I really enjoy Alec's videos and quirky humour. Beware though, this kind of channel is a massive, addicting time sink. Other similar channels that HN might appreciate (the first two are quite well-known, the last one I found recently and is definitely bingeworthy):
Stuff like this is why I have difficulty giving up YouTube. I've given up basically every other algorithmic feed (except for HN, I suppose) but the amount of passionate and engaging education and curiosity content on YT will have me hooked for a while I suppose...
YES! I'm also really struggling with giving up YouTube due to all the wonderful content available there. I wish there was a widely-used alternative like Floatplane to double-post your videos for your supporters.
I watched that last night. Alec's a tremendous educator and come a long way since his early videos (which are also still good). I really enjoyed his deep dive into how dishwashers work :)
Frankfurt airport has a huge split-flap display to show departures[0]. I love looking at it changing the data, and hope they'll never replace it with a digital one.
Genuine question: what are some practical advantages of a split-flap display over, well... any other kind of dynamic character display?
The only two I could think of are:
- they don't emit light, so maybe there could be some kind of light-sensitive situation where this would be required (though in that case e-ink would surely surpass it).
- the sound passively-alerting those present to updates. It's not obtrusive or attention-getting like a bell or spoken announcement, so only those wishing to notice the updates can listen for the sound, while others can go about their business unperturbed.
- Visibility angle is much better than any other alternative. This is crucial when one big display is used in a very large room instead of many small displays.
- Image sharpness and contrast is excellent. You can read text on an old flip display far away that you would struggle to read on a modern LCD if you have mild astigmatism.
- Perfectly readable on a sunny day.
- They can show colored logos and symbols (e.g. train type, airline, highlight warnings)
- Better failure mode than LCD: single digits fail independently from each other and can be replaced. But it's less likely that the entire display fails.
- They are either working or stuck - no fading out like CRTs.
Nostalgia, really -- I have fond associations of these with childhood train journeys.
I could just about imagine a genuine use case in some kind of hostile environment. You could probably make one of these very resistant to water/temperature/radiation, with the controller kept safely out of the way.
I used to have one of these, but left it above the radiator. All flaps warped beyond repair :( So if you're going for resistance to harsh environments you'd have to source a better material for the flaps, at least.
Hm, after watching and listening to videos of these displays at airports and train stations (don't think I've ever seen one in person), I thought the flaps in the commercial ones were made of thin sheet metal because the flipping sound was somewhat sharper than the OP version.
I think they'll work... Somewhat? As the card is pushed through the face, there's a small amount of spring energy that it stores by bending, which gets released when it makes it through the opening and propels it forward + downwards. Normal gravity just adds to that motion and keeps the flap moving down. Zero g doesn't help as much but the card should already have momentum out of the gate. In zero g the problem may be more dealing with the rebound as the flap hits the bottom of the frame - it may bounce back up and float. May need some dampening or magnetic trick to keep it down...
The traditional uses of split-flap displays were sports arena scoreboards, airports and train stations.
i.e. places where you want a very large display that's readable in direct sunlight.
If you were equipping a train station or airport in the 1980s, your options were split-flap displays, CRTs with sunshades over them (maximum size ~20 inches), a multi-CRT video wall ($$$$$$$$), a vane display (numbers only) or a flip-dot display. And a lot of places that were equipped with them in the 1980s still had them in the late 1990s.
These days you'd use an LED display, or a sunlight-readable TFT.
Worth noting not all LED displays are sunlight readable. Typically, outdoor requires old fashioned through-hole, separate LEDS, and at least 1/4 duty pwm refresh.
> And a lot of places that were equipped with them in the 1980s still had them in the late 1990s
Many UK rail stations had them into the early to mid 00s. If I'm not misremembering, the boards at London Kings Cross were not moved to different tech until at least 2005. Manchester Piccadilly had them replaced around 2001 (at the time I was passing through there a fair bit, travelling between York & North Wales).
As someone who used to travel a lot by rail in the not so distant past, every station that replaced one of these boards was a downgrade. Most stations had maybe 1 or 2 of these signs with the rest of the smaller ones already LCDs, so it’s not like there weren’t already places to get the info you couldn’t show on a sign like this.
Part of the problem with the LCD signs is they show too much information. It’s not really an advantage to show the next 30 trains incoming on the big board when only the closest 2 or 3 have gates assigned. They also tend to be themed in colors that are harder to read (Amtrak does blue on white, which is harder to read than white on black). Some stations tried replacing the split flaps with single-character LCD displays to mimic, but they always ended up breaking faster (and never getting fixed) for some reason.
