Mecha design; Hard corners and round cylinders

The reason why soda pop cans are cylinders has three main reasons; it lacks the weak corners of a cube, has the same silhouette as a cube and a sphere depending whether you’re looking from the side or above, and its manufacturing hits a sweet spot between the two aforementioned shapes. For a mecha design, while a recognisable silhouette is important, and the thematic motif whether or not you want your look robot to look industrial war machine or hand crafted master piece, you should keep your setting in mind whether or not to use cubic or cylindrical shapes.

It is much cheaper to build an object that is shaped like a cube. It is relatively fast, cheap and easy, and by standard your cube’s can deviate few degree. However, no matter what material a cube is made of, its weakest spot will always be the corners. You can have it on its flat side, but if you were to hit its corner, all the force from the impact would be on that tip. The corners, while not as weak, too are weak spots, as they are thinner in material than the rest of the cube. When metal is bent, it stretches. The stretching is dependent on the material and the angle. A very rough rule of thumb is that steel 1mm thick has to be 1mm shorther than its intended measures, as the material will stretch 1mm during bending. This stretching naturally wears and stresses the metal down, and it now being thinner, is weaker. This can be countered by having additional material welded into the corners inside the cube, but this extends production time and adds mass.

Of course you could weld the material together by its corners, but that would introduce heat and additional material. As the rule of thumb goes, when you heat metal up, it hardens, tempers. This might introduce a weak point. This is why you often see the weld seam to be in the middle of a side, because then the corners won’t become tempered by the heat. Then again, welding in itself is kind of intentional mistake whenever it is done. It would always be best to have any and all pieces as one whole, but manufacturing such objects is often extremely expensive and nigh impossible. Mechanical and chemical bonding are also good options to consider, e.g. airplanes’ surfaces are nowadays glued together with epoxy adhesive rather than riveted.

Cylindrical shapes on the other hand don’t have similar weak spots as such, as the whole thing can rolled from one sheet. A robot can weld a large cylinder together rather quickly from its ends. Nevertheless, by standard creating a cylinder shape is somewhat more expensive to manufacture than a cube due to the sheer amount of control the rollers must be under at all times. Unless the process has been automated to a high degree, and the general shape is simple to manufacture in proper fast steps, producing a cylinder can become stupidly expensive. The history of soda can design has allowed the shape and the manufacturing to be extremely efficient and automated, something that is not probably possible in a mecha setting for giant robot limbs and body shapes.

While G1 animation largely ignored the whole issue of corners, the fact is that most mecha designs realise that having hard corners is not a good idea and camfer them down. It might add complexity to the work scheduling and design, but its a small price compared to the benefits

No fiction really concentrates on these things really. Its all about the visual flavour of the design. Mazinger Z has a body based on round torso and cylinder limbs, because that’s what Go Nagai went for in a comic that was quick to work on. Gundam in general uses flat, cubic surfaces because that is industrially more sensible to work with. If you look at the difference between The Federation and Zeon Mobile Suits during One Year War or, you should see that, in rough terms, Zeon has a high number of machines with round and bulbous shapes in their MS, which all are rather expensive to manufacture, while Fed’s mass produced units like the GM are very much made cheaply with cubic shapes governing the limbs and body. Knowing OYW’s setting and its technological level, no MS production is fully automated. Building just one MS takes hundreds of people, of which part simply work to shape the sheets into their proper shapes before a welder prepares the parts for a welding robot to make the final seams, before another person comes in to check how the welding robot has fucked up this time and how he has to fix. We can safely assume that one of the many reasons why GM’s production cycle is much shorter than Zaku II’s due to the geometry of its armouring.

Another benefit in GM’s overall shape and silhouette is that it has a smaller profile than the Zaku II. Much like with real world tanks, the GM seems to have balance between mobility and armouring. It has less mass to move around compared to the Zaku II, though the Zeon MS most likely has better armouring overall. Nevertheless, its silhouette has more room of error what to hit. However, in a world where you have particle weapons able to pierce armouring like its was butter, GM’s silhouette offers less a target to hit. While MS designs would grow in size with time, and get all kinds of shapes to them, these considerations really aren’t an integral part of Gundam‘s base design idea. After all, in real world all Gundam designs have to adhere to the fact that the franchise is model kit driven first and foremost. This dictates everything from colours base faction design differences.

Tanks are the best real world example of balance between silhouette and best of both worlds with shapes.

