Tesla EV charged with diesel generator still cleaner than conventional car

Print Friendly, PDF & Email

Over the weekend, we performed a test between a very fuel efficient diesel and a very energy hungry electric vehicle, charged from a diesel generator, to see which car had the best energy efficiency using diesel. Here’s what was found…

Electric cars can be charged from many sources of power generation, one of those being the reliable but no longer cost-effective diesel, good for remote and emergency situations, a favourite of regional roadhouses.

Up until now, the perceived problem with charging an electric vehicle from diesel generated electricity was the belief that it consumed more fuel than just driving a regular internal combustion engine on the same journey.

Of course, the EV critic always manages to compare a heavy Tesla to a lightweight 4 cylinder diesel car, the same critics also always appear to overlook the fact that most Australian EV drivers offset their energy use with solar power. Why wouldn’t they? It makes economic sense.

The best way to put this claim to the test is compare a very fuel efficient turbo diesel 5 seat sedan against one of the most energy hungry EVs around, charged from a 30Kva diesel generator over the same roads at the same time.

Our test turbo diesel was a Volvo D4, weighing 1500kg with a claimed fuel consumption of 4.9 litres per 100km. This car has less weight and power than a Tesla, but that’s the whole point – play by the EV critics rules.

The Electric vehicle tested was a Tesla Model S P85D: a dual motor, high powered, 2200kg barge, apparently a very energy hungry vehicle. This particular car is normally charged from solar power, but has the versatility to charge from any power source.

Step one in the experiment was to charge the Tesla direct from the gen-set using a 3 phase power lead, making sure the tank was completely filled before and after charging, and the fuel replaced measured as accurately as possible. Enough energy was added to the car’s battery pack to cover the equivalent of at least 100km.

Step two was to top up the Volvo with fuel. This was a very slow and difficult process due to the anti-siphoning trap door. It was critical at this point to get the fuel level perfect.

Step 3 was for both cars to depart as soon as possible, driving the same route and traffic conditions for a minimum of 100km. The plan was for a mixture of highway and suburban driving with air-conditioners in use. Once the 100km mark was passed, both cars returned to the starting point.

Step 4 was to refill the Volvo to the exact same fill level as before departure. This was easier said than done, and took much patience and steady hands. Once this was completed and the amount of replacement fuel recorded, it was time for some calculations and cross-checking.

The result:

The Volvo is an exceptional car that recorded consumption of 4.800 litres over the 104.6km trip, that’s slightly better than the claimed fuel consumption.

The Tesla is another exceptional car; it’s also a 2200kg barge with dual electric motors that can propel it down the quarter mile in 11.6 seconds. Regardless of all that, the Tesla electric car required only 4.460 litres of diesel to cover the same 104.6kms.

The video:

  • handbaskets’r’us

    Rob’s hard work and advocacy for electric vehicles in WA -and Australia generally is pure gold.
    From overseeing a rollout of charging locations statewide, liaising with government and industry, and supporting the local EV owners, we owe him a debt of gratitude.
    Thanks so much Rob, for all you’ve done for the change we want to see.

  • Annie Atkins

    Great article Rob, very interesting to see that the electric vehicle is more efficient even when using an inefficient generator to charge it and despite using a heavier higher powered car.

    • heinbloed

      Well, it isn’t really.
      The overall picture needs to be taken into account.

      • handbaskets’r’us

        The article cited states: ‘Electric cars are much cleaner than internal combustion engine cars over their lifetime.’
        ‘Battery manufacturing life-cycle emissions debt is quickly paid off.’

        This study does not take into account the FF energy required to produce, refine, transport and deliver FF in the first place.

        • Miles Harding

          In addition to FF energy used in refineries, they use an enormous amount of electricity.
          see here:


          • JonathanMaddox

            It’s a shame that this video is still shared so often. Its numbers are completely wrong.

            It’s a perfectly valid criticism of the petroleum economy that refining crude oil into useable liquid fuel takes a lot of energy.
            But Robert Llewellyn has looked up the *energy* consumption (loss) of petroleum refineries and simply assumed it is all in the form of electricity. It isn’t.

            Refineries’ energy inputs are typically 80-90% in the form of crude petroleum, 7-10% natural gas, and less than 1% electricity. Their fuel product output is typically 84-90% of the total energy input. So yes, their energy losses are large, up to 15% of the input energy.

