[DIY] Big Hero 6 Hiro's Fighting Robot - Megabot (Work In Progress)
Posted: 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.
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
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.
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.
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.
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:
For the destroy pose, I needed to be able to rotate the head upwards, the arms to extreme angles, and the legs further apart:
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:
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.
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:
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:
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.
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.
The Thingiverse design didn't incorporate either of these elements:
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:
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):
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:
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.
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:
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...
For those who enjoy knolling shots, here's all 39 parts (minus the 22 magnets):
Plus a gratuitous 2nd knolling shot with all the parts flipped so you can see the internal passageways:
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:
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:
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!!!
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
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.
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.
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.
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:
For the destroy pose, I needed to be able to rotate the head upwards, the arms to extreme angles, and the legs further apart:
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:
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.
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:
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:
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.
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.
The Thingiverse design didn't incorporate either of these elements:
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:
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):
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:
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.
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:
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...
For those who enjoy knolling shots, here's all 39 parts (minus the 22 magnets):
Plus a gratuitous 2nd knolling shot with all the parts flipped so you can see the internal passageways:
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:
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:
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!!!