EV Chargers Could Be A Serious Target For Hackers

Computers! They’re in everything these days. Everything from thermostats to fridges and even window blinds are now on the Internet, and that makes them all ripe for hacking.

Electric vehicle chargers are becoming a part of regular life. They too are connected devices, and thus pose a security risk if not designed and maintained properly. As with so many other devices on the Internet of Things, the truth is anything but. 


Sometimes, securing a certain system or device is as easy as disconnecting it from the network. When it comes to light switches and door locks, for example, we got by perfectly fine for years without accessing them online. However, in the case of EV chargers, it’s not practical. At the very least, connectivity is required to run payment systems. Additionally, being able to monitor the status and health of EV chargers remotely is a big help in keeping them available and operational.

Given that EV chargers must be connected, securing them is important. However, research by Sandia National Laboratories indicates that thus far, EV charger companies haven’t done the best job at protecting their systems. Researchers investigated a variety of attack vectors and vulnerabilities and found many areas where existing systems were simply not up to scratch.

Vehicle-to-charger interfaces were studied as a primary target. EV chargers generally communicate with vehicles over signals passed through the charge cable. This communication involves negotiation on power levels and charge time, among other details. However, it could also be a path for malware to infect an EV charger if the vehicle’s responses aren’t handled properly or sanitized. Researchers found that not only could data be sniffed from these connections, but that a low-powered attack with a software-defined radio (SDR) could stop a vehicle’s charging session from up to 47 meters away. These interfaces are often completely unencrypted, too, leaving them vulnerable to man-in-the-middle and spoofing attacks.

Card skimmer” by ThisIsntExeter

The user interfaces of EV chargers are also vulnerable. The simplest attacks mirror those used at gas pumps, where card readers are fitted with skimmer devices to capture card data. Other straightforward hacks include RFID cloning attacks for systems that rely on those for payment and account management. The smartphone apps used by charging networks can also be a target for hackers.

Much like other network hardware like printers, EV chargers often come with web-based configuration interfaces. And, just like printers, many of these inevitably end up accessible on the wider internet. Researchers found all kinds of EV chargers that had their configuration pages publicly accessible. Even worse, many had very weak credentials, often being configured with basic passwords or only requiring a serial number for access.

Even if the web services are secured, admin interfaces were still found to pose a serious risk. Often, EV chargers come with some form of diagnostic and maintenance port. This can be via serial, WiFi, Ethernet, USB, or Bluetooth. In many cases, chargers were found to have unneeded services like Telnet and FTP accessible over these interfaces, presenting a broader security risk. In some cases, chargers would readily allow firmware dumps or unsigned updates to be made, or exfiltration of log data. Worse, in many cases, the physical ports were poorly secured, providing easy access to malicious actors.

Potential Consequences

Vulnerabilities in EV chargers can present a variety of consequences in the event a device is compromised. Many of those consequences are minor in nature and limited in their scope. For example, a charger fitted with a card skimmer could lead to a criminal compromising the credit cards of a few hundred users. Open web configuration pages could let a hacker disable chargers or cancel user’s sessions in progress.

However, in some cases, the consequences can be far greater. A charger with compromised firmware could potentially be configured to disable certain safety features, placing users at risk. Chargers could be reprogrammed to energise cables prior to their safe insertion, leading to a risk of electrocution. Spoofed vehicle-to-charger communications could lead to a charger delivering excessive power to the vehicle. In such a situation, a properly-secure vehicle may shut down the connection to a rogue charger. However, ideally, both the vehicle and the charger would be secure enough not to pose the risk in the first place.

Sandia National Laboratory has prepared a “best practices” document on how companies can secure their EV charger hardware and operations. Credit: Sandia National Laboratory

There are larger-scale concerns, too. Modern EV chargers demand huge amounts of power from the grid. While home chargers of 7 kW and 11 kW were once the norm, today’s DC fast chargers run at power levels in excess of 350 kW. At those power levels, researchers fear there is potential to cause significant disruption to the power grid with the right attack. If an attacker could control enough vehicle chargers, simply getting them all to stop at once could threaten the frequency and voltage stability of the grid. At current levels of EV uptake, this isn’t a major risk. There simply aren’t enough vehicles placing enough load on the grid all at once to cause a threat.

