish that the sphere half would nest inside, and the dish would have four magnets installed on the perimeter glued in the correct orientation. That way, as I insert the magnets into the sphere, the magnets in the jig should hold them in place and in the correct pole orientation.


Magnet Fitment Issues
image.png
image.png (586.5 KiB) Viewed 5559 times

I also discovered that my magnet sliding slots don't have quite enough vertical clearance due to a printing issue. The 3D model has a sliding magnet height of 10mm in the design, which should be plenty for a 9.5mm magnet, but the hot melted plastic droops ever so slightly during this bridging maneuver. It's not a huge sag, only 0.5mm, but because the same sag exists on both sphere halves, the total loss in vertical clearance is a full millimeter. The magnet is just large enough that it is creating a gap between the sphere halves of about 0.5 mm, so if the sag wasn't there then the magnets would actually have about 0.5mm of clearance. I have tweaked this element in the design, adding a full 1.5mm of more clearance, so the two sphere halves will close and the magnet should move freely.


Other Design Tweaks

image.png
image.png (1.18 MiB) Viewed 5559 times

While I was tweaking the design, I went ahead and changed the center pin to a hexagon, which makes it much easier to align the two sphere halves in the correct orientation.

Lastly I added an addition 0.25mm of gap around the face. That's such a small amount, it's almost hard to notice. In the pic below, the left head is the latest with the ever so slightly larger gap. In person, it's just enough of a tweak to change it from making you wonder if you're seeing a gap or a weird shadow, to knowing for sure you're seeing a gap. The one in back in the original Thingiverse design that has no gap at all.

image.png
image.png (1022.61 KiB) Viewed 5559 times

Printing
The new parts are printing now. It's a 9 hour print to create the three Megabot body spheres, and of course I'll have to do this twice. So it will be tomorrow before I can play with it again.

In the mean time, I'll get to work on designing an assembly jig. Should be a fairly simple design, though perhaps I'll get fancy and create a magnet retention solution so I don't have to use glue, as these magnets are expensive...
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] Big Hero 6 Hiro's Fighting Robot - Megabot (Work In Progress)

Post by Pauven » Fri Apr 16, 2021 1:11 pm

While I was able to assemble a 3-magnet sphere (head or bottom sphere) by hand, it was extremely hard, and I realized that it would be impossible to do this manually without supergluing my fingers to the sphere.

For that reason I designed the assembly jig to work in both 4 and 3-magnet arrangements (at least, in theory). For the 3 magnet orientation, the 3rd positioning magnet is not secured, so I can slide it manually - this is important for sliding the internal magnet into the halfway point between the halves during glue up, as that channel must stay aligned to prevent from blocking the magnet. Even with a hex pin to hold the two halves in alignment, there's still some slight wiggle that could allow for the channel to be misaligned during glue-up.

image.png
image.png (378.32 KiB) Viewed 5555 times

The magnets fit well, and are secured easily.

image.png
image.png (876.89 KiB) Viewed 5555 times

And though it still wasn't easy, it was possible to get all four magnets positioned, ready to glue the two halves together. (Note, the middle hemisphere shown in the pic below was one of my rare failed prints from last night, reprinting now...)

image.png
image.png (1.5 MiB) Viewed 5555 times

And the second hemisphere slid on easily, so the alignment jig is working pretty great... almost:

image.png
image.png (1.08 MiB) Viewed 5555 times

It was only at this point that I realized my major flaw in the design. With 8 super powerful neodymium magnets holding this thing together, it's impossible to remove the sphere from the jig! There's no leverage, and there's easily around 10 lbs of force (feels like more) holding the sphere in the jig. I can't get the magnets out of the jig, and I can get the sphere out either. Facepalm.

Glad I haven't glued anything yet. I'll take it all apart, then drill a large hole in the center of the base. That way I can lower the jig over a wood dowel and force the sphere out after glue-up.

Unfortunately, the jig isn't working near as well for the 3-magnet spheres. Positioning the 3rd magnet on the side didn't provide any holding power to keep the other two magnets properly positioned - the sphere would slide down and other two magnets would tilt out, making it impossible to slide the second hemisphere into place.

image.png
image.png (1.67 MiB) Viewed 5555 times

I did find success if I placed the 3rd magnet on the bottom, so that the 3rd magnet reaches the end of the internal slide and holds the sphere in the correct orientation. This was unplanned, and may pose problems once I drill a hole in the base. I probably need to make sure the hole is smaller than the magnet, perhaps a pencil sized hole will work best.

image.png
image.png (945.75 KiB) Viewed 5555 times

I also explored using smaller, less powerful magnets in the jig. When I purchased the magnets for Megabot, I also got a selection of other sizes for future projects, including some 8mm x 4mm magnets. These are about half the strength of the 9.5mm x 6.4mm magnets. This still didn't allow for me to align the 3-magnet sphere using the 3rd magnet on the side, but did make the bottom position even more steady.