Thanks Wikipedia

The above PLA’s Type 99a has a showcase of three important bits; round shapes, flat shapes and angles. The design of the turret tapers at the front in order to direct bullets and other hits away from the front section. The same applies to the  front of the tank, where a hit in an angle has higher chances to bounce off. If you look at the GM above, its flat surface invite bullets to penetrate it. The humanoid form is not exactly the most bullet resisting shape, hence why some mecha design elongate the chest further forwards to create a sort of cone effect. The silhouette of the hull increases in size towards back for the same reason as the turret. While its side view might be rather flat, there’s nothing much that can be done there outside additional armouring, be reactive or just additional plating. You want to face your opponent face to face to offer them the smallest possible target with the best protection you can offer. Certain tanks can also lower their aim below the horizon, meaning they can stay behind a hill and offer even smaller target.

While tank battles and their function is not directly relatable to giant robots, a Realistic setting would take into notion the cost of designing and producing shapes that make a mecha. It’s alright to want to design a super sports car for more unrealistic setting, where cost of production is no issue, but a serious take on mecha warfare is required to recognise that shape must not only be dictated by use and utility, but also by material and cost. Tanks give some idea for good utility of industrially probable shapes, while fighter jets can be used as reference point for aerodynamics.

Makes you wonder how the hell did Kouji’s grandfather manage to build Mazinger Z in his basement.

How does all this affect the design you might have in your mind? Depending on your angle you take within the fiction, the overall governing surfaces are dictated by the intention. If they’re intended to be slow walking tanks, consider what direction is intended to be against the enemy fire and add proper slants or curved surfaces. Even if the angles around GM’s cockpit are laughably small, the fact that they are there gives some idea where the design was going. Sure, in reality it probably was just to conform to the humanoid shape, but we can assume the angles were calculated somewhat to encourage ricocheting. Zaku II armour seems to have designed under notion of heavy armour to withstand firepower overall, though as mentioned, particle beam weapons don’t really care unless you have a repelling surface. Consider the production angle as well; a highly mass-produced mecha wouldn’t necessarily have many separate armour segments and these segments would be whole pieces, like a car’s hood. Cars’ hoods and frames are designed to deform on impact, and rounder shapes are safer for this rather than straight surfaces with angles. There’s something on certain lines and shapes that are naturally “safer” to our sensibilities, be it from experience or otherwise, and often using decent logic and “what looks good” attitude with an intended function can yield a design that works well.

While you can completely ignore the realities of manufacturing and process of it, it would add a touch of realism if there is some consideration whether or not certain shapes and angles are probable within the fiction.

Three approaches in designing a mecha

The three approaches to mecha design this blog uses is based on their role and function within fiction rather than in-fiction. The first archetype is the Protagonist, a mecha that functions or acts like any human character and is treated as such within the narrative.

The Protagonist mecha as a character serves an integral role within the narrative. Initially they may seem like simple machines, like the eponymous Mazinger Z, yet they exhibit clear-cut human characteristics in actions and behaviour. Mazinger Z sunbathing in the original series Mazinger Z-series is this exact human-like behaviour the mechas are written with.

Here, the symbolic action of shaking hands is not represent the pilots themselves per se, but the relationship and role of the mechas

These type of mecha can also be explicit characters unto themselves, as it is with the The Transformers and Brave-series. These mecha are only separated from their human co-characters is their nature as giant mechanical beings. In cases like Beast Wars, there is no distinction between characters as such, all of them simply are the characters, but share the main characteristics of being human equivalent in different form.

The Protagonist has a unique role within the story. Not necessarily the main protagonist in itself, often sharing that role with another human character or another mecha. The same categories of heroes and villains apply to these as much as they apply to human characters.

In visual design, Protagonists more often than not share a humanoid body with strikingly human face. Heroman, by all intentions, shared all the previously mentioned points; a human-shaped mecha with human face and sits in a prominent role within the fiction as one of the main characters next to the main human protagonist.

American made in Japan

However, there is extremely wide variety of Protagonist mechas which toy with the concepts and ways to realise the main role. GaoGaigar, for example, in itself has no characters outside as it is an extension of Guy Shishioh; it less piloted as it is a giant piece of armour for Guy.

It must be mentioned that most Protagonist mechas are found in media aimed at younger audiences with healthy amounts of toys, and tend to have connections to the Super Robot side of mecha. This is not to degrade from the fiction itself, only an observation.

Naturally, the opposite of human-like characters would be the lack of humanity, as it tends to be the with the second archetype, the Machines.

The utilitarian approach to mecha design has always been there, though it gained most of its popularity in the 1980’s. While Mobile Suit Gundam certainly paved the way for Real Robot as a sub-genre, shows like Armored Trooper Votoms and FLAG have taken the concept to its more natural direction due to lack of needing to sell toys as much.

FLAG‘s HAVWC, High Agility Versatile Weapon Carrier, is equipment.

Unlike with the Protagonists, a Machine has no nature to speak of. To make a blunt comparison, they are toasters. Their use is largely utilitarian. The form is made and designed for a purpose first and foremost, following the necessities over flavour.

