I was looking at the specs for the Kia EV6 after someone brought it up in a discussion:

DC Fast Charge Time (10-80% @ 350 kW via Electric Vehicle Supply Equipment) Level 3 Charger: Approx. 18 min.

If you're not familiar with EVs or other similar equipment you might think that this draws a constant 350kW, but charging a 77.4 kWh battery from 10-80% at 350kW would take only 9min so it can't be that. Instead, EVs are smart: they communicate with the charger to draw varying amounts of current depending on how quickly the battery can accept charge.

So then you might think that 350kW reflects the peak current the car draws. But no: when P3 Group measured it they found it peaks at 235kW, before throttling back when the battery gets to 50%.

This isn't unique to Kia: I started going through P3's report looking up advertised rates, and drawing much less than advertised was the norm. Even starting from a low state of charge and preconditioning the battery for fast charging:

Make Model Advertised Measured Ratio
Kia EV 6 350 kW 235 kW 0.67x
Hundai Kona 100 kW ~75 kW 0.75x
VW ID.3 [EDIT: 120] kW 103 kW 0.86x
Hundai IONIQ 5 >250 kW ~215 kW 0.86x

The two exceptions I found were very small cars that didn't make large claims: the Fiat 500e (85kW) and Mini Cooper SE (50kW) did test as advertised. [EDIT: Tesla is also fine; I originally misread the report and thought it peaked at 146 kW, but it does hit it's 250 kW claim.]

On the other hand, peak current doesn't actually matter very much: these cars have a range of charging curves and what actually matters is how quickly a charger extends your range. P3 also ran that test, looking at how many miles you can go with 20min of charging from 10%, and on this metric the EV6 came in first, hitting 192mi. So while I don't want to be too hard on Kia—they do seem to have made a car that can charge very fast—listing 350kW when they only peaked at 235kW is weird.

Comment via: facebook, mastodon

New Comment
3 comments, sorted by Click to highlight new comments since:

On the Kia EV6 page you link first, I think the position of the 350 kW value you quoted being part of the initial conditions rather than an expected draw is pretty clear. The interpretation I'm pointing at is “if connected to a charger with a capacity of 350 kW, the expected time is approximately 18 minutes”—the 350 kW is on the LHS of the conditional as signaled by its position in the text. By comparison to nearby text, the entry immediately above the one you quoted states 73 minutes under the condition of being connected to a Level 3 charger (the same fast-DC type) with a lower capacity tier of 50 kW, and other entries above that one display corresponding “if provided with” → “duration will be” pairs for weaker and easier-to-install charging pathways, down to household AC taking the longest duration. This would make no sense under an interpretation of each of the wattage values all reflecting the battery's internal acceptance rate. Note that the text as you quoted it is not visible to me as a straight-running block of plain text in the page linked; instead, “DC Fast Charge Time (10-80% @ 350 kW via Electric Vehicle Supply Equipment) Level 3 Charger” is the opening line of an expandable box whose body content is “Approx. 18 min.”, and that interaction/presentation provides more clarity to the conditional structure.

The Tesla Model 3 page states “max” for its 250 kW figure, whereas the P3 page is clear that the car “only achieves an average charging power of 146 kW” (emphasis mine), and the associated graph does show 250 kW being drawn at the initial charge state of 10%, then decreasing as the battery charge increases.

The Hyundai Kona and IONIQ pages are similar to the Kia EV6 page: the kilowatts are in the if-part of the conditional, and the measurements listed in the body cells to be relied on as outputs are the minutes.

The VW ID.3 brochure I also read as having the kilowatts in the if-part, though the text is laid out less clearly. Also, I only see your 125 kW figure mentioned in the context of the 77 kWh battery option whereas the P3 report specifies that they used one of the 58 kWh battery options.

In general, what information flow will the median consumer use? Not “do a bunch of division that's going to be wrong because of other variability in how batteries work”. “Show me how long it takes with this type of charger” is the information that's closest to their planning needs. The Tesla page is unusual here compared to the others, but “will a higher-capacity charging feed than X reduce my charging time” is a plausible second most relevant question and is answered better by the max than by the average (if you provide less than a 250 kW capacity, some part of the charge curve will take extended time by comparison to the nominal one).

if connected to a charger with a capacity of 350 kW

I don't know, do you think they're trying to say that if connected to one of the 250kW chargers they wouldn't be able to hit their maximum charging speed? Because I expect they would, and if it's just "what can you expect when using a charger at least this powerful" they'd want to list the lower number.