            Refineries do use a lot of electricity, but their throughput in terms of fossil fuel energy is more than a hundred times greater. Most of their energy loss (consumption) comes from the fossil fuel feedstocks themselves. Electricity use is less than a tenth of that.

            As for the one refinery that uses more electricity than two of the UK’s smaller cities: sure, it does, this is less than 1% of the UK’s electricity consumption. But that refinery also processes enough liquid fuel to power half the UK’s vehicles. The energy in that fuel is equivalent to some eighty times as much energy as the refinery’s electricity consumption. If that fuel were used only to generate electricity there would be significant thermal losses; it could still deliver fifty times as much electricity as the refinery’s electricity use.

            The real figure for a refinery’s electricity consumption per UK gallon of liquid fuel product is between 200 and 350 watt hours, not 4.5 kWh. (Equivalently, 50-66 Wh per litre).

  • Antony Day

    Well done, Rob and Jon, I’m sure Craig Kelly will be admitting how wrong he was sometime soon…

    • handbaskets’r’us

      Unfortunately, facts don’t change what people believe, and people with hobby horses never get off, -or there’s nothing to ride.

    • Joe

      Just put the Kelly in the pilots seat so he can feel the ‘cleanliness’.

  • Chris Jones

    It’s been pointed out elsewhere, but I’m rather impressed by the diesel generator’s efficiency – 4.46 litres of diesel to generate 18 kWh of electricity stored in a battery. The chemical energy contained in the diesel is about 44 kWh, so an efficiency of ~40% is impressive. Still – why would you when the sun is free?

    • Electric Boogaloo

      Generator engines run at a constant speed, which allows the engine designers to optimise economy for that speed.

      Large marine diesels like you can find in container ships and cruise liners have even higher efficiencies.

      • Mark

        I believe the current world record is around 50% thermal efficiency for a piston diesel engine. Allowing for the generator losses that is not far above 40%.

  • N Page

    Now do a Zoe!

    • Chris Jones

      Even better I would think, although I heard the charger (basically the drive inverter running backwards) is less efficient than a dedicated charger.

      • heinbloed

        The shorter the range capacity(the size of the battery) the cleaner the e-vehicle becomes.
        The Telsas are the filthiest.

        • Farmer Dave

          Thank you for the references: very interesting. I did not see in them the alarmed tone suggested by your comments. Indeed, the reports reinforce the message that even with the current European grid emissions intensities, substituting internal combustion engined vehicles with battery electric vehicles will reduce emissions over the life of the vehicle. They also point out that as grids are decarbonised, the benefits provided by electric vehicles will improve.

        • reecho

          Heavier vehicles use more energy to push them along. What a novel concept…

    • reecho

      Now if you could actually get your hands on one….

  • Phil Gorman

    That’s pretty impressive but in the interests of balance there are other factors to consider. The embodied energy that’s gone into producing each vehicle is an important consideration when comparing their credentials. That should then be offset against projected lifetime energy consumption. It’s already been done but there are mind bending variables and a-priori assumptions involved.

    The Union of Concerned Scientists found 51% to 53% of battery electric vehicles (BEVs).…/gasoline-vs-electric-global-warming-emissions-953
    See also:

  • Steve Woots

    Good to see the results of a fair test. However, I think we are at the stage where EVs don’t need any help. The public are waking up and the neanderthal trolls – like the other neanderthals – are dying out.

    • Greg Hudson

      There are a few remaining neanderthals commenting on this site. They are not extinct after all IMO 🙂

  • heinbloed

    A different picture with a broader view:

    referring to

    Being 20 or 50% better than a combustion engine won’t do.
    The longer the range of the battery-vehicle is the worse it is for the climate.
    Tesla leads the electric polluter list.

    • trackdaze


    • Greg Hudson

      Yet another FF troll ?

      • Farmer Dave

        Very possibly, Greg. Certainly, giving links to articles while misrepresenting their contents is a standard climate denier trick.

    • handbaskets’r’us

      You keep flogging this horse, but it just doesn’t run.
      Selling short?

    • reecho

      If you are going to troll, at least post some links with a few more untruths to try and firm up your case….

