This is a one what I plan to be a regular series of post-mortems on my projects, he says, with zero confidence about the regular part.
I’ve decided to do these so that when I have no motivation and wonder how I even produced some of these projects, I can see these posts and remember “oh yeah, it was actually really painful and not quite as miraculous of a magnum-opus as you remember, it was incredibly painful, you made many mistakes, and the process was about as fulfilling as pissing into a ceiling fan; but you did it, and you learned.”
Now. without further rambling: a post-mortem of the design of the Nintendo Switch | FS |
1: The First Inspirations (and my hatred of the smartphone market)
I’d actually been super-interested in making a Nintendo product for several years, I’d always loved their hardware, but after a writing a university-essay on blue-ocean strategy on Nintendo, I was absolutely floored by just how many risks they’ve taken with projects, and the really unique ways they embrace both new (and old) technology: it’s more about the goal rather than the means for their products.
It was so fresh and unlike any other company; for example, I absolutely resent the phone-industry’s tenacity with which they pursue new technology, compared to what they actually do with it; it’s like:
“Sir/Ma’am, we’ve made screens, that BEND. THEY EVEN FREAKING ROLL-UP, literally STAR TREK DIDN’T EVEN HAVE ROLL-UP SCREENS wOW”
“This is clearly incredible, inspiring, a new-era of interaction, what to we intend to do with this ground-breaking innovation?”
“we thought we’d make a regular smartphone, but now it gets smaller (: “
In fairness they’re trying to show-off a sort-off ‘L’ -bend feature where you have two things going on either face of the ‘L’, but it’s a bit weak-sauce.
It’s a shame because a folding-screen could be a door into so many different ways of interacting with products, and for an industry pumping billions into R&D it’s amazing how safe every design is.
From a design standpoint, I think Huawei has done by far the best job of embracing what makes flexible displays unique.
Sure, it could get scratched to shit almost immediately, but I love the sheer boldness it takes with showing off the postive aspects of the design; how the screen is almost edgless, the beautiful way the screen tucks into the ledge/bezel, where all of the buttons are neatly layed out.
The first time I saw it I remember that unlike the rest of the competition, this sparked off so many new ideas for me on how a product like this could be used: you could have really unique AR experiences since wrap-around screen could project the camera view two-ways, you could have a really neat new form of wraparound video…
“You could… stand it half-way.. in a “/\” shape, and have.. two seperate… screens…”
I grabbed my notebook.
2: Flexible-Screen hinges | It’s My Way or the Huawei.
It’s fair to say there’s quite a few crucial factors that is stopping flexible screens from being widespread; price mostly, but since my product’s a prospective design, meaning this is designed to be released at a time when these technologies are viable (or rather, I just needed a get-out clause) , but there are a few others, like the crease, ease of scratching, etc.
One of those crucial factors, in-fact, the factor that drove most of the designing-phase, was the hinge-mechanism. I couldn’t realistically just lift Huawei’s hinge mechanism, for 3 simple reasons.
1) Some nerd lawyer with their patents and their international-coorperate-lawsuits might have something to say if Nintendo took the idea, so I shouldn’t either.
2) It would be a very short project.
3) I’ve no clue how the hinge (the falcon-wing-design or whatever) on the Huawei Mate X actually works, since very few teardowns exist, and if it’s anything like Samsung’s, I don’t want to know.
I already knew that the main things I needed with this hinge was ruggedness, and simplicity: this thing was gonna go through more than just a journalists prying on a show-floor, so it needed to be able to withstand that.
I also knew that I dirt and ingress getting into the system and under the screen was a current big bugaboo in flexible phones, so I wanted to seperate the screen from the device as much as possible, hence why there’s the system acts as a sort-of ‘spine’ for the console in the final version.
When it comes to designing simple hinges on folding phones, Motorola’s folding phone really takes the cake. Whilst reviews on the hinge are mixed, I found that the simplicity would serve me well, so I used it as a base-point.
Ignoring all other elements, the hinge is pretty much just two small rods that connect the phone’s screen ‘wings’ to the hinge. Simple enough.
My design would fold the other way and didn’t need the other ‘air-pocket’-related elements on the Motorola, BUT my design would also differ in that the two ‘wings’ of my screen needed to be moved in-sync with eachother from the ‘spine’ that is the console itself.
This is where the vast amount of what I hesitate to call ‘sketching’ began.
(I never thought I’d publish them, it was just to keep my ideas from dissapearing, no judgement, just love <3)
One of the elements you might see recurring is this, sorta overlapping ‘kidney-bean’ shape, with 3 black circles denoting the hinge pins.
This essentially acts as the dual-hinge from the Motorola Phone, but instead of the hinge’s operating like a book, the two outer ‘pins’ would be fixed to the main console ‘spine’, and the central pin (which is horizontally fixed, but can move vertically) would run through both ‘kidney beans’ on a slot, such that when a force pulls the central pin downward (in this case, a spring of some description), the bean & connected ‘wing’ are rotated about their outer pins, and both wings are pulled horizontal.