From what I recall in an article about these signs, industrial versions were ungodly expensive to begin with, but the real reason they’re being replaced is that the companies that originally made them are all out of business and you simply can’t get parts for them. It’s cheaper to replace an entire LCD panel than it is to fix one of the mechanical signs once and you don’t have to wait 6 weeks with a broken sign for parts to be manufactured.
similarly to an e-ink display, it doesn't take energy to show the current/last status, it's required only for transitions. If there's a blackout, the display is still visible.
- they don't require energy to maintain their state - a property which they share with eInk displays, but a split-flap display is unrestricted in size.
- assuming a fixed output selection, a split-flap display can also achieve far greater visual quality than other competitors like flip-dot displays
- defective units can be automatically detected (by comparing the position of the indicator magnet), whereas stuck pixels in flip-dot, matrix or LC displays require a human to spot them
Zero power needed to retain image, not encumbered by e-ink patent and manufacturing monopoly, simple and repairable, and you can build it yourself at home
Readability and ultra low energy consumption (plus: they are awesome and produce a soft sound like a distant waterfall).
Also, contrary on what I have seen on LED matrices or other RGB/pixel based displays: they won’t display advertisement which otherwise manages to sneak in – too close to information.
I wonder how their energy consumption actually compares. Electric motors moving relatively heavy physical cards is far from free. I wouldn't be surprised at all to find that an eink display beats the pants of of a split flap display in terms of energy consumption, but I don't have the necessary equipment to test it.
There is the advantage of extremely high resolution and crisp contrast. The digits can be printed in any colour and on (kind of) any texture you like. It scales also very well to very large displays.
This is great! Love seeing hardware hacks like this on HN :)
Beyond just being a cool project, the detail in the README on Github is really amazing and I've bookmarked this as a great reference for releasing open-source hardware as a git repository. CAD files, a bill of materials, detailed usage instructions, etc.
I've always loved these - this company makes one that looks great, but isn't open source (there is an API that has yet to be released): https://www.vestaboard.com
It's also quite expensive, but has a clever design where the individual split-flap modules can be replaced.
There are also custom ones you can have made (probably even more expensive), but they're mostly targeting enterprise: https://www.oatfoundry.com/split-flap/
It'd be neat to have one you could hook up to some market API to make a little stock ticker.
Glyphs can be loaded in an arbitrary order. For a given corpus (e.g. city names, American English digraphs) how does glyph order affect the time taken to update the display? Is there an optimal ordering?
If your display could accommodate 5% more glyphs, which would you add and why?
(With apologies to anyone revising for an exam this summer.)
We made a small project with all the 6 letter words from the Italian dictionary (as an homage to the inventor and producer of the split flap display: Solari from Udine) ordered in a way to produce “the least effort”.
You can only flip in one direction, so the time taken to update is entirely dependent on the previous state. Worst case you'd need to display the previous character (B -> A for alphabetical order) and would have to flip through the entire alphabet to get to it.
> You can only flip in one direction, so the time taken to update is entirely dependent on the previous state. Worst case you'd need to display the previous character (B -> A for alphabetical order) and would have to flip through the entire alphabet to get to it.
This only applies if you have only one copy of each letter.
You can consider having more copies of frequent letters mixed into the stack, if the stack's thickness allows.
Wouldn't that only help with the average case, not the worst case? Either you have multiple copies of every character (in which case you're still flipping through the same number of characters as the original stack), or some characters don't have duplicates.
Suppose you are optimizing for the fastest average time to display a complete word/phrase. That means you will be limited by the slowest letter in each word. But the letters are flipping in parallel, which means you can use the time waiting for the slowest letter to move the other letters to an optimal display.
So have the alphabet ordered in descending frequency. But put a second block of the most common characters half-way through. For each word you first calculate the slowest letter, which gives you a 'budget' of how many flips you will take. Then each letter uses that budget to get to the optimal location of the character it needs to display, probably as early in the alphabet as possible.
If you know which word is being displayed _next_, you can also try to get into the best position to move to in preparation.
The ideal alphabet ordering might be different depending on position in the word -- for instance, some letters are more common at the start of a word. And you could think about digraphs, taking advantage of Q usually being followed by U.
Scott Bezek has a great video on "How a Split Flap Display Works" [0].
He also sells blank "flaps" on Tindie [1].
The documentation on how to build the Split Flap is also really awesome [2], including how to cut your own flaps [3].
I've been following this project for a while. It's pretty awesome. I estimate about $10 per character (as opposed to the $180/character (!!) another poster mentioned through Alibaba), with a large portion of the cost being the custom cut "flaps" (in small quantities). It seems like the cost could get in the $2-$4 range per character but I couldn't figure out how to do it without a significant amount of labor involved in cutting the flaps.
