The Insane Engineering of the Gameboy

The original Gameboy was launched 

in 1989 and was received with mixed reviews. While its success is 

ingrained in our cultural memory now, when it was launched it was a 

technologically inferior product. The Gameboy was designed 

to be a cheap, low-powered, portable gaming system. It was limited 

in many ways. No backlight for the screen and incredibly low installed 

memory available for coding games. Review magazines of the time viewed these 

features as a negative, but these compromises in design were exactly why the Gameboy 

succeeded. This was a console for the masses. Even with these limitations, engineers 

and programmers came up with ingenious methods to create games that have 

not only stood the test of time but launched some of the most valuable franchises in the history of the entertainment industry. 

TV shows, movies, toys, and even theme parks. This is the insane engineering 

of the Nintendo Gameboy. The Game Boy's simple design borrows 

much of its success from its older brother the NES. A straightforward 

and familiar controller setup. Nintendo knew that size and weight were the most 

important factors for a system to be portable. The Gameboy was almost half the 

size and half the weight of its competitors. Just under 15 cm in height and 3 

centimeters thick, it weighed only 220 grams. This 35-year-old console doesn’t feel oversized 

like the mobile phones of this era. Gameboy focused on user experience from the get-go, an 

ethos that has defined Nintendo to this very day. But how did Nintendo manage to make the Gameboy 

so much smaller and lighter? To begin, one of the primary technological limitations 

of the early 90s were these things. Alkaline batteries. While our Gen Z 

audience may recognize these as the batteries they have to replace in their TV 

remote once in a blue moon. These things were everywhere in the 90s. Costing about 50 

cents each, or about 1.16 in today’s money. I spent every penny of my pocket money getting 

these batteries to power my Gameboy in the 90s. Large, bulky, non-rechargeable, 

and expensive. Minimising their use as much as possible was going to give 

Nintendo an edge over their competitors. The Game Boy's main competitor, the Sega Game Gear, used 6 AA batteries. 

While the Gameboy used just 4. This of course saved space, made the Gameboy more 

compact, and saved money for the consumer. Especially as the Gameboy batteries lasted vastly 

longer despite having less energy available. The Game Gear’s 6 AA batteries supplied 4.5 

watts to power its electronics. Draining the 6 batteries in just 3 hours. Costing about 

2 dollars and 30 cents per hour of gameplay. The GameBoy, with its 4 batteries allowed up to 30 hours of gameplay. It cost just 

16 cents per hour of gameplay. Imagine being me in the 90s. Trying to explain 

to my father, who remembers when someone got a car for the first time in his village, that 

I needed money for a new set of batteries every two weeks. Well, for the Sega Game 

Gear that was likely closer to every day. One of the keys to Nintendo's success was 

recognizing this limitation and working around it. While the Game Gear featured a fully lit 

coloured LCD screen. The Gameboy featured a monochrome screen that was capable of 

displaying just 4 shades of green that were impossible to see in darkness 

because it didn’t have a backlight. While the Game Gear may have gotten better 

reviews with its power-hungry electronics, the Gameboy got the customers with 

a system that drew just 0.7 watts. The Game Boy's engineers were determined to use 

low-powered screens, and despite this screen being a huge part of our nostalgia today, it almost 

led to the cancellation of the entire project. The best available low-powered LCD 

screens in the 80s worked by having a passive matrix of electrodes 

that controlled a grid of pixels. A pixel consisted of some liquid 

crystals sandwiched between two perpendicular polarising filters. At rest, 

these liquid crystals twist the light that bounces off the backplate, which allows the 

light to pass through the set of filters. These crystals respond to voltage changes, 

untwisting as voltage is applied, when this happens less light can pass through. Early 

prototypes of the original Gameboy used liquid crystals that naturally twisted only 90 degrees 

at rest. These 90-degree structures slowly untwist with voltage with the amount of light transmitted 

being proportional to the voltage applied. However, there was a problem. This slope is 

not steep enough. This was a problem for the low-powered passive grid matrix displays 

used in the early versions of the GameBoy. The low-power screen used tiny changes in 

voltage to differentiate between on and off, and the difference in voltage needed to 

turn the pixels on and off was too large. A slight difference in voltage resulted in a 

very subtle difference in the amount of light emitted by individual "on" and "off" pixels. 