    • Nick Kemp

      So that paper suggests that IF batteries are used using electricity from coal fired power and IF EVs only last slightly longer than 150,000km and are charged from a grid using some coal fired power they are still better than ICE vehicles.

      In the meantime and the real world Lithium mines are being constructed alongside their own solar PV farms so that they can mine and process the lithium on site.

    • Nick Kemp

      “Electric cars are much cleaner than internal combustion engine cars over their lifetime. We find that a typical electric car today produces just half of the greenhouse gas emissions of an average European passenger car. Furthermore, an electric car using average European electricity is almost 30% cleaner over its life cycle compared to even the most efficient internal combustion engine vehicle on the market today.

      Plug-in hybrid vehicles, when driven on electric power for most trips, have lifecycle emissions similar to battery electric vehicles. In markets with very low-carbon electricity, such as Norway or France, electric vehicles produce less than a third of the life-cycle emissions of an average combustion-engine vehicle. This finding bolsters governments’ goals to promote electric cars as part of their decarbonization strategies”

      So the article you quote actually supports EVs – and that’s even while using some coal fired electricity to build and run them. So what is your point?

  • This test looks flawed to me. How could you know that “Enough energy was added to the car’s battery pack to cover the equivalent of at least 100km”? Could it be that the battery had less energy after the trip than prior to charging so that not all of the trip energy came from the generator.

    • Ferris B

      The Tesla records the energy consumed since last charge and/or last trip in tenths of a kwh, there is also cross checks that can be done.
      So you dont think the test was flawed due to the Volvo being 700kg lighter with 1/3 of the power and half the torque?

      • arne-nl

        “So you dont think the test was flawed due to the Volvo being 700kg lighter with 1/3 of the power and half the torque?”

        Sorry, that’s a cheap shot to a very valid question. This difference was already mentioned in the article.

        They should have measured the Tesla’s consumption by employing the same methodology as with the Volvo. By recharging it back to the same battery level as before the trip and then recording how much diesel the genset needed to do that.

        Now the test depends on some vague, unspecified assumption: “Enough energy was added to the car’s battery pack to cover the equivalent of at least 100km.”

        Any solid conclusions are therefore invalid.

        • Ferris B

          The vehicle was charged to a set amount of 18kwh, the vehicle completed the journey consuming no more than the 18kwh, the vehicle was charged beforehand whilst performing all pre-trip vehicle checks, tyre pressures, air con settings and all the other factors that need to be addressed so no car has an advantage.

          • How do you know the generator produced 18kwh of energy? If you used the car instrumentation it doesn’t record power in but state of charge and it takes more than 18 kwh of energy to increase the battery energy by 18kwh. Also, with a partial charge the increase in energy stored in the battery is only a rough estimate.

            As arne-nl said, the Tesla needs to start the test fully charged then after the run take it back to full charge and measure the diesel consumed. If you also want to know the energy that went into the battery measure it with a meter external to the car.

            Round tripping energy through lithium batteries is about 82% efficient see ( so for the claim to be true other losses in the diesel car have to be greater than 18% of the energy consumed by the heavier Tesla. That may be the case, but this experiment doesn’t show that because much of the trip energy could have been supplied by the vehicle battery rather than the generator.

          • reecho

            The Tesla does record power in as well as power consumed. Again you are overthinking the test…

          • I was advised by a colleague that has a Tesla that they don’t measure energy in though it wouldn’t surprise me if they do.

            If the measured energy in matches or exceeds energy out then the test is not flawed in the way suggested but the test does deserve close scrutiny as the results seem unlikely.

          • Ferris B

            Ken the point is how much diesel is consumed to move the vehicle the same distance as the other vehicle, generators work at different efficiencies for many different reasons, charging losses vary with cable length and quality, there are onboard charging losses but all these calculations are not needed as its a simple test.

          • The test is only valid if all of the energy to run the Tesla came from the generator. Maybe it did but “Enough energy was added to the car’s battery pack to cover the equivalent of at least 100km” isn’t very reassuring.