However, as EV uptake rises, the threat increases. With vehicle-to-grid chargers becoming a thing, too, there is also the potential for an oscillating attack method. Huge numbers of vehicles suddenly demanding power from the grid, then seconds later attempting to feed power back in could make it difficult for authorities to maintain control over the power network. Controlling the power grid is all about balancing electricity generation with power demand from the grid. Sudden changes in demand and supply from an EV charger botnet could cause widespread blackouts and even temporarily drive certain infrastructure offline. Doing so would likely require control of hundreds of thousands of chargers near-simultaneously, though, so for now, power grid engineers can continue to sleep soundly at night.

With EV charger roll-outs only increasing in coming years, the time to get security right is now. Thankfully, Sandia National Laboratory has provided a document on best practices for the EV charger industry. Much of it is straightforward stuff, like physically securing admin ports, verifying firmware updates, and keeping web config interfaces behind strict firewalls. Having a list of best practices all in one place, though, is an easy way for companies to ensure their products aren’t unnecessarily insecure.

56 thoughts on “EV Chargers Could Be A Serious Target For Hackers

  1. > With vehicle-to-grid chargers becoming a thing, too, there is also the potential for an oscillating attack method.

    That in particular has worried me, the other potential pitfalls are already common in other parts of our lives or really just a nuisance. Still should be taken seriously and protected against, but card skimmers can be anywhere etc. It is far too useful a concept to throw out though, and IFF the right effort is put in the whole thing should be secure enough but it needs to be more secure than the Banks really – As if a bank gets breached digitally and ‘money’ and identities are stolen nothing in the world really changes, no physical hardware that may be destroyed and with no direct danger to life.

  2. Take the damned things off of the network, and eliminate digital communication with the vehicle.

    I do not buy “it’s not practical”. “It’s not practical” is code for “it doesn’t squeeze the last tenth of a cent out of a hundred dollar transaction, it doesn’t allow us to operate everything on the hairy edge of failure, it doesn’t allow us to pretend machines can handle every possible contingency without human support, and it doesn’t shift enough economic power to the already more powerful side of the transaction”.

    Payment systems don’t need to be in the critical path. At MOST they need a pure output from the charging system telling them how much energy has been transferred. And the ability to identify which particular vehicle you’re charging is an anti-feature and a spying disaster waiting to happen. Just take payment from the customer the way feuling stations have taken payment since the beginning of time, through means OUTSIDE OF the actual power delivery.

    Monitoring and remote control? Most electrical devices are really reliable… if you don’t overcomplicate them with stupid misfeatures. Pretty much everything on the grid has been running without remote monitoring for like 100 years. And monitoring doesn’t imply control anyway.

    Want to shed load? Shed the load. You don’t have to communicate about it; when the power goes out, the receiving system will get the hint. There is no realistic way to make it into a negotiation, so forget about that pipe dream.

    Want to actively discharge the vehicle for grid support? Forget it; you will never get that to work right either. You do not have the information you need about which vehicles can be discharged without causing grossly inconvenient or outright dangerous unavailability effects. Any attempt to get that information will fail because of insupportable domain complexity, unavoidable error, and pervasive, intentional system-gaming from all sides. The communication channel is not the problem; the basic intractability of the problem is.

    Digital vehicle to charger communication is a frill. Maybe not even a money saving frill. Supply a nominal voltage at open circuit. SIgnal your capacity by rolling off in a predictable way as the vehicle draws more current. If the vehicle seems to be drawing truly crazy amounts of current, shut the power off until it’s manually reset by a human hand using a physical switch. That same switch can keep cables from being “energised prior to safe insertion”… which shouldn’t be a problem anyway if your connector design isn’t idiotic.

    And it should be IMPOSSIBLE for you to “deliver excessive power” to the vehicle, which should be deciding for itself what it can take at the specified maximum voltage. And both ends should have non-programmable, non-overridable, non-remotely-resettable hardware overvoltage and overcurrent protection anyway.

    1. >eliminate digital communication with the vehicle.

      Rapid chargers literally take over the BMS of your car in order to pull off their stuff. The car’s own charger is only designed to handle a few kW of power, so the system bypasses the car’s electronics and talks to the battery directly for feedback. Without that, you couldn’t pull off 350 kW charging.

