Building a panel out of e-ink electronic shelf labels | Hacker News Reader

Building a panel out of e-ink electronic shelf labels

rbaron.net

229 pointsfrxx37 comments

Love the simplicity of the circuit with the daisy-chained uart lines and the use of a time-to-live counter to determine the target so you don't need to give each one an ID.

Another approach to this, seen in addressable LED strips, is to have two commands - SET and RESET - where the first call to SET after a RESET will change the LED colour, and the next ones will just pass the value to the next node. RESETs will always be passed to the next node, resetting the whole strip. This doesn't require the hop counter.

The APA102 is probably even simpler, it acts as a 1 byte shift register register per LED and passes the bytes down the chain without needing to really know the total length, there's some special case to latch all the data into the output registers. I'm fond of that design because it needs no timing specific input, you could easily input the data with push buttons if you cared to.

GE Color Effects G-35 Christmas lights from about a decade ago used a similar scheme, except that they boot up in an "enumerate" state and use the first transmission to each figure out their own unique address (or pass the messages down the line). After that, they can each be addressed individually.

(Edit: jgrahamc beat me to it!)

I'm super interested in these ESL displays, but couldn't really find any (the few I found are very expensive per piece).

Are there any special keywords or sources for finding them in Europe?

Ditto, I'd love to get my hand on some of these at the prices he found them ($1.4/unit). There's nothing in the usual places (eBay, AliExpress), and the raw e-paper units (no battery or micro-controller, or case) start at $5.

I love stuff like this. I once went way too far down a rabbit hole with addressable Christmas lights.

https://blog.jgc.org/2011/11/turning-ge-color-effects-g-35-c...

These are also serially linked but the first step is to give each lamp an address and then messages just get passed on until consumed by the appropriate lamp.

I never quite understood why shops use those electronic shelf labels.

Most of them can't be updated by someone in the store headquarters remotely. Someone has to take a bluetooth programmer device around and program each one by hand. And if they're doing that, they might as well just take a handheld label printer or sheet of labels.

Benefit of the labels is they're clearer than e-ink, brighter, and more colorful (more sales)

I notice that the electronic labels are more common in poorer countries. Perhaps they help prevent fraud in some way (eg. shop assistant 'accidentally' misprices an item, and then their friend buys it?)

Is it less wasteful than throwing away labels?

They only last 5 years. The an electronic label has to cost more both environmentally and financially than ~500 paper labels.

Well, I don't know about environmentally, but nothing suggests that they cost more (financially) than paper labels. If there wasn't an overall benefit, why would the stores use them?

What's the failure mode?

I doubt it, especially as they're not going to keep these labels forever. I would be interested is seeing real numbers, but I would bet you'd need A LOT of price changes to a product to offset the (monetary, carbon, ...) cost of the electronic label.

It depends on how often you change prices, how many stores you have, and where the paper labels would be printed. For many stores now who have an online and physical presence where there's constant price pressures from all sides may change prices every day or even more often than that. The cost of the electronic label can be significantly less than printing labels, shipping them, then paying a worker in every store to manually change them.

These labels have big readability issues, but the LCD ones are the absolute worse. There's a supermarket that uses them and I never go there as the prices are so hard to read, even at a perfect 90º angle, and especially the smaller price per weight/volume numbers...

Maybe it's deliberate...? Shops prefer shoppers to just buy the most prominent item rather than hunting around for a bargain... After all, the bargain items tend not to have a very high margin.

I think of these as helpful side effects. It might not have been an explicit goal, but they sure as hell don't mind it.

I've witnessed quite a lot of them using IR-based beacons installed on the ceiling, as well as a few of them that use an expansion slot on certain Cisco APs to piggyback a transmitter to change tags.

I always assumed that any stores using these had them networked together.

So the argument being that they don't have to pay someone to go around and update it and instead can just do it centrally?

I think the Amazon stores do this, and I would assume that if anyone Amazon would set that up properly.

I also thought I saw somewhere the idea that the displays let them tinker with the price more often instead of on a weekly basis when the flyers come out.

BestBuys are pretty bright and they support red for sales. They can also show stock and additional info about the product. Really nice actually.

Usually for big supermarkets they are changed over a radio frequency. So they do not need people to go around and manually program them.

You can also do cool time limited actions without having an employee go and manually change prices. For example in Kaufland stores in Germany at 19:00 the price of the Fruit and Vegetables is automatically reduced.

You can also do cool time limited actions

I would hate to put something in my cart that was $5 at the time I put it into my cart only to have it be $8 by the time I get to the till.

That's not how time limited actions work.

In most of Europe at least there are strict laws on pricing so if you make a discount, it can't just time out after 10 minutes, it needs to last the end of the business day or until the stock is depleted.

These stric pricing regulations is probably why electronic shelf labels were invented in Europe, to make sure there will be no discrepancy between the price on the shelf/discount and the price on the register, otherwise the store is liable.

Which in turn makes dynamic pricing even less attractive for the shop.

If they decrease the price, then there are some shoppers with the item currently in their cart who will pay less for the item than they expected to pay.

But the reverse is disallowed by law.

Which means that for every price change, the shop effectively loses out a little. Better not change prices too frequently!

The idea is to do a discount towards the end of the day to help clear stock that would otherwise go to waste because it has a short shelf life. Supermarkets do this all the time, but it generally involves an employee going around manually relabelling produce.

Or more likely in many cases, putting up temporary discount signs to be collected at the end of the day.

Still a time-saver to automate.

In the store I worked in decades ago they were (LCD not e-ink) updated by two means IR and radio. Not sure why it had both. They were also kind of an ecological disaster. Getting crushed in the back of the store, tons of batteries... We probably replaced tens of them every freaking day (really large store).

batteries for eInk will last much, much longer.

>Most of them can't be updated by someone in the store headquarters remotely.

Yest they can, since the '90s (in Europe at least). Where did you get this misinformation?

The programing device you saw is only for initial pairing or resetting the label, but other than that, they can always be updated via 2.4GHz radio or infrared base-stations (depending on manufacturer) mounted throughout the shop.

Source: used to work in the industry many, many years ago

I can tell you that I have information indicating that RIGHT NOW major American retailers are losing LARGE amounts of money in stores where they don't have electronic tags. There is simply not enough labor to keep pace with price tag changes given the inflation.

This is cool. I like E-Ink displays and also have plans to build a wireless display like this, but was/am stuck on how to receive data while also sleeping as deeply as possible, so the linked page about advertising looks helpful!

I'd run it off of a LiFePO4 (bioenno or something), but that's the ham in me (also I'd probably give it powerpole connectors just to further justify how much I spent on my crimpers haha).

Really cool!

Just came to say that I am loving rbaron’s write ups - thank you introducing me to them @frxx.

I stumbled across them today. Glad I could make your day better.

What a coincidence that this [1] popped up in my youtube feed today.

[1]https://youtu.be/CLmotCeMlq0

That's really neat! I wonder if (and how long) they would run off a supercap or harvested wifi energy.

Reply on news.ycombinator.com