I still can't assemble yet. Since my middle sphere's had print flaws from last night, I have to wait on their reprints. I need to assemble the middle first, so that I can ensure all the other magnets are positioned in the correct orientation with respect to the middle, otherwise the model may repel it's appendages instead of attracting them.
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] Big Hero 6 Hiro's Fighting Robot - Megabot (Work In Progress)

Post by Pauven » Sat Apr 17, 2021 4:03 pm

While I'm waiting on some final parts to print out, I decided to work on a painting template. Megabot has two faces, the yellow happy, smiley face, and the red destroy face. Since the printed model's face is only about 1.25 inches wide (33 mm), and I wanted it to look as accurate as possible, I thought that creating a template would be helpful.

Normally I would just print a template on paper and cut it out, but with Megabot's curved face, a flat sheet of paper would not make a great painting template. Instead, I've decided to try 3D printing a template.

To do this, I started with a screengrab of a face from the movie:

Megabot Destroy Face.jpg
Megabot Destroy Face.jpg (59.61 KiB) Viewed 5546 times

Then I brought the screenshot into Fusion 360 as a background canvas, scaling it to the appropriate size, though due to camera perspective I couldn't quite get the all the sizes to match so I simply chose what I thought best. On top of the canvas, I sketched an outline of the black painted areas. Since white would be painted on top of the black, I included the white areas in the black outline.

image.png
image.png (1.97 MiB) Viewed 5546 times


I then created a "contact lens" shape that would rest on top of the face area, and cut out the eyes and mouth.

image.png
image.png (1008.53 KiB) Viewed 5546 times

To print the template, I flipped the model over so that the supports would touch the outside of the dome, so the inside would remain unblemished, ensuring a good fit against the model's face.

image.png
image.png (708.95 KiB) Viewed 5546 times

I was hopeful that my design would print and fit okay. It seems I'm getting pretty good at modelling, as it actually fit perfectly the very first try. While it might look like the nose is also cut out in the image below, that's just an optical illusion, only the eyes and mouth have openings.

image.png
image.png (898.19 KiB) Viewed 5546 times

Finally I repeat the process for the white irises and teeth, and do it all over once again for the happy smiley face.

image.png
image.png (902.37 KiB) Viewed 5546 times

I'm still a while off from painting, so I'll show more of these templates when I get to that step.
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] Big Hero 6 Hiro's Fighting Robot - Megabot (Work In Progress)

Post by Pauven » Mon Apr 19, 2021 3:07 pm

Finally, assembly time!

PLA is one of the more difficult plastics to glue together, and many superglues don't work at all. I've read many studies comparing different products, and Loctite's Ultra Gel Control Super Glue seems to be one of the best products. I've been using it for years now with great results.

image.png
image.png (1.65 MiB) Viewed 5540 times

Additionally, surface preparation is very important. I sand all mating surfaces (that I can reach) to rough them up and give the glue something to bite into. Additionally, I use a product from Aron Alpha called PP Primer, which wicks into the plastic and further improves its glue-ability - though the product labeling on this is very scary so choose for yourself if this nasty chemical is something you want to play with (death from breathing vapors, anyone?).

image.png
image.png (1.85 MiB) Viewed 5540 times

Instead of using the larger 3/8" x 1/4" neodymium magnets in my assembly jig, I decided to use the less powerful 8mm x 4mm magnets. The accomplished my goal of holding the model's magnets in place during assembly, and put up less fight when I had to remove the model from the jig. To ensure that the magnets were placed in the correct orientation, I used a test arm I had assembled with a magnet to ensure each magnet in the jig was positioned correctly.

image.png
image.png (970.89 KiB) Viewed 5540 times

I was also tired of hurting myself trying to separate magnets, even these smallish magnets are too strong to easily separate by hand. So I printed out a handy dandy magnet guillotine that I found on Thingiverse:

image.png
image.png (307.24 KiB) Viewed 5540 times

When assembling the 4-magnet middle spheres, I would superglue in the two stationary magnets first, just to help prevent any assembly issues, though this really seemed unnecessary. Once that set, I would place the other two magnets, add superglue, and place on the other half. In the picture below, you can see the surface scratches from sanding the mating surfaces for glue-up.

image.png
image.png (1.86 MiB) Viewed 5540 times

Immediately after joining the hemispheres I would push out the sphere from the jig using a pencil (note the all-important hole I drilled in the base of the jig), and firmly clamp the two halves together in my hands for 30 seconds. Then using the test arm I would wiggle the sliding magnets around to make sure they didn't get locked in place from any glue squeeze-out inside.

I has also briefly considered lubing up the sliding magnets with some lithium grease before glue-up, and in hindsight I'm glad I didn't. The sliding magnets slide easily enough, but also have just enough "grip" on the plastic to lock into place, so they only slide when force is applied. This seems pretty perfect.

image.png
image.png (763.84 KiB) Viewed 5540 times

After doing the middle sphere, I could then re-configure my assembly jig for the 3-magnet head/bottom spheres - this simply required taking out two magnets, and then using the freshly assembled middle sphere to position the bottom magnet with the opposite polarity.