        1. It is to the folks that insist their car has to be ready to do another few hundred miles in less time than it takes them to drink a cuppa..
          This one might be horrifying, but at least its a sane thought process grounded in the engineering from the designers to get the product people apparently want, not a paywall or subscription service to the hardware you ostensibly bought or solely as a method of making the user into a product…

          1. The point of rapid charging is

            a) intermediate round-trips just beyond your battery’s range, when you don’t have the time for a two-hour “lunch break”. This is a problem especially in the winter when the driving range drops dramatically, especially for affordable EVs with smaller batteries, because you would struggle to travel from city to city and back within the same day – which for a normal petrol car would be trivial.

            b) unplanned trips, when you may start with a spent battery and have to go somewhere, now, without knowing whether you can charge at the destination or whether you have the time for it.

            c) people who have to park on the curbside and detached parking lots, who don’t have a local charging option, so they have to top up at the nearby charger. The charger has to be fast, or else it can’t serve enough customers and it will be difficult to schedule your turn at the charger.

          2. >The point of rapid charging is…

            Which is my largely the same as my point. I guess you can argue my phrasing leaves open the option to change their lifestyle a tiny bit more when you are taking longer trips, but seems like you felt the need to elaborate on the point of ‘It is’ worth it with reasons…

            So just to elaborate on a similarly well defined thought – I think you would find for the bulk of Europe if it is cold enough to actually loose meaningful range its cold enough the trip wouldn’t be trivial in anything on the roads – for most of Europe its almost never that cold, so when it is with almost everyone driving on the wrong tyre, not having a clue how to drive in slippery conditions etc makes the trip rather less than trivial anyway… And for those in the areas of Europe that do tend to such cold it seems like the distances between population centers are generally very short so even with a loss of range they would probably be fine (though I still wouldn’t choose a battery EV for there).

            As for your point C I don’t see that is likely to be a problem for all that long – as EV take off enough to make it required the lampposts and other street furniture will start sprouting at least slow charge points so you don’t need your own personal spot to keep topped up. At least in the UK its already happening now in some areas.

        2. It’s fun knowing I can replenish the range of my vehicle to the tune of about 800 kilometers in 90 seconds time. Takes longer to process the payment than dumping 50 liters of fuel in the tank.

          1. Does depend on how you measure it – how far did you divert to put yourself in the situation to fill up so fast? Yes the same will sometimes be true of battery electrics, but for much of the time the magic pixies should be able to fill ‘er up while you sleep/work/shop etc.

    2. >the vehicle, which should be deciding for itself what it can take at the specified maximum voltage

      It can, by telling the charger how much current it wants, so the charger can regulate the voltage to keep that current and you don’t need those extra expensive power electronics built into every car. This is not a trivial problem: they really do cost a lot, significantly more than what you’d be willing to pay anyways.

      Imagine if regular cars were built to siphon gas out of a dumb hose at the service station, using their fuel pumps backwards to lift fuel up from a tank several feet under the ground. It would take ages to fill up, unless you oversized the fuel pump by 10x.

    3. ” Pretty much everything on the grid has been running without remote monitoring for like 100 years.”

      And yet power companies have been able to sneak fiber broadband past the current “we don’t like competition” incumbents because it’s adjacent to their current monitoring aka smart grids.


    4. Resistive loads are easy. Hence 100 years of ‘no problem’.
      Voltage drops, so does current, hence power.

      Switching power supplies are a bitch, voltage drops, the controller draws more mains current, to maintain the regulated voltage.

      Big old switching power supplies are just difficult. Your power provider counts on allowing the voltage to vary while resources ‘spin up’ or excitation current is adjusted. If the grid in sum starts to draw more current when voltage drops, bad things will happen. The grid needs to get smarter before that day happens.

      This is obviously an arm wavy simplification. Hackaday isn’t what it used to be, ‘promised no math’ types are here now.

      1. > If the grid in sum starts to draw more current when voltage drops, bad things will happen. The grid needs to get smarter before that day happens.

        Is that even fixable? Honest question; I am no longer being arrogant for the moment.

        Even if most loads were willing and able to do something about it when you said, “Look, you HAVE TO draw less current”, it seems like the communication and negotiation latency would be a problem.

        I assume the way it’s worked in the past is that loads draw a bunch of current, and voltage (and frequency) just naturally start to slump. Naive generating devices can independently notice that and open up the throttle, but if you have a lot of them you have to coordinate so they don’t start to oscillate or something. It takes them some amount of time (seconds? minutes? hours?) to coordinate and crank up, and you hope nothing heats up too much or slows down too much while they’re doing it.

