Appeal to Authority: When ‘Industry Experts’ are Confidently Wrong /  by William Watts

I have been following Chace Barber on TikTok since long before he started Edison Motors. I’ve always enjoyed his Canadian-flavored insights into trucking, and I’ve tracked his grassroots foray into building a diesel-electric logging truck with admiration. He provides real insight into what it is like to be a logging truck driver, and has won many hearts and minds with his no-nonsense, pragmatic approach to truck electrification. 

Instead of following the likes of Tesla and Nikola into full electrification, he relied on  crowdfunding to build a prototype hybrid truck more akin to a diesel electric train than to a Prius. It’s a novel and sensible approach, particularly for the logging truck use-case, and his scrappy, DIY-build has won him a cult following. 

So I was startled by Barber’s response during a recent video interview posted to TikTok, when he was asked whether he would ever build fully electric trucks. He began with the standard,  reasonable concerns about the challenges posed by truck weight. But then things got weird. To explain his views on the limitations of renewable energy sources, he miscalculated the power generation capacity required to support 5000 fully electric logging trucks.

To be clear, it wasn’t the math that startled me—it was what his attempt at calculation revealed about his understanding of the distinction between power and energy. He doesn’t have one. 

Here’s the video clip, plus a transcript of his remarks:


To give you an example, logging trucks in B.C., that’s a niche industry, there’s about 5000 logging trucks that haul logs at 2.5 megawatts of consumption per day. That’s 12.5 gigawatts of power. Site C Dam, has been under construction for the last, oh I dunno 15 years at a cost of $20 billion and that has a 1.1 gigawatt. So a 20 million dollar dam that takes 15 years to build has a 1.1 gigawatt capacity and logging trucks, just logging trucks alone are using 12 and a half gigawatts. You would have to flood an area of land the size of Wales to produce that hydropower.

This seemingly prepared argument is wrong from start to end, and the errors are fundamental. Notice the only units Barber cites are megawatts and gigawatts, which are measures of power. Energy, in contrast, is power over time, and is measured by watt-hours, kilowatt-hours, megawatt-hours, and gigawatt-hours. So 5000 logging trucks that each consume 2.5 megawatt-hours of energy per day would together require 12.5 gigawatt-hours of energy per day. He is correct that Site C dam will generate 1.1 gigawatts of “capacity” if he is referring to peak power. But he clearly confuses this for energy generated over the course of a day and completely misses that, over 24 hours, the dam could thus be expected to generate up to 26 gigawatt-hours (1.1 gigawatts x 24 hours) of energy per day. It will not run at full capacity all the time, so expected average output is closer to half that. Even so, his example dam could quite neatly power all the logging trucks in B.C. His claims about needing to flood an area the size of Wales to power this fleet are complete nonsense.

But it’s this nonsense that gets repeated over and over to reinforce people's biases against electric vehicles. A look through the comments on the video shows the depth of the damage one man’s confident misunderstanding can do. Whether the debate is about the relative environmental damage of EV versus ICE, the real environmental impact of energy that comes from fossil fuel sources, or the potential ‘strain on the grid’ resulting from EV adoption, people who have public credibility but lack understanding as to the technology and terminology are exacerbating an already confused consumer base with misinformation. 

I have written before about the state of  EV illiteracy,  which will have to be addressed through public education on basics like energy and power. Most people are unfamiliar with the terms, units and jargon associated with EVs and hybrids—and there’s no shame in that. There is a lot to learn, and kilowatt-hours are nowhere near as intuitive as gallons. But those holding themselves out as sources of information and as industry insiders should expect to be held to a higher standard and to be called out when they leave folks more confused than when they started. 

Of course, thanks to social media, these errors have a way of compounding. Back in September, for example, I saw a tweet from Tesla blogger Sawyer Merritt claiming that the new hybrid F150 has “7.2kWh onboard power” and quoting a presentation by a Ford representative who joked that owners could “help a stranded Tesla driver power up their car on the side of the road.” This claim was strange on its face, not just because a kWh is a measure of energy, not power, but also because that would be quite a large battery for any non plug-in hybrid. After the most cursory research, I confirmed that Merritt was wrong; the F-150 hybrid has a 1.5kWh battery and can provide 7.2kW of power continuously. I commented in his tweet thread and went on with my day. 

About an hour later, however, Merritt doubled down, tweeting the same incorrect unit but with a link to a video of the presentation. Watching this, I realized that the error originated with Ford. I was stunned to see none other than John Emmert, the General Manager of North American Trucks for Ford, in a presentation at the Detroit Auto Show, repeatedly use kWh instead of kW when referring to the power that the truck battery was capable of providing. Fortunately the deck projected behind him used the correct units, but it wasn’t hard to understand the ensuing Twitter confusion. 

In a social media-powered world, one person’s confusion becomes the evidence that thousands use to inform their decisions or reinforce their biases. If consumers are to understand this new world of vehicles, we need the supposed experts in the industry—the people engineering, building and selling these vehicles—to get the most basic concepts right. At the very least, they shouldn’t be staking out positions and drawing false conclusions based on a fundamental misunderstanding of how stuff works. 

For Road & Track: What You Need to Know About EV Charging at Home /  by William Watts

For October’s Road & Track article I covered the ins and outs of EV charging at home, from evaluating your home’s power to purchasing and installing a charger. A section on the advantages of dynamic load balancing is below.

”Of particular value for those who have 100 amp service is the addition of dynamic load balancing. In all homes, you are actually able to have more amperage worth of circuits in your breaker box than your total service can handle. This is acceptable because of ‘load diversity’ or the fact that it is unlikely that you use all of the electrical devices in the home at the same time, so 150 amps of connected load is highly unlikely to reach even 80 amps in practice. Dynamic load balancing takes maximum advantage of this, allowing your EV charger to monitor your whole home’s electricity use and ramp its own use up and down to ensure that your home stays under its allotted maximum load. This means that you can potentially connect a more powerful charger to your home without worrying that you will trip your main breaker.”

