Live data from Hacker News

The tiny chip that powers Montreal subway tickets

righto.com

271–280 of 525 posts

Re: The tiny chip that powers Montreal subway tickets

#272

QR codes seem like a better ticket medium. This is like a 100 bytes, a qr code can be over 2kb This is a cheap plastic substrate with ink printing over the top. A qr code is just ink - or some other even cheaper printing process if you prefer. This needs a specific ticket technology supplier over the next 10’s of years. QR codes can be drawn on screens or printed on paper and you can change suppliers for every compon…

The problem with QR codes is that they are read-only. I don't know about Montreal, but Moscow public transport uses similar paper tickets, also Mifare Ultralight, except you can get them for different number of trips. When you use your ticket, the turnstile or validator would increment those one-way counters so that the next one would know how many trips you have left. You can't do that with a QR code without either…

Once I searched how Japan do the synchronization of the data in the IC card because I can't imagine how they handle all of the traffic for millions of people especially in the rush hour.

So the solution is the transportation card is writable, and each train station acts like a small data center. They sync the data periodically to the main data center.

I think the syncing tech is getting better, Japan train companies are going to experiment with QR code soon. So read only is feasible.

Re: The tiny chip that powers Montreal subway tickets

#273
post #32

> Presumably, the makers thought that making the card look like a smart card would help people understand it. The card actually uses an entirely different technology. It’s kind of the same, though. The physical communication layer is different, but the higher protocol layers are basically identical. Smart cards with contacts follow ISO 7816. These MIFARE contactless cards are ISO 14443 Type A cards, and their protoco…

It would be more analogous to your wifi antennas looking like 6" cat 5 cables with an RJ-45 painted on the end :)

Re: The tiny chip that powers Montreal subway tickets

#274
post #201

Earlier quoted context omitted.

At peak hours it can definitely be a problem, you really need the entire pipeline to work well. There's going to be somebody behind you most of the time, and you really don't want people stuck at the gates. In Japan credit card transactions routinely take a couple seconds. Imagine each person taking 5 seconds to go through the gate! I think what trials for credit card payments in transportation services there are doi…

Yes, the TfL system does a partial authorisation. It checks the card is valid and not blocked etc but doesn’t necessarily do a real-time authorisation all the way to the issuing bank. If you try to use a card that is valid but has no available balance/credit, it might work for the first ride but then be blocked when you try to use it for the return trip. Fares are batched throughout the day and you are charged once,…

Wonder how they block the card, my impression was that tokenization was meant to make it harder for card chargers to be able to track a card through multiple taps like that.

Re: The tiny chip that powers Montreal subway tickets

#275

Earlier quoted context omitted.

The problem with QR codes is that they are read-only. I don't know about Montreal, but Moscow public transport uses similar paper tickets, also Mifare Ultralight, except you can get them for different number of trips. When you use your ticket, the turnstile or validator would increment those one-way counters so that the next one would know how many trips you have left. You can't do that with a QR code without either…

Once I searched how Japan do the synchronization of the data in the IC card because I can't imagine how they handle all of the traffic for millions of people especially in the rush hour. So the solution is the transportation card is writable, and each train station acts like a small data center. They sync the data periodically to the main data center. I think the syncing tech is getting better, Japan train companies…

> So the solution is the transportation card is writable, and each train station acts like a small data center. They sync the data periodically to the main data center.

That's also how the two subway systems I'm most familiar with do it here in Russia. In both Moscow and St Petersburg, the data stored on the refillable tickets (Troika and Podorozhnik respectively) was thoroughly reverse engineered. People who did it, of course, tried to write them too — for example, you'd make a dump, enter a station, then restore the dump with your old balance. It worked, but only for a day or two, after which the card number was added to a blacklist that all turnstiles check cards against. The conclusion was that there's a server on each station that turnstiles talk to, that syncs with some central server each night (when the subway is closed), where all system-wide transactions for the day are collated, and if anything is off, the card is blacklisted.