        If you have to coordinate with the *loads*, the time must get at least somewhat longer. First you have to notice that you’re *going* to need to let the voltage slump. And maybe you have to make some guesses about why and for how long. Then you have to communicate that to thousands of eally diverse loads. Maybe you have to somehow figure out which ones most need to keep getting full power, and which ones can actually slow down, and how fast they can do it… in the presence of incentives to lie. Then the power supply in each of those loads may have to signal something downstream before it can actually safely stop drawing so much current. And you still don’t want all of them to do anything abrupt at the same time.

        Can all of that communication, and maybe really meaningful computation, actually happen before something gets really screwed up, and without any horrible feedback effects? How fast do things have to respond? How do you know you haven’t built in some horrible cascading failure mode?

        Does it really boil down to needing large loads to power down gracefully if the grid can’t keep up, without negotiating or prioritizing or trying to tell them how long it’s going to be? And if they can’t do that fast, needing them to buffer some energy locally to cover their needs while not over stressing the grid?

        Because if it’s just going to be that kind of one-way downstream signal to stop hogging so much juice, isn’t that fairly similar to just using the voltage slump itself as the signal, and telling large loads they have to scale back their actual power draw if they see one?

        1. Basically ‘all of the above.’

          Small switching supplies (like wall warts) are built to new standards. That allow a lower regulated output voltage for short periods. Dimmer switches are also modified to not all ‘chop’ the same way, intended to make them, as a group, better behaved.

          Bigger smarter power supplies (like car chargers) notice the brownout and slow charging in response, pulling _less_ power when voltage drops. They’re not keeping anybody alive, so charge rate doesn’t really matter. They don’t volunteer to do this however, contractual compulsions are usually sufficient. Not the government’s job, they’d screw it up for sure.

          Utilities install more combustion turbines and battery packs to do ‘instantaneous’ power regulation. Some of which just generate ‘imaginary’ power (they just phase shift current/voltage).

          Demand side management (DSM) is generally done more quickly and with foresight from computer models with integrated real time system conditions (my personal experience, collecting the grid state, feeding the model and reducing the output data for the operator).

          As you note some DSM has a built in delay (e.g. it takes a half hour to gracefully power down). DSM contracts have prices attached that reflect the costs and timing. They are dispatched, more or less, like generation resources. (Protip, wrap your AC DSM in aluminum foil.)

          Utility control rooms are some of the oldest big computer clusters around. The guys sitting the stress desk have decades of experience. Sometimes things still cascade fail. Look at the big failures of the USA’s eastern interconnect in the 20th century. People get PhDs in control theory.

          If the weather forecast is good, load forecasts are surprisingly accurate. Which is why they sell weather financial derivatives. Florida in particular is a problem. They have for forecast gas burn by the time it takes for gas to flow down the pipelines. Miami has to get it’s load forecast right by about 4-5 days, or they have to flare gas, burn oil or shed load.

          It’s reliable because it’s time and stress tested, to the extent it is. Nobody bats 1000.

          1. EV chargers tend to stop charging in a brownout, because they can’t know whether the loss of voltage was because of a grid problem or a cable/connector issue which could spell the beginning on an electrical fire.

            That creates a bit of a problem if the grid is unstable, because leaving your car to charge overnight might leave you with no charge in the morning.

        2. >I assume the way it’s worked in the past is that …

          The way it worked in the past was that the power company used the power grid itself as a carrier wave to send a low frequency signal to devices which were equipped with load shedding capabilities, and they would be shut off just before an expected demand peak in a controlled fashion, so there wouldn’t be a big transient in the grid.

          The same equipment was used to signal day/night pricing so people could put certain loads like hot water boilers behind a separate meter and only run them when the power was cheap.

          The system was rather dumb, since it could only signal everyone in an entire grid section to turn on/off, which is why they upgraded it to a one-way radio system and eventually to two-way wireless, so they didn’t have to send the meter man to your house every month.

          1. Also, the reason they switched over to radio is because, when there is a grid failure like a tree fallen on the lines, you can’t reach the devices like remote operated circuit switches on the other side of the fault through the broken wires.

            These days with everyone adding solar panels and batteries, generators, at the end of the grid branches where there used to be none, you have to be able to reach over the fault and check whether there’s power still on the line before the rescue or repair crew gets electrocuted.

    5. OK, I yield, having been heavily educated by numerous people. You evidently can’t charge the thing that fast without some negotiation, and I should not be so arrogant about stuff I don’t understand. I apologize for that.