FYI, there's also a company called Oat Foundry that sells large split-flap displays [4]. I seem to remember the cost being in the $10k+ region but I don't see a price and I could be way off base.
Oh, interesting, I see Vestaboard also has a (smaller) one for around $3k [5].
> I've been following this project for a while. It's pretty awesome. I estimate about $10 per character (as opposed to the $180/character (!!) another poster mentioned through Alibaba), with a large portion of the cost being the custom cut "flaps" (in small quantities). It seems like the cost could get in the $2-$4 range per character but I couldn't figure out how to do it without a significant amount of labor involved in cutting the flaps.
Laser cutting a bunch of sheets at once could be one option, but material selection might be a problem: PVC releases chlorine when laser cut, which will do all sorts of horrible things to optics/the meatbag operating the machine. Maybe a cutting plotter will work, since you're using really thin sheets?
I don't understand. Just send it to a service. You don't make your own PCBs do you? There's lots of things a hobbyist cannot do in-house, and even if you can just make far more sense (at hobbyist scale) to contract out.
There is the wazer, but it doesn't live up to its hype.
Laser cutting is straight out for the hobbyist, because of the chlorine, as you correctly point out. I think the way to do this is either with a CNC or with some die punch solution.
Doing a single cutout of PVC with a CNC would be fine, but doing 40x23x7 is probably not.
Doing a die punch is maybe possible but I don't really know how to create a custom punch and what tooling to use.
I've looked into buying ready-made solutions. Vintage split-flap mechanisms go for close to $1000 per character on eBay. There is a company I found in China that manufactures these, but its still around $300 or so per character. Displaying a tweet would cost close to $100k after shipping, etc...
The randomness factor is quite mesmerizing, each flap search ending at a different timing along with the mechanical aspect make this a very fun thing to watch. Old tech was so cool..
These screens are awesome. I did a design a few years ago, aiming for super low cost and easy build. Made a few units but it needs more refinement: https://splapr.com/
They're intended to be laser-cut, which limits the materials.
Those flaps and wheels really want to be injection-moulded from ABS, with round pins at the flap ends. That would be cheap if you were making enough of them.
(I miss TechShop. They even had an injection molding machine good enough to do this.)
yes i very much miss TechShop. i was a lifetime member. too bad they tried to grow too big, too fast. i think they misunderstood their business, at a fundamental level.
I've long hunted for split flap modules on aliexpress and I know folks who buy old train boards to get these modules. I wonder if there's someone out there scaling up the manufacturing? I'd love to buy modules at under the DIY ~$50 mark and build cool stuff at home!
My local makerspace has a really good laser cutter which members can use. It works well for that kind of usage as there's not too many ways you can kill it or yourself.
Generally cheap machines have compromises which mean more work for you - less reliable, less accurate, weaker laser, worse software - all of that means a lot more trial and error, a lot more burnt material and a lot more stuff you just can't do.
If you want to laser cut thicker materials, but only occasionally, you'd probably find it much more cost effective (and low effort) to either outsource or find a makerspace which has a decent machine. Laser cutting is quite easy to outsource as it is very easy to fully specify what you want (2D drawing + material specification is all you need).
Do you have a suggestion? I feel like the last time I looked into laser cutting, it was either SEOd hits that were crazy expensive or shops that looked dodgy. Nothing in-between.
Stay away from anything cheaper that 1k at the moment.
Generally speaking, for cutting anything above paper or balsa wood, you will need module that costs around $500, higher power modules need decent power supply.
There are some laser cutters from china that claim to have such power for less, but there is not jet "consensus" on any of DYI forums/reddit's that any of them are good jet.
I bought a 50*70cm "60"-watt unit online for about €2000 shipped. It is the size of a chest freezer. It works reliably, except that if you don't power it up with the laser armed, and instead flip the safety switch after power up, it will not fire the laser. I had one Hall effect endstop fail, but everything apart from the enclosure itself is COTS commodity stuff, so I just got a pack of new endstops on eBay and fixed it. Cooling with an Ikea tub full of water, vent with a hose out the window.
For home gamers, keep in mind the price of the machine is only one part. Large laser cutters are expensive to ship (from china), and take up a big space. Most materials will also require adequate ventilation.
As others have suggested, jobbing it out or joining a makerspace is often the best choice for occasional use.
We need to vent our laser cutter to the outside. There's also considerable danger that what you're cutting catches on fire, so they need to be supervised. It's not just the cost of the machine--they're not just "plug and go"
I wonder how easy and cheap i could make these? Would be neat to slowly grow a stack of these on the wall with some Pi controller and feed Wifi messages to it.