For Road & Track: How Long Does it Take to Charge an Electric Car? /  by William Watts

In September I continued my EV series for Road & Track, this time covering how long aspects that influence how long it takes to charge an EV, both when at home and on the go. A section discussing true vs peak fast-charging speeds is below.

”At higher power levels, lithium-ion batteries charge on a curve, typically receiving less charge as they reach higher states of charge. This phenomenon is apparent with smartphones, as they charge from 10 percent to 20 percent much faster than from 90 percent to full.

This curve declines so sharply that the commonly stated metric for fast-charging speed is the time it takes for an EV to charge from 10 percent to 80 percent because ‘fast-charging’ above 80 percent isn’t fast at all. Below, you can see the different charging curves for various EVs, showing the relationship between the car’s state of charge and the power that its battery can receive.

It’s important to know that how fast a car charges in the real world is not a function of the peak power the car receives. What matters is the average charging speed”

For Road & Track: What it Really Costs to Charge an Electric Vehicle /  by William Watts

I recently wrote another article about EVs for Road & Track, this time exploring the various factors that determine how much it costs to charge an EV, whether at home or on public charging networks. A section about EV-specific rates offered by utilities is below.

“Inevitably, EV ownership means that total electrical usage will go up, so it may be difficult for customers to take advantage of the allowance-limited rates. Electrical utilities generally want to encourage EV adoption, so many, like PG&E, have introduced EV-specific rates. These rates generally trade a higher peak rate for a much lower off-peak rate, recognizing that EV charging can be as much as half of your electrical bill but can be done easily during off-peak hours. Under this scheme, our 1000-mile Model 3 example charged off-peak would cost just $67.50 per month—quite a savings.”

For Road & Track: How long do EV batteries last? /  by William Watts

I recently wrote an article for Road & Track diving into the data we have so far on EV battery longevity, the factors impacting battery degradation, and how drivers can, with small behavioral changes, get the most life out of their batteries. A section on how charging habits can affect battery longevity is below.

“A third way owners can preserve their EV batteries is to limit the speed at which they regularly charge. Very high-speed fast charging may be convenient but can cause a phenomenon called ion plating on the anode. That’s bad: lithium ions are supposed to migrate into the anode's graphite layers, not accumulate on the anode’s surface. Plating seriously degrades battery performance over time. Fortunately, for most owners, fast charging is rarely necessary and can be reserved for road trips, and the occasional fast-charging top-up is not going to measurably damage your battery. On the other hand, if public fast chargers are your primary charging option, because you live in a city or otherwise lack access to home charging, consider whether you really need the fastest 350 kW chargers for your weekly charge or have the time to opt instead for slower, and less damaging, chargers in the 25 to 70 kW range. Some years ago, Tesla taxi fleets in Amsterdam were rapidly degrading their batteries through daily use of 120-kW Superchargers. Tesla later provided taxi fleets with private 60-kW Superchargers and has since installed 72-kW “urban” superchargers in dense cities around the world where owners may rely on them as their primary source of energy.”


Resetting the Standard: an American Coalition Chooses NACS /  by William Watts

Late last month, Ford and Tesla announced a two-prong partnership that rocked the North American EV charging world. First, starting in 2025, all Ford EVs will be equipped with only the Tesla-designed NACS connector, in place of the current CCS connector. Second, all Ford vehicles will have access to most of the Tesla Supercharger network, through an adapter for older models built with the CCS connector.

Before I had a chance to post this piece yesterday, GM announced that it is following suit. GM will also build NACS into all its EVs starting in 2025 and will offer an adapter for legacy vehicles, with access to the Supercharger network starting in 2024.

This is a huge deal. To date, discussions about America’s EV future have been heavily bifurcated. On the charging front, despite its market dominance, Tesla has been treated less as a standard bearer than as a prominent exception. For years, it was widely accepted that Tesla had largely solved the problem of EV charging, but only for its own customers. The proprietary nature–and sheer success–of the Tesla connector and Supercharger network meant it was easy to set Tesla aside and focus on solutions for the rest of the EV market. Tesla’s direct sales model and anti-union stance only served to further set it apart, with dealership lobbies pushing legislation banning direct Tesla sales altogether in certain states and making Tesla an unpopular candidate for federal funding—or (for a while at least) even federal acknowledgement.  

The Ford and GM announcements change all that. With a stroke, the three largest American EV manufacturers have committed to NACS and the Supercharger network, which can no longer be dismissed as ‘Tesla only.’ It’s a seismic shift and there’s a lot to unpack here. Below I consider how Tesla got to this point, what this means for federal funding, and why this is a good time to have a frank discussion about connector quality. I also comment on the fundamental changes Tesla will have to make to accommodate the higher voltage of many non-Tesla vehicles.   

What Tesla Did

The Ford and GM announcements were made only two weeks apart but are not an overnight development. Make no mistake: Tesla has been building to this moment for a long time. Last November, when Tesla first opened up its connector design and renamed it the North American Charging Standard, it did not simply invite other manufacturers to copy its homework. It revealed that it had crucially, but quietly, changed its communication protocol to match that of the CCS standard and laid the groundwork for a particular solution to the developing contest between the continent’s two major charging standards: one that matches CCS’s open communication protocols with the fundamentally superior hardware design Tesla pioneered and has been using for almost a decade. Around the same time, Tesla began to offer CCS adapters for its cars, and developed a hardware retrofit to older vehicles to allow them to communicate using the CCS protocols and similarly enabled the newer V3 Superchargers to speak the same language. These steps towards interoperability largely went unnoticed, but made the recent moves by Ford and GM possible.