      Nobody’s said it has to be on the Internet, though….

    6. You do know petrol stations all have ANPR and don’t unlock the pump until the car is identified? And I believe they check the car is taxed and isn’t flagged by the police.

      Petrol theft is big business.

      1. “Petrol theft is big business.”
        Here in Minnesota (USA), the gas/petrol stations have switched to pre-pay at the pump. They lost revenue from their “convenience store” sales, but made up for it with fewer drive-aways.

    7. I’m happy that I am not the only one seeing that problem – we must stop sacrificing the technical foundation of our society on the altar of the god named digitalization. Now is the time to save the knowledge on how to do things without a global network of microchip and data distribution. This does by far not only apply to the power grid.

      That said, none of the mentioned “problems” is unsolvable without using a computer, some problems don’t even exist without one. Some long solved problems become *much* harder when using digital networks on them (for example power balancing, the introduced latency kills the system stability – see https://mainsfrequency.com for a primer; all the necessary data is already available to every instance connected to the grid, no need for internet).

  3. These systems aren’t ever going to be designed in a way that is focused on public good, as much as we want to think/believe they are.

    Cost optimization and rushing to market to beat your competition with the newest and greatest buzz word are practices that don’t just reward you for doing it wrong, they punish you for doing it right.

    It would be great if we could design it right and maintain it properly.
    But that costs money.
    So we will just end up spending 10% less for 25% of the product right now, and never think about it again until it collapses, explodes, steals enough money, or kills enough people and we can’t ignore it anymore.

    How else are all those poor executives going to guarantee they get their bonuses this quarter?

  4. So here’s a website that thinks “hacking” is a good thing, even though the definition is incredibly broad. But then this website uses “hacking” to mean malicious. Isn’t there something wrong here?

    1. Not really. It´s a fact that “hacking” has both meanings.
      Moreover, “bad” hacking can be used to demonstrate a design is flawed, which is “good” do raise awareness and bring corrections.
      It can be complemented with “black hat” or “white hat” to precise the *intention* of the hacker. But in itself “hacking” does not carry an intention.

  5. “Spoofed vehicle-to-charger communications could lead to a charger delivering excessive power to the vehicle.” – that is not how it works. A “charger” is not really a charger, it’s merely a switch. The charging circuit is in the vehicle.

    1. What you say is true for level 2 chargers. I believe that the Level 3 chargers are providing DC at a voltage and current that is negotiated between the vehicle and the infrastructure device. A compromised infrastructure device could agree to a delivery level and then instead deliver much more voltage or current. Safety mechanisms in the vehicle would *probably* prevent damage to the battery system, but possibly at the cost of needing replacement of high voltage fuses and disconnects.

      1. … and this design presumably saves 10 bucks on IGBTs or something in some of the vehicles. At the cost of 5 bucks worth of computer, $4.95 of NRE on the software, unknown but postponed and largely externalized failure and security breach costs, and some customers not being able to find compatible chargers after the next standards update…

        Or maybe it just let them end some some 18 hour committee meeting without having to hold a knife fight between the 240VDC people and the 480VDC people or whatever…

        1. Oh, I forgot the most important part. You’re really not paying for the computer and software, because you needed them anyway so you could charge the customer $100/month for “fast charging”. And the $4.95 comes out of another department’s budget.

        2. You seriously under-estimate the cost of the hardware.

          The power inverter to the motor isn’t built to handle that much power in most cars. Adding another one just to charge the battery would increase the price of the car by thousands of dollars, not “10 bucks”. They’re going for 900 Volts these days…

          Did you also know that the charging cable of the newer rapid chargers is liquid cooled on the inside? Otherwise it would have to be made too heavy to lift by a person.

          1. I think he might be talking IGBT as a disconnect (ie: smart fuse / resettable circuit breaker), and he does not mean an entire inverter meant to do the regulation. Such a thing would, yes, be very expensive indeed.

    2. Eh close but not exactly. There are giant 1KV 350A ABB power supplies in stacks in vaults behind the stations connected by red and black wires as big around as hotdogs each. They’re controlled by a fiber optic interface from the “dispenser” and will put out whatever they’re asked straight to the plug.