The mechanical design is far more industrial as opposed to organic contours, than anything else among the Machines. Take Heroman above for an example. Most of its shapes are round to further accommodate its humanoid visual. While at a first glace HAVWC would fit this as well, its shapes are equivalent that of a car, lines made to increase aerodynamics. Heroman is not exactly an aerodynamic character, and its not supposed to. That is a tertiary concern at best. In order for it to be more aerodynamic in its forward position, it would require some sort of wind-breaking apparatus around its chest to lessen drag.

However, FLAG is an example of the more more adhered end, similar to Metal Skin Panic MADOX-01, which has been described as equivalent of mechanical pornography due to its attention detailed opening.

The Machine comes in many varieties, all of which share multiple characteristics. Mass production is one, where the mecha can be or is mass produced. Scopedogs are a dime in a dozen in Votoms and are easily replaceable. Round Vernian Vifam is another example of a show, where mechas are tools, and the cast goes through numerous units during the course of the show.

Valkyries from Macross, despite often gaining a prominent role as a single unit or a customised main character vehicle, are all from a production line of similar units. While later entries in the series have made an effort to give most characters their own unique snowflake Valkyrie, in the end all of them are more or less faceless machines that showcase no human characteristics, outside the genre-defining four limbed humanoid shape.

Specialist roles are not exactly uncommon among Machines. Full Metal Panic!’s Arm Slaves, while mostly consisting of non-unique units, the units used by the protagonist Sousuke Sagara deviate from this mould in form of Lambda Driver, which allows the pilot to turn their willpower into physical force. This specialist position, be it due to extra equipment, prototype role or simply because the mecha is a protagonist’s unit, is a common trope. This position does not change them into Protagonists per se, unless human characteristics are applied. It is not uncommon for people, fictional or not, humanise their devices to a large degree and treat them accordingly.

Vehicles technically fulfill this spot,

However, it’s not uncommon to see the the aforementioned archetypes mixed either.

The Hybrid approach takes characteristics from both sides of the fence in a happy mid-ground. Perhaps the most well-known examples of this would be the Evangelion units of Neon Genesis Evangelion. While treated as equipment and something that can be mass-produced, each EVA-unit exhibits overt human-like characteristics from in-universe and in their role. EVA-01 is effectively one of the main characters while still serving the role of a toaster. Its design goes for utilitarian, but only in terms how the EVA-unit itself allows this in-fiction. The base design idea was, after all, a monster barely controlled by humanity.

A some sort of purple mom bot

Another method to give mecha character is by keeping the core mechanics itself intact in terms of its role though the use of Artificial Intelligence. Jehuty from Konami’s Zone of the Enders series of games is exactly this.

Jehuty in itself has no conscience or awareness within fiction, no character to speak of. Its actions and behaviour are determined by its pilot and support AI, A.D.A. In principle, A.D.A. could be embed into whatever Orbital Frame would support the addition.

These three approaches are more or less starting points, more or less. While at first it may seem arbitrary to make a category of three, one of which is effectively just combining the first two, they serve their role in setting the proper mindset for design work. That is, the nature of the mecha rather than the end-visual the designer ends up making. That is up to the designer’s own style and research into the subject materials.

For further reading on expanded subjects, such as combiners, basic design tips, controls and similar, please visit the Robot Related Materials section.

Mecha design: Animal transformation

While majority of the transforming  mecha are of vehicles, Beast Wars engineered biological transformations as we understand it nowadays. We did have transforming robots that had animal like modes, but these were always more or less mechanical in nature. Be it Transformers‘ Dinobots or Tobikage‘s Ninja Robots, organic transforming mecha were a rarity. Beast Wars didn’t just push toy engineering, design and manufacturing further, but also had a television show that is of high quality. It’s 3D CG looks outdated by modern standards, but we can’t blame the staff for not using technology from the future.

Organic transformations use the same idea as before; break down the individual components and reform those into new form. However, we can have two approaches to this, depending on what is our alternative, or Beast Mode. Well, Beast Form is more likely what Hasbro could use nowadays due to Marshawn Lynch trademarked Beast Mode for his clothing line.

The first approach is to have no excessive parts, no unnecessary shifting. To stay true to the idea that we are talking about a mecha that is so organically intertwined with its beast form that the transformation scheme flows from one to another. This approach can simplify the transformation quite a lot. To the point in which an animal pretty much stands up. Let’s use Boxtron’s brother, Anitron, as an example.

New canvas3
Animatron? You couldn’t come up with a better name?

I’ve broken down a generic four-legged animal into more geometrical elements. Breaking organics into chunks of irregular geometric shapes works better than oversimplifying them. This is assuming we want to keep some of natural shapes around, rather than mechanise the beast form completely.  To keep with the first approach, the easiest way we can have the first step in a humanoid mode is to make Animatron stand.

New canva1s1.png
Are you serious?

