The cost to upgrade public transport ticketing system to contactless - how much?

A tap doesn't provide the card number. All it gives you is a one-time token that can be presented to the "bank" for payment of the amount you asked for. As it's a one time token, it's very hard for a merchant (i.e. TFI) to cross reference until the transactions have fully cleared, which can be 24 hours or more for foreign cards, and the card can refuse to supply the token if you've exceeded your limit. Storing the card numbers is the greatest sin you can commit in payment processing, so all of the systems guide you away from being able to do it. The card details are actually on the magstripe, but not the cvv2 number that's printed on the back of your card, and making transaction requests without cvv2 is prohibitively expensive... but swiping cards is only done in the technological backwaters of the payment industry like the USA.

Running without Internet was, and still is, essential for fast and reliable operations. Consider a train station turnstile at rush hour: it takes one second to pass the gate, so anything that takes more than one second to process will become the bottleneck in getting people onto the train.

Even today, online card transactions take around three to five seconds to process. That's "fast" if you're buying a coffee, but far too slow if you're trying to get a thousand passengers onto a platform. But the big problem is if internet connectivity is lost. The Mifare card systems (of which Leap is one) were designed to be usable as a standalone system, with no need for a live network connection. This was a great idea in 2004, and to be honest it's still a very good idea in 2024.

There are ways to get contactless card payments to run faster, but to do that you need to take on more of the job yourself than you do using a typical merchant terminal. You also need to take on more of the financial risk - taking on that risk is actually not a problem with something of the scale of Leap, which probably has 200 million+ transactions a year, but you still need someone to produce and maintain a secure system for you.

The reason is quite simply twofold:

• The shortfall of drivers in Capwell - Bus Éireann are 27 drivers short of the number that they need to deliver the timetable - it’s not that they’re sick, but rather that they just don’t have l enough drivers full stop - people aren’t viewing it as an attractive job in sufficient numbers anymore.
• Buses getting stuck in traffic and having to be curtailed to get back to where they ought to be

They are implementing a reduced timetable from Sunday week across the 202/a, 205, 207 and 220 routes which they hope should be deliverable.

The driver shortage is across the industry, and the housing crisis is not helping. All bus operators are suffering the same problem, but BÉ in Cork are suffering an acute staff shortage.

Modern smartphones are never really off, unless the battery is flat. "Off" is just a state where the communications and user interface aren't running, and the software does very little except monitor the power button and perform a very small number of other tasks, such as controlling the battery charging.

The NFC chips used in Leap cards can be phantom powered by the reader device… a wire coil collects enough energy from the terminal's RF transmission to power the card IC for just long enough to perform the operation. The transmitter in the reader is deliberately power inefficient so that the card can harvest that "waste" energy. (RFID tags are a simpler form of NFC, where the on-tag IC just broadcasts back a code when the coil is energised)

It would be possible to do the same phantom powering on phones, but it would need a standalone NFC chip and an additional coil (in addition to the one already used for wireless charging) and no phone maker would spend the extra money needed to accommodate this when the device is always on anyway.

I've a fair idea of other transport projects and software backend development, and I'd put the cost in the hundreds of millions over a 10 year term. My gut feeling is that €1000 per terminal is about 3x too high, even including fitting costs. The bare electronics and computing piece of the terminals can be implemented for about €50, or even less for vehicles that offer on-board wifi (and thus have an existing internet connection), but that's only a part of the hardware cost.

A lot of the costs wouldn't be technology per se: anything involving payment systems attracts a lot of very expensive certification and auditing costs.

If you were paying me to do it, I'm reasonably confident I could spec you a terminal at €50 bill of materials cost for the electronics. Any €10 microcontroller can handle the transaction logic, a €15-20 IoT modem will handle the comms (TLS included), and the rest of the cost goes on passives, power, NFC air-interface and the PCB. Just needs a ruggedised enclosure, a 12 volt supply and fitting. cost of enclosure and fitting would bring things closer to €300, but the point was that €1000 is not a realistic price.

The electronics part of it is really cheap these days. The big costs are in the installation, backend software interfacing, certification and maintenance. Bit those are not related to number of vehicles.

Regarding your Luas example, the ideal system is just "tap on, tap off", so you don't have to worry about ticket zones.. this is pretty much like how Oyster works in London (to pick an example people would be most familiar with). Your only obligation as a user is to tap your card (Leap or payment card) on entry to the bus, or on entry and exit at a tram platform or train turnstile... the backend software system does the rest, figures out where you've been and what you've used, and ensures that you never pay more than you have to for public transport.

Right now we've already got that level of convenience for Leap card holders, but adding general payment cards to that would make transport much more accessible to more casual users.

TfL take the money off your card at the end of the day, its not instantaneous. I assume there is some threshold value that the machines "charge" instantaneously and if that's unsuccessful it refuses entry?

You can't just cobble together a card reader from a kids science kit, install it on thousands of buses and hundreds of train & tram stations and expect it to work. There are major obstacles to that.

The first is certification. Before a reader can start accepting payments, it has to pass PCI PTS security certification to make sure that it can't be compromised to leak or (improperly) capture card details. Then the software on it has to pass acceptance certification by the acquirer. The second test is straight-forward, the first is not.

The second, and maybe the harder problem, is that is has to work in a hostile environment. The on-board units must work 18h a day, 7 days a week, attached to a pole that vibrates constantly (but not evenly). It has to carry on working even though it's being shook all day long. You also have to account for it taking a wallop several times a day as people hit it on the way on/off the bus, probably with the anvil collection in their bags. Depending on the age of the bus, it probably won't have a reliable power supply so you'll have to deal with that as well.

And once you solve all that, your software will need to receive updated configuration constantly, it might need new firmware occasionally (also acquirer certified). You'll need a management application to handle all that and a gateway to process the taps from the reader and get them to the acquirer. That gateway needs to be PCI DSS certified and, depending on the protocol used (transit has its own card acceptance protocol), it might need to be scheme certified.

Your reader will probably die at some stage (hardware has a tendency to do that) and will need to be replaced (not repaired - the security requirements make them hard to repair) so you'll need to budget for a supply of spare hardware and people to fit it. If it's lucky to live to the ripe old age of 5, it will probably see it's PTS certification expire so you'll need to build a new reader that meets any new requirements.

Luckily no-one takes that approach so they buy off the shelf readers that already have their certification and have dealt with all those problems. But you'll pay a premium for that. The management software is generally licensed on a per-device basis so that cost will scale with the size of the fleet.

The other part that people are ignoring is that the total project costs will presumably include the card processing costs for the lifetime of the contract. Acquirers charge a fixed fee per transaction and a percentage of the transaction value. Gateways will do the same. That means a small percentage of everyone's travel costs will be paid to those companies every day for over a decade and that will probably be included in the figure being mentioned by Newstalk. DB alone carried 146m passengers last year. If they had a (totally invented) average fare of €1, then the card processing costs could easily be over a million euro a year.

The NTA are already paying card scheme fees as part of the Leap card, it's just that those fees are paid when the Leap card is topped-up, not when the passenger travels. There's no change because of the new ticket type.