  6. The biggest threat with electric vehicles is that they are always online on the GSM network and they can be remotely told not to charge at all. Car will refuse to drive, refuse to charge, dashboard will show some cute emoji and message box explaining that this is “because of higher than usual electricity demand in , please try again later”. If you short RF section of built-in GSM modem and GPS receiver then after about week of no connectivity it’ll also refuse to drive, complaining with some cryptic error code and asking you to contact the dealer so they can replace entire part.

    EVs are nothing but overpriced rental scooters.

    Note: some of that also applies to IC cars but fortunately an engine is just a bunch of mechanical parts moving about. Unlike 150kW inverters It’s not a rocket science to create a custom ECU to read throttle, timing, operate injectors and fire spark plugs in the right order.

    1. I’d say ECU to control an ICE is rather more rocket science than controlling the Electric motor, and you don’t have to use fast charging at all – just plug it into the domestic socket where there is no smarts to negotiate with. It just takes longer.

        1. Indeed, not saying its all that hard, just that electric is generally easier, less things to worry about and make an error its less likely to be destructive.

          The point was meant to be that I wouldn’t call either a major challenge – both are well established technologies, but of the two ECU are in many ways the more complex.

    2. EU is mandating speed limiters on all new cars, called Intelligent Speed Assistance (ISA), on all new cars starting this year. They say it’s not a speed limiter because you can “temporarily override it”. Otherwise it acts as a cruise controller, not allowing you to exceed the local speed limit.

      It means cars have to be online and equipped with GPS to get the local speed limit data. You may use a camera sign detection system instead, but there’s a catch: it’s unreliable as heck so nobody will want to use it.

      They also made it mandatory to have an event data recorder (EDR) built into the car, which collects data on what the car thinks is the speed limit, and whether you’ve been bypassing the ISA over it. The insurance company can then punish you for repeatedly “speeding” whenever the computer gets the speed limit wrong.

  7. Yet another reason to hold off on buying an EV.

    These vehicles have been designed with the utmost disdain for the consumer, and there is no driver’s reason for all of this complexity. I don’t want infotainment systems that hide common driving controls behind multiple menus. I don’t want proprietary chargers that pop up in the middle of nowhere and have no paid staff to man them. I don’t want touchscreens and electronic door handles that fail in extreme heat or cold. I don’t want my car receiving OTA updates. I don’t want subscriptions for my heated seats or radio stations.

    I want a car that is light, economical, built to last, and can be quickly refueled at existing stops along the highways. Plenty of ICE vehicles fit the bill, but I’ve yet to be impressed by any EV, as it’s always included at least one or more of those deal breaker complaints above.

    All I wanted was EV propulsion in the cars that I enjoyed driving. But if it has to come with all those strings attached then you can forget it, chuckleheads.

    1. The Nissan Leaf matches pretty much all of that – works on most charging networks, all controls are physical buttons, there are no OTA updates, and the heated seats are built in like on a regular car.

      1. Company is on the ‘Are you crazy? Hell no!’ vendor list though.
        I’d sooner outsource a critical project to EDS (HP clownshow or whatever they’ve renamed themselves). I’d sooner install Oracle apps. I’d sooner ‘lick all the toilets in grand central station clean’.

        Nissan is just a marketing badge hiding French cars. Like Chrysler and FIAT, but longer term.

        Don’t buy any other Japanese cars using their CVTs either (Jatco). Hot garbage. Guaranteed to fail every 50k (Miles if lucky, km more likely).

    2. Many of those complaints are also true of new ICE powered cars, older ones too…

      And light is something a battery EV isn’t ever going to be in the near future, as practical range dictates a fairly substantial mass of battery. But not being light doesn’t really seem to have much in the way of drawback as the change in powertrain and how low the mass can be means it should perform similarly to the much much lighter car..

      I’d also argue that refueling is for most people entirely irrelevant – the longer range EV can already do a perhaps even a whole week of the daily communte without charging. And charging at home off the standard domestic plug even with the weak juice Americans seems so fond of can easily cover that. If you are the sort that does a several hundred miles in a single journey many many times then the battery EV are not for you (yet anyway), if you can’t charge it at home or work at all again perhaps not. But on the whole for most users the EV is the car you never have to go out of your way to fill up as the magic pixies did it while you slept.

        1. You still need enough range to be a useful vehicle which means a substantial energy capacity no matter how low mass the vehicle gets in total. And non of the options on the horizon right now are energy dense or light enough to best the efficient little ICE engine and the huge energy density of its fuel – so if that is your point of reference…

          Better than current battery pack does looks very likely in the near future though.