In other words, the contrast was very low. This got worse as the passive matrix 

created an interconnected set of pixels where voltage could leak into neighbouring 

pixels. So neighbouring pixels would also be slightly activated resulting in a blurry 

image that looked even worse from the sides. When Nintendo's President Hiroshi 

Yamauchi tested a version of the Gameboy with these 90-degree twist screens 

he actually cancelled the entire project. However, a breakthrough occurred in 

the late 1980s. SHARP perfected a new type of LCD screen known as Supertwisted Nematics. These screens used crystals with twists 

between 180 and 270 degrees. These extra twists made a sharper transition 

between on and off possible. This is what a super twisted crystal 

transition curve looks like. The transmitted light drops off rapidly 

with a much smaller voltage change. This technology resulted in sharper black and 

white pixels, with the green colour of the gameboy being a byproduct of the polarising filters tint, 

but how did the gameboy create 4 shades of green. It was not possible to create these shades 

with 4 different voltages settings. Instead the gameboy created different shades 

by quickly pulsing the pixels on and off. Faster pulses result in darker shades, 

while slower pulses result in lighter shades. This is the same technique that 

LEDs use to brighten and dim. We can’t perceive the pulsing with our 

eyes, but cameras can pick it up. The quest to make the system as cheap as 

possible of course created limitations elsewhere. The 8-bit CPU could only handle 64 kilobytes of 

memory, less than a single frame in this video. Programming a game like Super Mario Land with so little memory available required 

some creative problem-solving. All of the Gameboy functions, maths, and logic happened by simply reading or 

modifying those 64 kilobtyes. Some are read from the Gameboy itself while others 

are read from the inserted game cartridge. These 48 numbers, for example, are read from 

the cartridge every time the Gameboy is turned on and every licensed game cartridge has to have 

the exact same hard-coded data at this location. This is the data it reads, just numbers. But, by rearranging them and converting them 

to binary we can start to see a familiar pattern. Turning off the pixels with ones we 

can make out that nostalgic logo that dropped into the screen before any game. Inside the 

Gameboy, there is a copy of these same numbers. During the boot-up process, the Game 

Boy displayed the logo stored in the cartridge while comparing it to the 

one in the system, byte by byte. If a faulty connection caused a byte to be read 

incorrectly, the Game Boy would not start up. Unintentionally, this sparked a 

magical tradition among kids worldwide. A technique that transferred across 

cultures and continents before the internet existed to share that 

knowledge. Take the cartridge out and blow on it to remove any dust 

that may be causing faulty connections. For this byte-by-byte comparison, they could 

have used any numbers or any image. But they intentionally used the trademarked 

Nintendo logo to curb bootlegged games. If you were an unlicensed game 

developer, this forced you to display Nintendo’s trademarked logo, and 

if Nintendo did not permit you to use it, you would be breaking trademark laws 

even if the games themselves were not. However, using individual bytes to create the 

image, the way the Nintendo logo was displayed, is not a very efficient way of 

populating the full screen for games. If the 160 by 144 pixel wide 

screen had to address each individual pixel it would need 

a list of over 23,000 numbers. Dedicating a whole 35% of the available directory 

only to set the screen makes no sense. The real amount of space dedicated to creating images 

is only 12.5% of the available directory. But how did such a small memory create 

graphics? The key here is the use of tiles. These are the tiles for the game Super Mario 

Land 2, a classic Super Mario scrolling game. Each tile consisted of a square of 8x8. 

Rather than building the frame pixel by pixel, The Gameboy system rendered the 

screen in a three-step process. The CPU would first assemble a 

background made out of 32x32 tiles. But the size of the Gameboy screen only 

fits 20 tiles on one side and 18 on the other. So a viewing box has to be 

placed on top of this background. This view box could move along the 

background enabling smooth scrolling. It also has a local coordinate system 

that allows non-movable information, like lives or scores, to be visualised 

consistently in the same location. Movable objects like Mario or 

goombas that can interact with the background have a special 

name, they are called sprites. Sprites are just 8x8 pixel-wide tiles that can 

be flipped or rotated. For larger characters like Mario, a set of 4 sprites was 

needed to make the full character. Once the frame was ready to be visualised, the Gameboy went line by line setting the pixel 

values on the screen. This is called a line scan. This practice was a bleed-over from the 