This is nice and simple, but quickly found that it has a pretty big flaw if used on it’s own.
This is an issue for a lot of hinge mechanisms, but even more-so here, is that the perimeter-length of the screen casing gets shorter when it unfolds, since the screen doesn’t magically stretch or contract, it’ll make a massive hump in the middle.
If it helps, imagine a laying each edge of a rectangle out flat so it’s one long line, and measuring the length of that line. Then repeat this with an identical rectangle, except this one has rounded corners: the one with rounded edges will be a different length, even if the general dimensions are the same (that’s a bad way of explaining it, but hopefully it makes some sense). Our screen when folded has a rounded top, which is made flat when we unfold the screen (our mechanism also overlaps, which is even more an issue), so our screen bunches up in the centre, which is no good.
Simple enough solution, have a compression spring where the hinge meets the screen-chassis that compresses when folded, thereby pushing the screen apart and keep it taut when it’s unfolded. In-fact these springs could be used as the means to push and unfold the wings entirely.
I make that sound like it was a simple fix, in reality it took nearly a week to work out and even now I’m not sure whether it totally works.. I tried everything from flexible rods that ‘push’ them apart, to utterly ridiculous hinge designs.
I should mention at this point that the screen isn’t being ‘pulled’ such that’d undergo any strain on the screen itself. The screen is, much like Samsung’s Fold, reinforced with two layers of a) perforated aluminium, and b) aluminium plating. The aim being that it’s these elements that bear the brunt of the stress.
3: General Dimensions | Size Matters.
Now that I had the rough hinge mechanism figured out, it was time for sizing.
From the get-go, I knew that if I couldn’t design every internal element, I could at the very least use the current Switch to provide a rough guide as to how much space I’d need to leave so that they’d fit suitably.
Since this now includes two folding screens that need to sit flush, so the thickness of this product would undoubtedly be a lot higher. Using iFixit’s Teardown as a guide, as-well as putting my calipers to every square-mm of my own Switch and extrapolating the internal part measurements, I worked out my rough dimensions, which included a needed 15mm extra thickness for the screens to fit-flush, bringing my total thickness up to a whopping 29mm, over double the Switch’s current thickness, this also accounting for better components and a larger battery (7000mAh over the 4310mAh current). The other dimensions were larger but comparable, 5-10% differences in length and height, which seemed appropriate for what is marketed as a more premium system.
Another thing that became increasingly apparent, is that orginally I figured the screen, when unfolded, would be 16:9, just like the two seperate screens. My brain just hadn’t made the connection that this isn’t how geometry works.
Conviniently, however, when including the top part, the unfolded version works out at almost 4:3, such that with a bit of tweaking to the radii, you could get it pretty much perfect.
As an aside, I think 4:3 BOTW actually looks nicer than 16:9, there verticality of the world feels a lot better realised.
Weirdly I didn’t really think all that much about the thickness (we shall soon see that this was a mistake), and proceeded to hand-draw a 1:1 drawing to use in my CAD model.
This quickly became some the most difficult part. I’d done a lot of the design on paper, but there were so many elements that I needed at-least the basic elements made virtual so I had a decent base to build off.
You can see right from the get-go that it’s way too thick, nothing about this screams ‘portable’ or even ‘practical’, but at the time I’d figured that it only seemed this way as an orthographic model, and that when rendered with joycons it’d be fine.
If you look really carefully near the ends of the top in this last shot too, you can make out the two rods either end connecting to the centre hinge-pin, this is what moves up and down to unfold-fold the screen. 2 tiny electronic locks inside the Switch keep them held in place until being released, reengaging the lock to keep the Switch in Split-Screen mode if needs be.
There’s also a small plastic ridge running the length of the top, this adds a bit of stability and makes it harder to feel the ‘kidney-beans’ beneath (although the metal reinforcement of the screen should help with that too.)
You can see in this next shot the exact moment when, after quickly mocking up a joycon to use as reference, that I realised the product was, indeed, a bit too chunky.
But, undeterred, I remembered a technique that I’d seen Apple use on their iMacs; where they curve the back so that, from most angles, you’d be forgiven for thinking that it was only 5mm thick, when in reality most of it is just obscured by the front face.
This… partially worked. From certain angles the console did seem a lot thinner, however by the nature of the console, from other angles it still looked far too thick.
I’d have to start the whole design from scratch (this was the 3rd time doing this), and rework every dimension again.
This was the closest I came to abandoning the project.
I had been out of uni for quite some months, only landing small gig-work in local-branding and feeling incredibly doubtful in any of my own abilities. This project was meant to be a kickstart to my confidence as a product designer, and all I had at best was a non-functional, ugly, useless, economically unviable concept of an product.
I chose to carry the project on because I knew that it’s potential was at-least worth exploring, and wanted to show to myself that I could still finish projects, even if I wasn’t proud of the result.
So, attempt 3.
So, this thing was 30mm thick, I needed to shave of a considerable amount; I also decided that I needed more space inside the product for better components and a thicker battery.