Re: The tiny chip that powers Montreal subway tickets

#276

QR codes seem like a better ticket medium. This is like a 100 bytes, a qr code can be over 2kb This is a cheap plastic substrate with ink printing over the top. A qr code is just ink - or some other even cheaper printing process if you prefer. This needs a specific ticket technology supplier over the next 10’s of years. QR codes can be drawn on screens or printed on paper and you can change suppliers for every compon…

The problem with QR codes is that they are read-only. I don't know about Montreal, but Moscow public transport uses similar paper tickets, also Mifare Ultralight, except you can get them for different number of trips. When you use your ticket, the turnstile or validator would increment those one-way counters so that the next one would know how many trips you have left. You can't do that with a QR code without either…

I think semi-persistent internet is not such a big deal for a lot of providers. For Opal here in Sydney AU it uses a connection for some features. Cards can be anonymous or assigned to an account, the account shows all your trips online, so there's some required connectivity for that. The cards have a smart chip, but credit cards or phones can now also just be tapped directly, so it needs to match up when you tap off and determine where you started and how much to charge. This includes buses, which use 3G IIRC, and may drop offline sometimes.

To support the above, they did away with multi-trip tickets like you describe. It instead tracks and discounts once you hit a weekly limit (yes, you have to give up your privacy for this with an account or use your credit card directly). Not great for intermittent travel.

For systems like this the question is: if the internet is down and a few people get a free trip, does it really matter? You don't always need 100% accuracy if it makes everything else simpler, like removing paper tickets, printers, litter, etc.

Re: The tiny chip that powers Montreal subway tickets

#277
post #258

I am apalled at the crazy amount of waste this creates. Millions of tickets with chips inside them? My understanding is that they are one time use? In New Delhi metro, India, they used to use plastic tokens with these chips, but at the end of the journey, to exit the station you have to give the chip back. Nowadays, they use a printed QR system, and they have even gone paperless. I can buy the ticket with my mobile a…

I find the plastic tokens quite clever, but QR codes are not a great option for mobile ticketing at transit gates: Often it takes people forever to pull up the code in the right orientation and dial up the brightness enough for the scan to work. Not something I love to deal with when I want to catch a train.

If the QR code is in an app of some kind, any one I've seen maxes out the brightness while it's onscreen. Orientation data is built into QR codes anyway, so that's a problem with the reader.

Re: The tiny chip that powers Montreal subway tickets

#278
post #273
post #32

> Presumably, the makers thought that making the card look like a smart card would help people understand it. The card actually uses an entirely different technology. It’s kind of the same, though. The physical communication layer is different, but the higher protocol layers are basically identical. Smart cards with contacts follow ISO 7816. These MIFARE contactless cards are ISO 14443 Type A cards, and their protoco…

It would be more analogous to your wifi antennas looking like 6" cat 5 cables with an RJ-45 painted on the end :)

...and now I want one.

Re: The tiny chip that powers Montreal subway tickets

#279
post #2

Author here for all your NFC chip questions :-)

I'm curious about how the unique ID is programmed into each chip. Presumably all the chips on the wafer come out identical - at which point in the process are they individually selected and given a unique personality? Is it done with direct electrical contact that is then fused off, or using the near field link?

Re: The tiny chip that powers Montreal subway tickets

#280

Earlier quoted context omitted.

There's a noticeable delay between contactless cards and oysters. Some people I know prefer oyster cards simply because they open the gates faster, in spite of having to top them off all the time. I'm looking forward to not having to choose one trade-off over the other.

It’s noticeable if you’re used to the instant response of Oyster, but we’re talking about a few hundred extra milliseconds. Not something that bothers you once you’re accustomed to it. It’s still fast enough that it will read my Apple Watch before the gate starts to close from the passenger in front of me. One saved trip to an Oyster top-up machine will make up for a lifetime of contactless NFC latency!

>but we’re talking about a few hundred extra milliseconds. Not something that bothers you once you’re accustomed to it.

Wrong. With the traffic volumes normally seen in Tokyo, those few hundred extra milliseconds will cause huge delays at the fare gates. There's a reason the systems here use the Felica card which processes in 100ms: it's really needed for this kind of pedestrian volume.

Post reply on HN