  8. And what of the reverse attack path, namely using a compromised charger to, in turn, compromise the vehicle, planting software that can control power delivery and any other system in the vehicle…

  9. On the bright side, there are enough competing (bickering) parties in play that any widespread attack is likely to be limited. The real fun will start when/if there are a lot more charge stations singing from the same song book. Maybe climate change will condense enough vapour to fill a truck stop with Tesla Semis dragging gigawatts. That could get messy.

    Also, whilst a more limited option , couldn’t the vehicles themselves present a method of attack or propagation? Data flows both ways and many vehicles ‘top off’ in work places or retail car parks.

    1. There are only three real flavors of charger out there today. In the USA we have CCS Type 1, and in the EU they use CCS Type 2. Go to any level 2 or 3 charging station and you can charge any* EV there.

      ( * Teslas not included.)

      And then there’s Tesla. Instead of giving up on the proprietary Supercharger format as they had said they would, they just pulled some typical E-Lon shenanigans. They announced last week “look, we’re opening up the Supercharger format for free, so everyone can experience the engineering joy that is Tesla’s chargers.”

      The car industry is likely to deafen the world with their collective yawns. Nobody but Musk and his fans give a rat’s ass about Tesla’s “standard”. A few EV charger makers will be happy to make Tesla aftermarket chargers, but that’s about it.

      Even most Tesla owners would rather be able to use any convenient charger anywhere, and not be limited to just the supercharger network. And to that end, they also covered their bets. Just in case the world didn’t roll over and suddenly convert all chargers everywhere to Tesla’s mighty “standards”, Tesla also announced a $250 dongle that will enable their cars to be charged with an *actual* standard CCS plug at any commercial charger.

  10. I’m just waiting for someone to hack a charger’s charge rate negotiation and effectively make a large scale version of a USB Killer device. Imagine going to a gas station and having your car burst into flames 5 seconds after you put the fuel nozzle into the tank, just because someone messed with the pump’s software.

    1. Has that been a problem for fuel hoses?

      Granting those asshole hacks require physical presence and hence video evidence.

      Chargers are limited by wire gauge. They’re pushing all the current they’ll handle as they sit. It wouldn’t be as easy as you think.

  11. As an EV driver I can promise that the state of affairs of the software of both the front and back end are pathetic. One prominent Southern California charger operator uses just the serial number of their RFID cards to authorize charging and and and blah blah long story short I can also use my LA Metro card to initiate sessions. Nice.

    One other interesting thing is that recently a law was put in place that requires chargers to accept credit cards, among other access requirements, and I can promise that it was implemented hastily. I would imagine that this isn’t helping the security station.

    Hey even Electrify America technicians say the software is a steaming pile and that most station faults are because of it.

    I feel that one nice thing would be to put a simple coin and bill acceptor in certain chargers in the right places ( laundromats, etc) where they’re used to handling cash. The fees and membership that the charging networks take from station operators are outrageous and it’s nearly impossible to recoup those without charging outrageous fees to drivers.

    And station maintenance has gone sharply downhill since pandemic, plus parts backorders months long once they DO get around to it. The whole public charging infrastructure thing is a wreck.

  12. Just wait until they start transferring GPS data from the vehicle to determine “taxation” and permit malicious governments to know where you go and when as de facto geofencing for parallel construction.

  13. The communication between electric vehicle (EV) and supply equipment (EVSE, aka Charging station) is different in different parts of the world.
    Europe and US use 15118 for DC charging, which is a real TCP/IP connection (with IPv6!). This can be encrypted, but encryption is only mandatory, if it is also used for billing (aka Plug-and-charge). You can also charge AC over this connection, but in most of the times, the AC current a EV can draw is determined by the pulse width of a 1kHz rectangular signal. Hence: no communication here, no real thread.
    Japan (chademo) and china (GBT) are using a CAN-Bus for the EV EVSE communication. There is no encryption and no billing feature.

    The bigger attack surface is given by the backend communication: OCPP. Written probably under the influence of a lot of genever by some dutch guys, it has no mechanism to use more than the ID of an RFID-Card to identify the customer. Since these IDs are mostly used consecutive by the operators, you can just add a bunch to your ID and charge on some other poor guys bill.
    In most of the times, it is not even encrypted, the only “security” is given by the fact that charging stations often are in private (APN) networks.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.