Broadly, the announcements have been received positively by the EV-owning public. The Supercharger network, the largest and most reliable DC fast charging network in the world, has historically been a walled garden, available only to Tesla owners. It was only late last year that Tesla opened select Superchargers in the US to the general EV public by equipping them with a ‘magic dock’ Tesla-to-CCS adapter, with the promise to equip at least 7,500 with this adapter by the end of 2024. Ford and GM’s agreements with Tesla go above and beyond that. They will be the first manufacturers to outfit their cars with the NACS connector for service by Tesla’s charging network, but they almost certainly won’t be the last. 

The Impact on Federal Funding

All this raises pressing questions about the effective exclusion of the NACS connector and Superchargers from the major federal funding programs authorized under the Bipartisan Infrastructure Law. The $5 billion NEVI, $2.5 billion CFI, and $51 million Ride and Drive Electric programs have all centered on supporting the CCS standard. Tesla, despite being by far the most popular EV manufacturer, has had its standard largely excluded from these programs. 

These exclusions reflect a long-running two-track approach to the development of EV charging infrastructure, under which Tesla has been treated as a closed, self-sufficient ecosystem and public funds are funneled into Tesla alternatives. For example, the Volkswagen emissions scam settlement that funded the creation of the Electrify America charger network specifically called for chargers to be built with the CCS and CHAdeMO connectors, though the only car in America built with the CHAdeMO connector is the Nissan Leaf. As reported by Verge last year, only about 170,000 Leaf EVs were sold in the US over the last decade, while 2022 alone saw the sale of 564,743 Teslas. Notably, since last year EA has been phasing out CHAdeMO support from future sites, except in California where the state’s clean air agency (CARB) still requires it. So it’s hardly surprising that the NEVI program, a first-of-its-kind commitment of federal funds to the installation of EV charging infrastructure by private entities, imposed similar specifications requiring only CCS and allowing CHAdeMO connectors. Following Tesla’s NACS announcement in November, the NEVI Final Rule allowed NACS connectors to be installed at qualifying sites if and only if each charger also had a CCS connector. 

Tesla, Ford and GM together control almost three-quarters of the EV market. Under the current NEVI rules, most American car buyers would be unable to use chargers funded by American taxpayers without first purchasing an adapter. This is obviously problematic. NEVI and other programs will have to revisit their technical standards to require direct support of NACS. 

That said, it has taken years to develop the technical specifications set out in the regulation, and the funding is already flowing. Charging hardware manufacturers have spun up ‘NEVI Compliant’ offerings and have started to move their production to the US to comply with the made-in-America requirements that will be in effect in 2024. A shift in the technical requirements will disrupt those plans and supply chains, but the sooner the changes are made, the better. Disrupting the program while it is being deployed will inevitably be less painful than going back to retrofit hardware that is out of date before it is even installed.

Not everyone would view amending the funding legislation as a positive, however. CharIN, an industry group that exists to promote the CCS standard, is predictably dismayed by the total disruption of what the FHWA described, as recently as February, as the industry’s verified “mov[ement] to adopt CCS as a market standard.” A week after Ford’s announcement, CharIN released a statement of its own criticizing Ford’s move. CharIN claims NACS is not a standard at all, and explicitly calls for its exclusion from federal funding. But the statement also acknowledges CCS’s continuing reliability problems and fails to persuasively address the reasons Ford and GM have pivoted to Tesla for their charging solutions: a better connector and a bigger network. 

And it doesn’t seem like NACS’s advantages are going to dissipate any time soon. Most know It’s no secret that Tesla has the largest DCFC network in the US, with 63% of the ports in the country, but the more telling statistic is that Tesla continues to deploy chargers at a rate that far exceeds the rest of the market combined. In view of these trends, it’s obvious why Ford and GM would want their customers to be able to tap into this network. It will soon be a significant competitive disadvantage not to be a part of this coalition.

So what is next? I think it’s likely that over the next six months, we will see at least a few other manufacturers adopt NACS. Rivian and Stellantis will almost certainly move over very quickly. If that happens, there will be no way for federal programs to justify not accepting and supporting NACS in their deployments. The European and Japanese car makers may prove to be slower to react, but if the federal funding shifts to accommodate NACS, they will likely fall in line. In my opinion this would present a best-case scenario, one where all of the cars in the country would be able to use all chargers, and where the technical standard would be world-leading. 

Charging Standards

This shift in the EV charging landscape–and specifically, Ford and GM’s apparent willingness to backtrack on their own EV programs to take advantage of the merits of the NACS connector–should force a long overdue conversation about what a US standard could and should be. I mean that from the perspective of technical superiority and functional design. 

If the U.S. government were to mandate a standard connector for all vehicles (for example, by conditioning certain EV manufacturer subsidies on the connector’s inclusion), the NACS would be an excellent choice for technical reasons. This would not be the first time a government has mandated use of a particular plug type. The UK mandated CCS (a different version than the one we use here), and China has its own standard called GB/T. These standards are, in my opinion, inferior to NACS due to their large size and lesser current carrying capacity, limitations that result from designs that require different pins for AC versus DC charging(or in the case of CHAdeMO and GB/T, entirely different connectors). But Tesla proved it possible for cars to receive AC and DC on the same pins, resulting in a compact connector that has unignorable advantages when it comes to reliability and accessibility. A world-leading standard doesn’t have to look exactly like Tesla’s NACS, but it’s hard to imagine a rival standard of similar quality emerging and overtaking Tesla’s decade-long lead.

As I’ve mentioned previously, there are essentially three charging connector standards in the US. CHAdeMO, a protocol created by Japanese car manufacturers, debuted in 2010 and was the first widely deployed DC fast charging connector. The popularity of the Nissan Leaf helped drive the popularity of CHAdeMO, but its relatively low initial charging power 62.5kW meant that it was not a real solution for long range EVs.