So, that’s more space inside a now-thinner product. Achieving both was desirable, but either would do.
First step was pretty straightforward: realising where I’ve redudant space.
The OLED display, even with all the reinforcements, is barely 3mm. I’d devoted 7mm to the screen-chassis, since I wanted it to be as resilient as possible.
I decided to scrap having the connection point between screen-console being two points near the centre, and move them right to the edges of the switch, such that on the right and left sides of the screen-chassis, the thickness was 6mm for stability and also for containing the 4 spring-loaded rods that extend the screen.
In total, the thickness was reduced to 24mm, this was as far as I could reasonably push the thickness before it became too fragile/lacking room internally.
However, as a result of this change in depth/layout to the screen-chassis, most of the Switch’s console actually increased in internal thickness by 4mm!
Getting back to where I was last time, once employing that sneak iMac tactic again, I felt the depth was suitable enough, and that the internal console space could also work.
I knew I wasn’t techincally minded enough to work out PCB boards etc. but given enough space, the hope would be people more talented than me would be able to work with the space given.
Many tweaks, revisions and adjustments later, I could begin to see the product coming together as I’d envisioned in my head.
I’d added some small, sprung-open kickstands that could potentially be fitted with a rachet so it springs open when pressed in once, and then can be clicked shut after.
One big issue that I’d initially glossed over, but needed to fix with increasing urgency, was the vent.
Since the top can’t be used to house the fan anymore, the bottom really is the only option, which is fine until you start using it in ‘split-screen’ mode.
Blasting hot air directly into the floor really isn’t any good, and I’m not overly fond of my solution of extending the vent round to the back of the product, since it’s not doing much more than blasting air at the floor, but also vaguely giving the heat the option to escape out the back.
Either way, at this point I was happy enough with the placement of eveything to begin rendering it up.
I was originally going to use Blender to render with, (which I recommend way more than Keyshot btw if you can take the learning curve, it’s free and the results are just as good, if not better), but with screens and displays specifically, it can be a bit more fiddly than Keyshot, so in this case I chose the latter.
Then it was just a case of finding some rad shots to best show the idea. I chose to keep the internals and hinge off a lot of it, simply because I didn’t feel it suited the style of Nintendo to advertise those elements over it’s fun-factor.
Like any of my projects, there’s a lot of things I dislike and like about the finished result:
- That I finished it. I think any other project would have made me abandon it completely. This project was coming up to 130-odd parts including Joycons, even without a lot of the internal features; and with that many potentially moving parts I was struggling to keep up, and so was Solidworks. Coupled with my lack of motivation, it was an important project for me to see the to the end.
- I like that even if I don’t touch on it much in marketing material, the hinge mechanism has some vague sense of actually being a mechanism; despite not being seen all that much, it was a nice simple solution that I’m like 65% could work, mostly since it’s Motorola’s with extra beans.
- I also like the time I took to make it easily replacable. The easily scratched back is a justifiable complaint, so this seems an fair trade-off.
- The general look. It still feels like a Nintendo product. There’s a slight retro-ness to it that really helps sell a more family-friendly feel.
- It still feels ‘concept-y’. I wanted in the beginning to have a product with fully layed-out internals, a full BOM, manafacturing plans etc., but I think I came to a comprimise that if Nintendo have however many employees working in tandem and millions of dollars of R&D to produce all of those elements, I could cut myself enough slack to say “as long as I take the time to be realistic with the design, and use the current Switch’s components to provide a rough guide on what sort of space & design is needed etc., that will be enough”.
That’s what I told myself, anyway.
- The front-facing buttons. With the top-part of the FS predisposed, working out where the rest of the buttons like ON/OFF, Volume etc. was a bit of a pain. Originally I wanted the product to only have buttons on the bottom or rear faces and keep the front face clear from clutter, but in the end, it made the most sense.
- I’m also really annoyed that the two ‘unfold’ buttons are even buttons at-all. It just seems a little off, I’d have prefered some sort of mechanical switch that didn’t rely on a micro-controller; but that’d be another moving part and another point of failure.
- The bottom vent. There’s not really anywhere else to put it and whilst I did push as much as possible to have the exhaust closer to the side, it’s not really that effective a solution and it still blocks air ventilation.
- The little kickstand things. They seem too flimsy, if I had another crack I might try and find a better way of supporting the screens.
- My split-lines are dumb.
- A whole bunch of other things here and there.
This project was very much a ‘crossing the line to see where it lays’ study.
I took the project with an idea to be as ambitious as possible for a fairly inexperienced newby designer, but to recognise when there were parts I couldn’t manage. Countless times I’ve given up on something because I knew I couldn’t do it, only to realise that someone with a different skillset could fill out the parts I wasn’t capable of.
No, the hinge isn’t perfect, the internals might not be workable, even the things I’d normally do weren’t as fully fleshed out as I’d be used to; ergonomics, material selection, manafacture etc.
I just wanted to rekindle the simple joy I’d had for most of my life, that came with the craft of having an idea and realising it as you’d envisioned it.
And, if nothing else, I can say that has been a success.