The Tesla Model 3 page ... only achieves an average charging power of 146 kW ...

Thanks for catching this! I misread the report and have updated the post to match.

I only see your 125 kW figure mentioned in the context of the 77 kWh battery option whereas the P3 report specifies that they used one of the 58 kWh battery options.

Thanks for catching this as well! The same PDF gives 120 kW for the 58 kWh battery model. I've updated the post.

I don't know, do you think they're trying to say that if connected to one of the 250kW chargers they wouldn't be able to hit their maximum charging speed?[1]

Let's rephrase the question from the publisher side: when compiling a specification sheet, which sets of conditions do you choose to include? This is naturally a balance, where comprehensiveness and comprehensibility trade off, along with other factors like difficulty of testing. For a more extreme example, consider a world in which Level 3 chargers regularly showed up in capacities advertised to 1 kW granularity[2]: does the manufacturer take every data point between 1 kW and 350 kW and put in a massive table and graph (and hope that J. Random Carbuyer knows how to read it)?[3]

The Kia EV6 spec seems to base its independent-axis points on anchors of public perception of Level 3 chargers at the time of publication. 350 kW is the top capacity of Level 3 charger commonly available to my knowledge (this shows up in various factoids on the Web for instance), and 50 kW I think may be the lowest or among the lowest. The consumer is, yes, left to guess what the curve in between looks like, but those bounds are good enough to establish an intuition and make rough plans from. These also correspond loosely to two useful anchor edges of the situation envelope: “living in a proven electric-vehicle-friendly area where 350 kW chargers are likely to be built if they're not here already / wanting to know what kind of charge time I can expect if I make sure to go to the good station” and “living in an area where only weaker chargers are getting any traction thus far / wanting to know what kind of charge time I might have to budget for if I just choose any old place”.[4]

A specific motivational factor that I would imagine here is implied range of observation. This comes into play both on the factual level, with facts like “they've tested that this car works with a 350 kW charger at all”[5] and “they know that this charging time stays true for a 350 kW charger”, and on the social status and emotional comfort levels that derive from those, along the lines of “this manufacturer is up to date with the progress of charging technology and is likely to know what they're doing with the modern stuff”. The latter, while to some degree self-serving to the manufacturer, is not misleading provided that the design and/or test conditions did in fact exist.

Comparability is another motivational factor that I would anticipate here—that the consumer would want to see “how do cars X, Y, and Z behave under whichever charging conditions I'm currently holding in mind”[6]. However, there wasn't much evidence of this in the set of specs under discussion. Possibly this is because they come from different manufacturers, where I would expect comparability to be a stronger force for multiple specs from the same source.

So: there's reasons the 350 kW condition could be advantageous as a point to provide, and I see no pointed, strong reason that they would want to put a 250 kW row in as well. But that doesn't mean they can't, either! Alternate approaches do occur in your list of examples. Let's look at two of them:

  • The VW ID.3 brochure is more technical-looking overall than the Kia spec sheet, and provides its DC3 (which I believe is the same as “Level 3 fast DC charging”) charge time based on slightly different charger capacities for the different battery options. I don't know why exactly, and I think this makes it require more mental steps to compare them, but assuming the reader will infer an inequality is a valid alternative and seems closer to what you're expecting.

  • The Hyundai IONIQ page specifies the inequality explicitly: “>250kW”. Quibbles over the use of > vs ≥ aside, this again seems closer to what you're imagining above as far as picking a tier beyond which further benefit cuts out as the reference point to provide.

From a more speculative plausibility-modeling perspective, I see that you calculate (and I also calculate) an 86% ratio between the P3 peak measurement and the charger advertised capacity condition for both of those cars. Let's suppose that the charging capacity conditions in the spec sheet were in fact based on an anticipated benefit ceiling per above. I didn't see in the P3 report where exactly they were measuring the power usage, but that ratio could easily be explained by transmission losses between two different measuring points, safety headroom, or similar natural discrepancies. If we assume that Hyundai is taking into account a 50 kW granularity of advertisement of charger capacity, then for their Kona model, the ceiling to that granularity of 75 kW / 0.86 also matches the 100 kW figure in the consumer spec sheet—harmlessly leaving extra room in the same way as with the 350 kW anchor in the Kia spec, since maybe a 90 kW charger would be just as fast.