CCS has been the leading ‘open’ standard in the US and Canada, and has been widely adopted by manufacturers. The CCS stands for “Combined Charging System” and is so named because it quite literally combines the J1772 level 2 connector with two additional large pins for DC charging. This bulky arrangement allows cars to be built with just one connector capable of handling both AC and DC charging (in contrast, CHAdeMO cars require two separate ports). 

Around the same time CCS was developed, Tesla developed its own connector and, in 2012,  introduced it with the launch of the Model S. The Tesla connector (since rebranded the NACS connector) proved to be superior in pretty much every way. It is less than half the size of the contemporary CCS and easier to handle, but has higher current carrying capacity. The main way Tesla was able to achieve this was to use the same large pins for both AC and DC charging. This reduced the overall pin count and therefore size of the connector. The one downside is that Tesla owners have always had to use an adapter to use the public networks, whether those be J1772 Level 2, CHAdeMO, or CCS chargers. 

Image: Tesla

Looking at the two connectors, it's obvious which is more elegant, but the benefits are not just skin deep. The larger CCS connector must be aligned just so to connect properly, and is more unwieldy to use, presenting a challenge when stretching the cable close to its limit to reach a charge port that is far away from the charger. This can be particularly challenging for smaller individuals and those with disabilities or physical impairments. It is also easier to package the NACS connector in a car, with Tesla famously hiding its port in the taillight of its vehicles. 

Limited Voltage

For the reasons outlined above it may seem pretty clear that moving to NACS is the right decision for Ford and GM, but there are certainly some potential kinks to work out as well.

The major immediate issue is the question of voltage. Tesla vehicles, and therefore Superchargers, operate on a 400V battery architecture. While this means the NACS connector itself is capable of handling higher amperage, and therefore the power levels on higher-voltage cars, the inverters that power the Superchargers will need to be upgraded to support those voltages. This may be relatively straightforward for the V3 superchargers, which use technology developed through Tesla’s commercial solar business, but older V1-2 Superchargers will almost certainly not be able accommodate higher-voltage cars (as they are essentially rack-mounted versions of the vehicles’ on-board chargers) and will need to be replaced entirely. The result is that vehicles with higher-voltage packs may not be able to charge at all on legacy Superchargers, or may charge very slowly. 

Ford’s 2025 switch to NACS won’t raise the voltage issue if Ford intends to continue with its 400V architecture for the foreseeable future. This would represent a departure from GM and Dodge, but would allow Ford to take maximum advantage of Tesla’s existing charging network. Most would argue that 800V systems represent an advantage, particularly when it comes to charging, but the high current-carrying capacity of the NACS connector largely negates that advantage (at least for smaller vehicles). 

The more likely possibility is that Tesla plans to move to support 1000V DC charging with its Superchargers in the very near term (which it would likely have to do anyway for the larger battery used in its Cybertruck), and that the V4 (or upgraded V3) Superchargers capable of this will be rolling out sooner than expected. 

GM’s announcement seems to confirm this course, as it has already committed to an 800V architecture for its EV lineup. Either way, Ford and GM vehicles will have access to Superchargers starting in the spring of 2024, so Tesla has a year or so to make the necessary changes to the network to support higher voltage. Given GM is likely already tooling up its 2025 NACS cars, I’m certain Tesla is quickly upgrading the Superchargers to support them. 

Driving Data - NYC Curbside Pilot Report /  by William Watts

Source: NYC.gov

As I mentioned in my last post, I was impressed with the NYC/FLO curbside L2 chargers near my apartment before and after a recent road trip. The chargers were installed as part of a pilot program that rolled out 100 chargers from June 2021 to July 2022 through a collaboration between the city and Con Ed, with FLO serving as the provider responsible for charger installation and maintenance. The admirable goal of the program is to accelerate EV adoption, particularly in low and moderate income neighborhoods. Earlier this month, DOT Commissioner Ydanis Rodriguez released the pilot’s first evaluation report. After reading it, I think the thorough approach the city has taken towards measuring success is at least as impressive as the chargers themselves.

The city has done a great job capturing data. That the DOT is releasing this report at all is a sign of its data-driven approach to EV charging and marks an encouraging shift from the norm. This is the most detailed analysis of L2 charging I have seen, and includes information that I have never seen from a charging provider, let alone a municipality. This is the kind of data that allows governments to understand whether their initiatives are actually driving EV adoption. I highly encourage anyone who is interested in EV charging to read the report in its entirety, but I’ll highlight some of my favorite stats here.

Source: NYC.gov

1. An example of unusual data: DOT actually measured the rate at which chargers are ICEd (blocked by an Internal Combustion Engine car), and concluded it’s about 20% of the time. While most EV owners and EVSE operators generally know that ICEing can be a problem, DOT took this on not only by issuing 3,200 tickets, but by installing time-lapse cameras (and presumably tasking someone with reviewing the footage) to get a sense of the scale. Even during my time at Tesla, where I had highly detailed charging data, we didn’t have accurate information on ICEing frequency or duration. Though the report shows that incidents of ICEing lasted, on average, less than half an hour, it still represents a huge impediment to increasing utilization of the chargers. Hopefully we can chalk this up to teething problems as people come to accept the idea of dedicated EV charging spaces, but it is certainly worth tracking. 

Source: NYC.gov

2. ICEing and other issues affecting charger availability may take on new importance given another interesting detail in the report: the relatively high utilization of the chargers. Here, utilization is defined as the overall percentage of time when a charger was plugged into a vehicle (excluding ICEing). On average, the chargers reached 34% utilization by December 2022. A third of the locations saw over 50% utilization and some reached as high as 69%, which is a particularly impressive feat considering the ICEing problem. This is an unqualified success, since 60%+ utilization is about as high as can reasonably expected of any public charger considering the relatively low turnover overnight.