Four-legged animals are a very easy example due to them having somewhat remotely similar body structure to humanoids. It just needs some tweaking. Let’s throw the animal head back, push the front leg’s paw’s back while flipping hands out from the forearms and turn that tail into a gun.

New canvas2.png

There are multiple ways to handle the animal head. Here we see an example where it’s essentially a hanging kibble on the back. While it’s dirt cheap method, it has its use when handled properly. More often than not, the head of the animal is pushed to the front so that the face is on the chest. This gives the mecha a crest and a thematic approach. Take a look at Cheetor’s toy control art to see how they’ve flipped the head forwards.


Cheetor’s transformation is not the same as Animatron’s (there is more twisting of limbs and whatnot), but the point still stands. With a shorter neck, the head dangles on the chest. There is also an option to leave the animal head as-is and use that as the mecha’s humanoid mode head as well. You can also do what Voltron did and have the face be hidden inside the animal’s mouth, leaving the top of the animal head a helmet.

This approach is very straightforward and can be adapted to pretty much any animal shape, except to those that do not necessarily lend themselves for the humanoid two-arms two-legs form. In cases like this we need to start looking how to either break down elements in order to force them a new shape, or how we could create the animal form to function as a shell. Shellformers are a thing, and here’s the worst (best?) example; Break.

Cliffbee‘s review says it all, really.

You could really just pull all the penguin parts and have a robot with no forearms. The design’s not something that would win prizes, but not all designs are required to be of top percentage.

But let’s combine the shellforming and “standing-up” transformation for the second approach, where we have to tweak some things in order to work. A T-Rex is always a popular theme for robots, so let’s use that. An intentionally bad example, so I can touch upon this sort of volume-centric scheme later on.

New canvas111.png
Let’s try to have a smoother T-Rex than breaking it down to geometrical elements

A dinosaur like this poses two problems compared to the four-legged animal above; the forearms are short and can’t make humanoid arms, and tail is far too long to be integrated easily without adding too much bulk. In order to achieve a similar transformation, we need to add few more steps into the mix.

New canvas3131.png
Animatron is pleased, yes

In order to house human proportion arms, we can stash them inside the torso underneath the robot mode head. We turn the torso’s sides into shoulder armours and have the arms lower from there. Then we can have the neck take over now empty section of the torso all the while pushing the head down to become the chest. Straighten the legs and turn the tail into a turret attached to the pelvis, and shorten it by splitting it open and covering top and bottom.

There would have been multiple ways of achieving a transformation. We could’ve have gone the opposite direction and split the head and neck open to form legs, legs becoming hulking arms and tail becoming the weapon again. This approach, where elements are derived from the volume.

Volume-centric transformation schemes work best when you have enough volume to work with, where you can turn things inside out to reveal something that the mecha was hiding underneath. This isn’t transforming one shape into another anymore, like we did with Boxtron’s example. Here we are revealing completely new elements from within the mecha itself that have no relation to the beast form with the arms. I could’ve done this a lot more, like have the legs transform into more mechanised ones instead of keeping the T-Rex legs as-is. The tail-gun, which looks absolutely retarded above, is another far simpler example.

The two approaches are almost inseparable to some extent. More often than not, you need to use both volume centric and shape transforming approaches when it comes to organic objects due to the nature of the beast. In a way, an animal may force a smoother and flowing idea of a transformation compared to a machine, where industrial efficiency in shape and function reign.

Mecha design; From cube to humanoid

The previous post about mecha design was all about the basic ideas that yours truly tends to use when it comes to transforming or shape changing robots. As mentioned, they are not definitive and many would probably contest them, but they work just as well. However, all transforming mecha follow one essential thought pattern most of the time; from inhuman shape to humanoid shape. This shape can be whatever. Cars, planes, guns, dinosaurs, trains… pretty much everything has been turned into a robot. Hell, there used to be a saying on imageboards that the Japanese can transform anything into a mecha if they just want to. Of course, there are those that simply change utility shape between modes and never become humanoid. These are relatively rarer in scale of things, but the overall discussion follows the same pattern overall. You have a shape that you want to force into another.

The title of this post is misleading. The term that I should be using is cuboid. However, I am going to break any and all good language practices and keep mixing cube and cuboid to label any cuboid shapes. This would an example of marketing of sorts when you get down to it, as many companies want to use cube in a similar sense. Nintendo’s Game Cube being one, with it being a cuboid even when the Game Boy player is attached.

As with any matter like this, there is no one correct way to do anything. The examples here are simply just for the sake of examples and being as simple as possible. Expanding on basics and building on them is really the only way to get around.

The core idea is to take a cube and “spread” it to the similar breakdown as human would be, if we’d draw human with simple geometric shapes.

Continue reading “Mecha design; From cube to humanoid”