But that wouldn't explain the three cars where (after revision) the figures match better, so let's look at those one by one:

  • The Fiat 500e… well, I actually can't seem to get at the spec page anymore after some kind of misgesture that causes it to redirect to the Fiat USA homepage, which doesn't list a 500e. I don't have time to figure out how to untangle right now, so that one's inconclusive for now.

  • The Mini Cooper SE brochure presents the figure you cite in a different place from the Kia and Hyundai spec sheets. Taking the yellow horizontal bars as the if-then separator, the preconditions are just Level 1/2/3, and the “up to 50 kW” text is below the “Level 3” bar, in the corresponding result cell. Note specifically that this doesn't say “if you connect it to a charger that advertises a maximum of 50 kW, the peak power drawn will be 50 kW”! If that turned out to be false, you could even make a reasonable case that this statement were more misleading from a consumer perspective (with some wishy-washiness around whether the consumer should be assumed to be aware of natural discrepancies per above). I didn't see a part of the P3 report that described attempting those conditions, so I don't know how the Mini Cooper would physically perform there.

  • The Tesla Model 3 page does something similar and is subject to a further vertical integration constraint: the heading of that cell doesn't say “Level 3” but rather says “Supercharging Max”, where I believe Supercharging refers specifically to Tesla's charging methods and facilities. Presenting a front in which the consumer has less discrepant information to reconcile from different suppliers is one of the benefits of vertical integration, and it seems reasonable that Tesla would take advantage of that.

From a broader communications-analysis perspective, I think the intended audience of the P3 report is different. If I look at the front page of the P3 website, it does not seem designed to provide a Consumer-Reports-like publication to me, and their “About” page implies a B2B/intra-industry context with “We advise our clients strategically in the areas of technology strategy, business process optimization and organizational development.”. So aside from the promise vs observation and independent vs dependent axis confusions, there's a big difference in communication context there.


  1. I actually think it is plausible that it would not! I don't know the details of how e.g. voltage negotiation, reference points of power measurement, or safety handling around transients work for high-capacity electric vehicle charging, but in general I would expect the effective tier requirement for the charger to be higher than the power delivered to the battery, with the uncertainty mostly revolving around how much higher. But I don't think this is necessary for my argument that their spec sheet is constructed reasonably; I think my primary position holds even if we assume that the same maximum charging speed would be reliably attained by a Kia EV6 connected to a 250 kW charger. ↩︎

  2. A rough glance around the Web suggests that something like 50 kW granularity of ‘tiers’ of Level 3 charger is common. ↩︎

  3. Longer datasheets with more specialized and technical audiences do do this sort of thing, though! Electrical components with current versus voltage curves, or power consumption versus clock speed, for instance. ↩︎

  4. Level 2 and Level 1 charging times are sufficiently far off, and their environmental prerequisites sufficiently far off as well, that they correspond to distinct behavior types and distinct rough planning regimes (at work / at home overnight rather than gas-station-style top-ups). This is presumably the motivation behind the level categorization in the first place. ↩︎

  5. You could consider this redundant with other marketing messages surrounding interchangeability of charging stations, but for anecdote, I personally observed a relative several years ago fretting over whether a higher-capacity USB flash storage device might require more experience or finesse to operate properly than a lower-capacity one, and I would be unsurprised if analogous forces came into play for a large chunk of the target market of household electric vehicles. (The route is even more plausible in this case, if I'm allowed to speculate; what if, say, a 350 kW charger puts out a within-spec but sharper voltage rise time than the previously tested 250 kW one, and this pushes the envelope of part of the charging circuit in an unanticipated way?) ↩︎

  6. From a marketing-ease standpoint, the conditions the consumer anchors on could well be based on whichever sheet came up first, and the reading-mechanical strictness could even get to the point of “scan the sheet for a specific number and give up or feel frustrated if it's not there”, though neither of these are necessary. ↩︎