3. The most surprising part of the report is its claim to 99.9% uptime. Although I have seen very few complaints on Plugshare about inoperable NYC/FLO chargers, and the status lights on the chargers in my neighborhood have never indicated anything other than a functional charger, 99.9% sounds almost too good to be true. The design of the FLO chargers that locks the plug to the post until a session is initiated may have some effect on the reliability and discourage vandalism. It’s also unclear whether the city is using the same uptime calculation as the NEVI guidelines, which excludes factors like electric utility service interruptions, internet or cellular service provider interruptions, and outages caused by vehicles. In any case, anything close to this reliability is encouraging for the feasibility of a wide scale deployment of curbside L2 in the city. That said, it has only been a year or so, and I will be interested to see how this number changes over time. 

4. Perhaps most unsurprisingly, the chargers located in neighborhoods with higher median household incomes (Manhattan, Williamsburg, Park Slope) tended to have the highest utilization. Less wealthy neighborhoods, which are less likely to own EVs, such as the Bronx, Eastern Brooklyn and Queens, tended to have lower usage. This is despite a higher density of EV chargers inside of paid garages in the wealthier neighborhoods. So we have confirmation that if the city installs chargers where EVs are popular already, owners will use them. Measuring the impact of the program on the stated goal, driving EV adoption in low and moderate income areas, will take time. Most Americans only purchase a car every six years, so patience is necessary when quantifying the effect of this infrastructure on EV adoption. Whether and how those trends change will be interesting to watch in the coming years, but the instrumentation is in place.

Measuring the impact of the program on the stated goal, driving EV adoption in low and moderate income areas, will take time. Most Americans purchase a car only every six years, so patience is necessary when quantifying the effect of this infrastructure on EV adoption. Whether and how those trends change will be interesting to watch in the coming years. 

For example, according to the report, the city has seen a steady increase in EV registrations of 2-3% per month. That is about a thousand new EVs per month in a city where fully half of all vehicles are street-parked. It’s clear that there was pent-up demand for curbside charging in some neighborhoods, and the availability of curbside might encourage EV adoption in others. That’s true for me, anyway–the presence of available L2 charging in and around my neighborhood is certainly going to inform my next vehicle purchase decision. One way to measure the program’s impact on adoption would be to not just track EV registration growth but also to compare rates of change in neighborhoods with curbside L2 versus those without. 

Either way, this report is encouraging both because of the preliminary success of the curbside program itself, and because of the rigorous, data-driven approach the city has taken to its rollout. Given the success of the pilot, I look forward to the expansion of the program and am eager to see how these figures change in the next edition of this report.

An EV Roadtrip in 2023 /  by William Watts

It did not go 100% to plan.

I used to spend a lot of time driving long distances in EVs. At Tesla, I traveled up and down the eastern seaboard to scout charging locations, negotiate with property owners, and represent Tesla at planning and zoning board meetings. In those days, between 2016 and 2019, there really weren’t third-party fast chargers to speak of. I relied exclusively on the Supercharger network to get me where I needed to go. 

I haven’t done a long trip in an EV since the pandemic, so when a friend asked me if I’d be interested in joining her on a drive from Florida to New York to bring her dog back home, I figured it would be a good opportunity to evaluate the EV charging state of the union. To that end, I decided to rent an EV and make the drive from NYC to Orlando and back in six days. 

Because the Tesla Supercharger network is a known quantity–the chargers are plentiful, reliable, and seamless to use–and my goal was to understand what the charging experience is like for the rest of the EVs on the road, I initially sought to rent the fastest charging, most reasonably priced, non-Tesla EV available. Right now that means one of the Hyundai/Kia offerings: the Ioniq 5, Ioniq 6, or EV6. Unfortunately those are not currently available from Avis, Hertz or Enterprise, and for an interstate road trip, Turo rentals are impractically expensive. Instead I rented a Tesla Model 3 Long Range from Hertz and resolved to utilize third-party chargers wherever possible. So I purchased a third-party CCS to Tesla Adapter on Amazon (ordering one from Tesla wasn’t an option because the company requires a Tesla VIN number). Just before leaving, I was also able to get hold of an official Tesla adapter from a friend. 

This turned out to be an enlightening trip in ways I couldn't have expected. All in all, I drove 2669 miles in six days. I compared the charging experience at eighteen DCFC chargers and four L2 chargers across three networks. I experienced the sinking feeling that comes with experiencing charging failure with my vehicle at a single-digit state of charge (SoC), and the latent anxiety of wondering how much my range would drop overnight in the cold. I realized the challenges of route planning outside the Tesla network, and in particular, the importance of topography to the accuracy of range estimates. I learned the very real difference between having a long range versus medium range EV. I learned how resilient an EV can be in the case of an accident, as well as some of the pitfalls of renting one in these early days where it’s still perceived as somewhat exotic. And I was impressed by how cost-effective EV driving can be, even when using the relatively expensive electricity offered by public charging stations. 


Charging Networks

Electrify America. The single biggest change to EV charging since 2019 is undoubtedly the rapid expansion of the Electrify America (“EA”) network. The first EA stations were just going live when I left Tesla, but now they are relatively ubiquitous, at least along the I-95 corridor that made up most of my southbound trip. These stations are spaced every 80 miles or so along major highways, and typically offer between four and ten relatively fast DC fast chargers, generally a mix of 150kW and 350kW posts. The overwhelming majority of my non-Supercharger charging was done at EA stations and I had a largely positive experience, with some minor snafus. 