With that complete, I then began assembling the arms. One interesting observation is just how closely I designed the magnets towards the very edges of the model. In the picture below you can see one of the printed arm tips (finger-tips?), which looks completely solid:

image.png
image.png (208.28 KiB) Viewed 5540 times

But when backlit, you can see right through the very thin plastic in some areas, probably about 0.2mm thick - if the magnet was any closer to the surface there wouldn't be plastic there at all. This translucency was hard to capture on camera, in person it looks much thinner.

image.png
image.png (327.33 KiB) Viewed 5540 times

To ensure that the Megabot model could successfully link in the snake shape, I needed half of the finger-tips to be in the opposite polarity. To do this, I first glued-up half the finger-tips (6, since I'm doing two complete Megabots), then used those to control the magnet orientation for the other 6 before gluing them in place.

image.png
image.png (802.54 KiB) Viewed 5540 times

Then I glued the finger-tips onto the arms:

image.png
image.png (946.29 KiB) Viewed 5540 times

Before gluing the arm bases onto each arm, I would set the magnet orientation using the 4-magnet middle sphere:

image.png
image.png (1.5 MiB) Viewed 5540 times

With all the gluing complete, I was finally able to check out my creations - Megabot lives! Below I put the face painting templates on to give you a hint of how they will look once painted:

image.png
image.png (1.54 MiB) Viewed 5540 times

I was also able to arrange Megabot in the snake form as I had planned. Unfortunately, the magnets aren't quite strong enough to hold it dangling from the top - it can only handle two sets of arms, as soon as I get to 3 it falls apart.

image.png
image.png (1.15 MiB) Viewed 5540 times


Failure(s) Analysis:
As much as I would love to exclaim this build completed perfectly, it does have some flaws. Yes, a lot when right: the magnets slide as I wanted and allow for the offset positions I had hoped, and the model looks amazingly realistic to the movie design. But it could be better.

One minor assembly flaw is that two of the sliding magnets got superglued into place. I was able to free one of them by running another magnet back-n-forth over it for a while, but the other one is impressively steadfast. This happened in spite of me working the sliding magnets around during glue set-up - I guess I should have done this for longer. But compared to the bigger issue, that's pretty minor.

Something I hadn't put enough thought into is how two sphere's held together by magnetic forces would spin freely on each other, like a spinning top on a table. I originally noticed this while spinning some arms, finger-tip to fingertip, and thought this would make a nice fidget spinner toy. But what seemed like a cool feature to play with makes it impossible to pose the models. Because there is zero friction to stop rotation, the three body spheres spin on each other and fall limp like a rag doll. The same goes for the arms, while I can angle them offset like I planned, they then flop around like they're broken.

So while the model can achieve the poses I had laid out in my original plans, it can't hold them on it's own. All I can do is lay it flat on the table. This does explain why there is a 3D printable stand for this model on Thingiverse, as external supports are the only way to make it stand up as designed.

While I have a potential solution, it would require a full redesign and reprint of the entire model. The solution would be to replace each individual magnet with two magnets side-by-side in opposite orientations (one pointing north, the other south).

With only a single magnet on each side of a joint, I can control the joint's position, but not the joint's orientation. But two magnets on each side would allow for both position and orientation. I could still keep the sliding feature, so you wouldn't be locked into just a single pose. If this works the way I think it would, then you could actually pose a rigid model that could stand or squat on it's own, holding whatever pose you select.

The harder part of this design change would be sizing all the channels for two magnets, plus downsizing the magnets since there's not enough room for two 3/8" x 1/4" disc magnets, those were already about as large as you could go in a model this size. I think this would require 1/4" x 1/4" magnets (6.35mm x 6.35mm) which when paired up would produce about the same pull force as the current 3/8" x 1/4" magnets. I'm not sure if there's enough room, though, and I'm also not sure if going any smaller would work as power drops off quickly with size. I feel the model needs about the same power as what my current design has, less is probably not enough.

The last failure is that I've also come to realize that this model doesn't make much of a fun toy to play with, mainly due to how floppy it is. There are some cool strong-magnet effects that are fun to play with for a few minutes, but since the model won't hold any poses, it's more frustrating than fun to play with it.


What's Next
I'll probably go ahead and finish the model with painting - it will be good painting practice if nothing else. It will also give me a chance to see if the face gap is large enough to survive priming without getting clogged, and if my face painting templates even work.

And I'll probably print out some stands so I can display them. At that point, I'll see if I'm happy enough with the project to call it done, or if I want to redesign for dual-magnet joints and try again.
President, Chameleon Consulting LLC
Author, Chameleon MediaCenter

Post Reply