NES, which was designed to be used with the tracing rays of cathode ray tube 

screens. CRTS work by altering the path of a beam of electrons to hit against 

a screen coated with fluorescent chemicals. This technique allowed programmers to create 

animations. At the end of each line scan, Nintendo gave the programmers the choice to pause the line scanning mid-frame to adjust 

the position of the viewing window. This is the intro to the Links Awakening 

game. This was all created using a static background. Once the background was assembled 

the tiles and the screen location were set, and the line scan would start. Here a pause would 

happen and the viewing window would be moved a tiny bit. Then the line-scan would restart the 

drawing and the end product emulated movement. The enemies in Link's Awakening like this or the 

intro to some games like TITUS were all created using these techniques. Even racing games used 

mid-frame pauses to create the curves in the road. This design ideology of simplifying 

also affected the audio of the console. The Gameboy came with only one speaker 

that was controlled by only 4 channels. Two square wave tone generators, one white 

noise maker, and a separate channel that could load any custom waveform that is 

stored in the game cartridge. That's it. Lets create a song by sending the desired frequencies and timings to the 

first two square wave channels. Now lets add our custom chipped triangle wave to the fourth channel with it’s 

frequency and timing parameters. Now, the final touch, a little 

percussion to highlight the beats, made with the white noise channel. This style of music is a huge part 

of our nostalgia and love for the Gameboy. I can hear the intro to the 

pokemon games in my head to this day. But games are more than just images and sounds, they are fully fledged stories 

that need data and space for logic. Of the 65,000 numbers that the Gameboy reads, only half of them are read from the cartridge. 

This worked fine for simple games like Tetris, where the full instructions and data needed 

to run the game was less than 32,000 numbers. Limited data was common in the 80's so game 

developers developed a technique called memory banking where the game divides the data 

into smaller sections or banks. Essentially, the game dynamically switches between 

different banks of memory to access a larger pool of data than the 

hardware originally allowed. The Game Boy's hardware can only read 

32 KB of data but Pokemon Red/Blue has a memory size of 373 kB. The data 

had to be divided into 44 banks. As the player explores different areas, 

the game seamlessly switches between these memory banks to load and unload the relevant data. This is controlled with a small 

chip inside the cartridge. When the Pokedex was opened the chip 

would access “Bank 2B” where all the 151 Pokemon had a 100-character description 

that was printed on the screen using tiles. If the player entered a Pokemart 

the chip would access Bank 1 to get the prices of each item. As the 

player moves between towns, locations, or activities, the game continues to 

manage these memory banks dynamically. The engineers in Nintendo made 

a choice that allowed them to get consoles into the hands of gamers 

around the world. For many, like me, it was their first experience of video games. 

With a launch price of just 89 dollars it was significantly cheaper than either of its two 

main competitors, and vastly cheaper to run. This ethos of player first is what defined 

Nintendo as a company. While its competitors focused on ever increasing hardware specs, 

Nintendo focused on accessibility. The Nintendo Wii with its motion controllers introduced 

hundreds of thousands of older people who weren’t familiar with traditional game controls to gaming. 

The Nintendo switch doubles as both a portable gaming console and docked home console, with 

detachable controllers that have allowed me and my friends to have impromptu mario kart sessions 

in airports and hotel rooms. Nintendo are masters of interactive design and the Nintendo Gameboy 

was a generational defining piece of design. Devices like the GameBoy were 

designed for a simpler time, when the only way to add software was 

a physical cartridge and the only way to input or output information from the outside 

world was a link cable. Decades later any device, even if only intended for gaming, will 

require some sort of account login connected to personal data and will constantly 

transmit your data with a variety of servers. In this hyper connected world, collecting user 

data is big business. Data brokers specialise in collecting every bit of public information 

available about you to sell to marketing agencies or in some cases, more malicious actors. Data 

breaches are getting more and more common too. You might have noticed an increase in 

spam calls and spam emails as a result, and getting out to these data brokers 

list is a tedious process that requires contacting each one by one and using 

the exact legal language necessary to force them to remove it. But you don't 

have to, thanks to our sponsor Incogni. Incogni reaches out to data brokers 

on your behalf with removal requests, and deal with all the annoying objections. I get emails for crappy Bitcoin and gambling 

YouTube sponsors to an email address I don’t have listed publicly anywhere. Clearly this 

email address has been sold to data brokers and they somehow know what I do for work. 

This email address is private for a reason, I don't want it spammed with nonsense, so I 

used incogni to start clearing things up and over the last couple of days it removed me from 

an astounding 58 different data broker lists, and is working on removing me from even 

more without me having to lift a finger. They then keep up with the pressure to keep 

your data out of their hands permanently. If you want to check them out, you can get a 

massive 60% discount off Incogni’s annual plan by going to www.incogni.com/realengineering, 

or by following the link in the description.