Electrify America - Emporia, VA

EA: The Good. In the plus category, I found the EA stations relatively reliable and usually placed in sensible locations. The chargers more or less worked as advertised, and while I did always spot at least one other car during my stops, I never had to wait to charge. I signed up for the monthly subscription for a very reasonable billing rate of$.36/kWh or $.29/min. I also very much appreciated the real-time availability data accessible not only in EA’s own app, but on Plugshare as well. While the stations were never full, a few times the sole 350kW charger was in use, and it was nice to be able to know that ahead of time. EA’s biggest advantage was the density of its network. I was generally able to plan a stop around an upcoming station without compromising my charging strategy too much. In fact, often they were located on the same highway exit as a Tesla Supercharger, allowing me to trick my car into preconditioning the battery by inputting the Supercharger as the destination, and simply going to the EA station instead. 

EA: The Bad. But EA had a few issues. On the software side, though I appreciated the real-time availability data in the app and on Plugshare, that data wasn’t always accurate. Sometimes a charger would show up as “in-use” on my phone, when it clearly was not in use. This wouldn’t be too big a deal, if not for the fact that initiating the charge session in the app was the most reliable way to start charging, and that can’t be done if the app thinks the charger is “in use.” The other method, the “tap phone to initiate” function, had glitches. Some posts simply refused to acknowledge the phone, others beeped but failed to start the session. 

Electrify America - Arundel Mills, MD

There were also general hardware reliability issues. Some were minor, like a charger that made slightly unnerving rattling noises when plugged in, or CCS connectors that were missing superficial covers or had suffered other obvious external damage. Other issues were more  serious: at about a third of the EA stations, at least one post or pair of posts was out of service altogether. Luckily, because there were multiple posts at every site this was only a minor inconvenience, but given all EA stations are relatively new, it was discouraging to see such a high failure rate. It also made me wonder how fast EA does repairs. As I write this, two weeks after I experienced a completely down charger, it remains listed as “under repair” on Plugshare, with second pair of chargers, a third of those available at the site, now broken as well. The NEVI guidelines require a 97% uptime for each port, which means 11 total days of downtime per year, so I’d imagine this may be a problem for EA soon if they want to take advantage of this federal funding. 

EVgo. The only other third-party DCFC stations I tried were EVgo, but I quickly learned to avoid these. That wasn’t hard; EVgo stations are pretty much exclusively located in and around cities anyway. This meant that along I-95, there were zero EVgo stations in Delaware, North Carolina, South Carolina, or Georgia. EVgo hasn’t yet truly built a charging network; what it offers are standalone charging stations that may be useful to those who happen to live near them and provide very little utility to those traveling through. I did go out of my way to try one EVgo station that was listed as 200kW on Plugshare, and found myself sorely disappointed. This station boasted a ‘100kW’ Tesla connector, two ‘100kW’ CHAdeMO connectors, as well as two ‘200kW’ CCS connectors between two posts. Unfortunately, despite that 200kW rating, the system must be limited to 175A, because whether I was using the Tesla connector or CCS with an adapter, my charge was capped at around 70kW. As I’ve written previously, this discrepancy between the advertised charging rate and the actual rate is a function of the misleading way hardware can be rated: note that these chargers are actually fully compliant with the 150kW NEVI standard, notwithstanding how slow I found them in practice. 

EVgo - Richmond, VA

Tesla. The Tesla network proved to be as reliable as I remembered, and far faster. When I left Tesla in 2019, the 250kW V3 superchargers were just being deployed, and the 2170 cells in the Model 3 and Y (which charge significantly faster) were just rolling out as well. This combination, along with the mile per kW efficiency of the Model 3, makes for a pretty incredible charging experience. On V3 superchargers I routinely saw 250kW at low states of charge, which meant getting from 15% to 35%, adding 72 miles of range, in 5 minutes. Ultimately the only minor disappointment was arriving at a Supercharger to find it was an older V2 station, capped at 150kW (and charging at full speed also required not parking next to a neighbor). 

Tesla Supercharger - Summerton, SC

Other DCFC Networks. I was unable to use any other DCFC networks during the trip. Even with a relatively low power filter of 70kW, it became clear that most of the other providers have almost no corridor charging presence, even along I-95. ChargePoint, which ostensibly has the second-most DCFC locations nationwide after Tesla, had a single station between NYC and Orlando. Blink had none. There are a number of non-networked DCFCs dotting the I-95 corridor, often located at Nissan, VW, Ford and Chevy dealerships, but they typically feature only one charger, and trying my luck at  one of these at a low SoC seemed too much like tempting fate.

L2 Networks. I charged at L2 chargers only four times on this trip, and most of those experiences were positive. The first and last sessions were at the curbside Flo chargers under the Brooklyn Queens Expressway near my apartment. These are part of a city-sponsored program and have been reasonably successful in bringing L2 charging to the curb. Other than some peculiarities with the Flo app, which failed to display the real-time charging speed, everything worked seamlessly, though it was a bit expensive at an effective $.52/kWh (Flo actually bills per minute). I also charged overnight at a friend’s home and once for free at a hotel. I tried and failed to charge at a public parking garage in Charleston, but there the problem wasn’t the charger but the fact that after a minor road accident (more on this below), Hertz had given me a second Model 3 that was missing its Tesla J1772 adapter. I returned to my hotel with 10% SoC, grabbed my spare adapter, and attempted to use the L2 charger at an adjacent car dealership–only to have an automated security system loudly warn me that I was trespassing. I hoped the car wouldn’t lose too much energy overnight, and arrived the next morning at the nearest Supercharger with 4% SoC, or 10 miles of range to spare. This felt more like the EV ownership experience I remembered from 2016…

Tru by Hilton - Asheville, NC

Trip Planning

Tesla. The other somewhat tricky part of the road trip was trip planning. With a Tesla, this is pretty straightforward if you plan on using the Supercharger network. You simply input your destination, and the car plans your charging stops for you (only at Superchargers), telling you when and where to charge, and how long you need to stay there. I do think the trip planner trends conservative, in that it doesn’t seem to take into account how much faster cars charge at a low SoC, but that makes some sense, given the high downside risk of owners not making it to their destination. 

Third-Party Trip Planners. There are third-party trip planners like evnavigation and abetterrouteplanner that do an admirable job of replicating Tesla’s planner solution using data about the locations and availability of public charging networks. But they are not fully integrated into the car, which presents two difficulties. The first is that you find yourself fooling around with your phone while driving to use them, which is not very safe. The second is that if you find that you are using a bit more energy than expected (because, say, you’re driving over the speed limit), you will need to adjust your plans mid-trip. For Tesla vehicles, at least, abetterrouteplanner has a solution to the route-adjustment problem: it allows you to pass real-time vehicle data to the app through your Tesla account (a feature I could not use while renting from Hertz). Hyundai also added charging stop planning to their navigation UI a little over a week ago, but I have not yet been able to use it.

Most of the time, trip planning isn’t crucial–you can simply look at the car display for how many miles of range you have left and find a charging station within that range. Occasionally, however, your route may take you through some high elevation change, which can have a pretty dramatic effect on range. I first noticed this the day I drove from Charlotte to Asheville. During a stop in Spartanburg SC, I noticed the Tesla trip planner was indicating that I needed to charge to nearly 100%, for roughly 250 miles of range, though my destination was only 150 miles away. Having been to Asheville before, I knew this was likely because the road passes through the Blue Ridge Mountains. Sure enough, the car estimate proved to be reasonably accurate and this leg of the trip used around 30% more energy per mile than the trip average, even though I traveled at slower speeds. This also happened to be a stretch where the only DCFC chargers available were at EVgo stations that were capped at ~70kW for 400V cars. It’s easy to see how a driver not using a trip planner could get in a tough spot–at best, experiencing a much longer trip than expected, thanks to slow charging, and at worst, getting unexpectedly stranded. While I didn’t experience it on this April/May trip, cold weather can have a similar effect on range, so it is wise to leave a significant buffer when traveling in the winter months. 

Medium vs Long Range EVs. After I took a hit from a flyaway truck bed cap (more on that  in the next section), I was forced to trade in  my Long Range for a Standard Range Model 3. On the bright side, this gave me a chance to directly compare the two on trip planning and charging. The Long Range Model 3 has an 82kWh battery and a rated range of 358 miles, while the Standard Range has a 50kWh battery and a rated range of 272 miles. For 99% of the driving most people do, the Standard Range is more than sufficient. Long road trips represent a tiny fraction of journeys; more than half of American car trips are less than six  miles and only 5% run over 30 miles. It takes an extreme situation, like driving 2700 miles in six days, to really notice the difference in range.

Perhaps counterintuitively, the biggest difference I felt between the Standard and Long Range cars isn’t ultimate range, but charging speed. As I’ve written elsewere, EVs don’t charge at a constant rate. The battery is generally the limiting factor on speed, and lithium batteries charge faster at lower SoC. Charging speed is also proportional to the size of the battery, with larger batteries able to take on more energy in a given time than smaller ones. This is why I saw as high as 252kW in the Long Range Model 3, but never more than 170kW in the Standard car. 

Most people tend to stop every 3 hours (200 miles) or so on long journeys, especially when there are children or animals along for the ride. Without accounting for charging speed, one might think that both vehicles would perform similarly under these constraints. In practice, the Long Range car arriving at 10% SoC would need to charge to 66%, a task that takes as little as 15 minutes, while the Standard Range would need to charge to 84% to cover the same 200 miles between stops, which takes twice as long due to the battery charging slower at higher SoC. Again, an extra 15 minutes for distances that represent less than 1% of trips taken is not a big deal for most users, but it does add up when covering nearly 1000 miles in a day. What might be more meaningful to owners is the effect of a medium range car on charging options along secondary roads like the aforementioned Charlotte to Asheville route, where there are 100+ mile stretches with zero DCFC options.



The EV Accident Experience

I count myself lucky in that I have never been in a car crash. That’s still true after this trip, but I did experience an incident that highlighted some of the advantages and inconveniences of EVs. 

While driving through South Carolina, I was horrified to see the folding truck bed cap of the GMC Sierra in front of me fly into the air and tumble, spinning like a leaf, towards me. Knowing there was a car in the lane to my right and that I could not entirely avoid the errant panel, I slammed on the brakes to give myself the chance to run over it rather than have it come through the windshield. Ultimately the car didn’t quite slow enough, and the cap managed to wedge itself in the narrow grille of my Model 3.

FOD Foreign Object Damage

Luckily, the driver of the truck realized what happened and pulled over. I called the police so they could file a report, then called Hertz for next steps. Hertz informed me that the nearest facility where I could bring the car and exchange it for another EV was 200 miles south, in Jacksonville. I know from experience how hard it is to get a single tow truck to travel that far, and so did the Hertz operator. She asked me whether the car was driveable. I confirmed that it was for the moment, but that the radiator was leaking, and there was a good chance it would not make the full distance. 

Cracked radiator

I decided to get as far as I could. The HVAC system blew hot, the sound system had been disabled (because of damage to an external speaker/horn wire), and the undertray had been partially torn from the front of the car, but I was more or less good to go. After the operator reassured me that Hertz was okay with me making the attempt, I used an extra shoelace to tie the undertray to the bumper to keep it from dragging on the ground and started on my journey. 

Roadside repair

The Good: Driving without a Cooling System. After about five minutes, the car registered that it was low on coolant and capped my speed to 65 miles per hour, then 60, but my top speed remained stable for a half an hour. As the outside temperature increased, the allowable speed decreased to 55, then 54. The car display confirmed that the car was “Safe to drive,” and it appeared to be regulating its power output to ensure the motors and battery did not overheat. I made it 150 miles before I needed to stop to charge, by which point my speed had dropped to 50 mph. 

Limping the car home

Because charging puts as much (or more) heat into the battery as driving, I suspected that topping off to reach my destination might result in the car locking up. So I waited until the last possible moment to charge–when I was under 10% SoC–to shorten the likely tow to Jacksonville as much as possible. That prediction proved accurate. Though the car charged normally, I made it only about 200 feet from the charging station before the car flashed the “PULL OVER SAFELY” message and promptly shut down. Eventually it did allow itself to be moved again, but it was unclear how far I’d make it. Four hours later, I was towed the last 50 miles to Jacksonville, where I was able to quickly exchange the damaged Model 3 for a new one (but unfortunately only Standard Range cars were available). 

While this might seem like a failure, the experience was significantly better than what would have happened in any ICE vehicle. The fact that I was able to travel 150 miles after damage to my EV’s radiator speaks to a distinct advantage of an EV over an ICE vehicle. Any ICE vehicle’s cooling system requires high pressure and is absolutely critical to the vehicle’s basic function. Driving an ICE vehicle even a few miles with such a badly damaged radiator would have resulted in overheating and even permanent engine failure if ignored. Meanwhile, some EVs (like early Nissan Leafs) have no cooling system at all, and the ones that do operate at relatively low temperatures and pressures. Though these systems serve a function–they increase the amount of power that can be put into or removed from the EV motor and batteries–without them, the cars can still operate at reduced load. 

The Bad: Beware the “Exotic” EV Insurance Policy Carve-out. Renting an EV does come with a potential downside, which did not become clear to me until later. Because my Chase Sapphire credit card includes rental damage coverage, I declined Hertz insurance on my Tesla rental. The Chase coverage is known to be good and reduced the rental cost by about a quarter. But it was a close call: only after the accident, when I googled “Chase Sapphire rental insurance,” to learn the process for filing a claim did I see the alarming fine print. Chase Sapphire Preferred excludes “high value or exotic” vehicles from the rental coverage policy: these are defined as  “Alfa Romeo, Aston Martin, Bentley, Corvette, Ferrari, Jaguar, Lamborghini, Lotus, Maserati, Maybach, McLaren, Porsche, Rolls Royce, and Tesla.” To be clear, one of these things is not like the others. The Tesla Model 3 and Model Y, which represent an overwhelming majority of Tesla sales, are hardly “high value” as far as cars go; they are just above the average price of a new car in the US. It was stunning to me that they were excluded, especially given that Hertz had just bought 100,000 of them. Fortunately, after stressing for the remainder of my road trip at the prospect of being on the hook for the repair costs of the Model 3, I realized that my card, the Chase Sapphire Reserve, offers a flat $75,000 of coverage, with no specific car brand exclusions. I got lucky, but the Reddit forums on this subject make clear that others have been burned by this seemingly arbitrary policy. 

 

Cost 

Cost was not a factor on my last significant EV road trip because I was in a company Model S with free supercharging for life. This time around I was interested to see how the trip would work out financially, on everything from the rental experience to charging at the public fast charging networks. 

Rental Cost. All in (excluding insurance coverage), renting a Long Range Model 3  for a week with unlimited miles cost $341. That made it the cheapest possible rental by a pretty significant margin. An ‘economy’ rental of a Chevy Spark or similar ICE vehicle for the same period would have cost me $564. Surprised by the difference, I happily booked the Tesla. (Hertz’s favorable rental rates for Teslas might have something to do with the lower maintenance costs.)

Driving Cost. As for the cost of driving, in total I spent $180 (including two free charges) to drive 2669 miles. Even factoring in prevailing rates for those two free charges, the total would have been under $200. On the other hand, if I had driven my personal 2017 Subaru Outback, which gets a respectable 27 mpg on the highway, the trip would have cost $352 in fuel alone (at the $3.57 national average regular fuel price). It is somewhat hard to believe that renting a brand new car and driving it 2700 miles came out to only $200 more than just the fuel cost of driving my own vehicle. 


Conclusions

On long journeys, EVs still require a bit more ‘savvy’ than ICE cars. On this trip, I occasionally applied the ‘Mom Test’: Would I reasonably expect my mom, who is more tech savvy than most, to do this? For a road trip in a Tesla, the answer is yes. The process of navigation and trip planning is sufficiently simple and automated that I’m confident she could handle it. Would I send her out in a Hyundai armed with Plugshare and abetterrouteplanner? Absolutely not. There is still significant work to be done on network density and integrating the user experience if we want road trips to be as easy in an EV as in an ICE vehicle.

But all in all, this was a fantastic trip. Before this, I already had 50,000 EV miles under my belt, but it’s clear to me that a lot has changed in the four years since I last regularly relied on public charging networks. As I mentioned, my personal car is currently an ICE vehicle, largely because I street-park in Brooklyn and I use my car mostly for my 250-mile trips to Vermont. I thought that having an EV(especially a non-Tesla) without dedicated charging here in NYC would be too big a hassle, adding too much time to an already long weekly drive. As it turns out, that view is a little dated for a few reasons. 

  1. The non-Tesla charging networks have gotten much better since 2019. EA, while not without its problems, never let me down on my trip and provided cost effective, truly fast charging every time. 

  2. Curbside L2 charging is really convenient, and the experience reinforced for me how critical they will be for widespread urban adoption of EVs. The stations I used near my apartment worked well, if a little pricey. I just hope the city doesn’t stop building them, as they were sometimes completely full. 

  3. Modern EVs charge fast enough. The Model 3 Long Range isn’t even the fastest charging car on the market, but the ability to acquire 200 miles of range in about 15 minutes is more than sufficient for long trips. I’m sure cars will continue to get faster, but for the first time this felt more or less the same as ICE pitstops, which is a significant milestone. 


This trip convinced me that it’s time to trade in my personal vehicle, so expect a future article on the used EV buying experience.