            It is a surprising result as the extra losses in the diesel car have to be greater than the losses of round tripping through the battery. According to the Which? tests reported at the Volvo S60 D4 is very fuel efficient, beating all hybrids except the Outlander PHEV (Plug-in Hybrid Electric Vehicle) and the Outlander doesn’t round trip through the battery whenever possible to avoid these losses. Also, as was pointed out, the Tesla is much heavier. Because the result seems unlikely there needs to be a high level of confidence in the testing.

          • Ferris B

            The test has to based on the energy added to the battery, using the higher generator figure would flaw the test in favour of the EV.
            The genset sits at close to peak efficiency for the charging cycle, there are variable losses all the way to the battery pack, once the car is driven it makes very good use of that energy, there is very little wasted. The ICE vehicle has a very good Diesel engine, the downside is the multiple speed gearbox and engine management system is fighting a continual battle to keep the car in the efficient power band, car companies spend tens of millions every year trying to improve these systems to use less fuel and yet it will never match the flexibility and efficiency of a electric drivetrain, Volvo have finally woken up and are going full electric and hybrid.

          • arne-nl

            They started off with a battery charge level of 53%. But did they end the journey with that charge level? More? Less? There is just not enough information.

          • Ferris B

            A smudge over 53%, importantly the energy consumed was slightly less than the energy recorded going in to the battery, the energy used is broken down into 1/10s of a kwh.

        • reecho

          Would make no difference to the overall result.

  • Peter

    It is not clear from the description, but was the Tesla also charged after the test run? Or how was it determined to have been charged to the same level before and after the test run?

    • Ferris B

      Getting a 100km plus drive was the first requirement, the second part was to not use more energy than added to the battery pack, the car could have also been charged a second time and the average of the two used but more than enough exhaust emissions were created for one day.

      • Peter

        If that is the case, I think the experiment is flawed because the energy replaced in the Volvo by diesel fuel corresponds to the energy it used on the test run. But in the Tesla, the energy used on the run has not been replaced and so measured. The correct way to run the experiment is to fully charge the Tesla (from the grid, if you wish) before the test and measure the fuel used to charge the Tesla after the run.

        The experiment also ignores well-to-wheel loss for the diesel, which would favour the EV.

        It also depends on where the electricity comes from. My EV (a hybrid, so includes petrol) alone would halve emissions in Tasmania, but in Victoria it increases them because of the Brown Coal. Installing Solar PV halves my emissions.

        • reecho

          You are over thinking the experiment. What is so hard to understand with this test?

          • Ferris B

            I agree with Reecho, too much analysis goes into generator efficiency percentage, charging losses and a lot of other factors that vary greatly, just way too much margin of error, real world testing provides a far clearer picture.

          • Peter

            So, allowing for “margin of error”, within 10% is a draw.

          • Ferris B

            Where to you see a 10% margin of error?

          • The problem is that that the diesel fuel consumption for the Tesla was calculated prior to testing. The energy consumed on the test course made no difference to the result. Look at the sequence “Step one in the experiment was to charge the Tesla direct from the gen-set using a 3 phase power lead, making sure the tank was completely filled before and after charging, and the fuel replaced measured as accurately as possible. Enough energy was added to the car’s battery pack to cover the equivalent of at least 100km.”

            The fuel was measured “as accurately as possible” but the amount of energy added to the battery was what the experimenters estimated would be required. So it doesn’t matter how accurately the fuel was measured, it is the accuracy of the estimate that is important and that estimate was made prior to testing. If 10% less energy was consumed driving the course, or 10% more, the reported result would be the same. Had the Tesla not driven the course at all, the result would be the same. So the experimenters estimated the energy required to drive the course and used that estimate as if it was the actual energy used to drive the course. That is not a test but an estimate pretending to be a test.

          • Ferris B

            The energy was only estimated to be sure as close to but preferably more than 100kms was achieved, there was no estimating in the final calculations after the 104.6km drive, X amount of fuel was consumed by the generator to travel X amount of distance, the result would have been the same regardless of charging before or after. If the Tesla had been charged after the drive only you would have found fault with that. I’m surprised you have not found fault with a list of advantages towards the ICE diesel.

          • The problem is that “X amount of fuel was consumed by the generator to travel X amount of distance” is probably not true. The number of kilometers that could be driven on the amount of fuel consumed by the generator was estimated prior to the Tesla running over the course. The energy actually used by the Tesla had no bearing on the reported result.

            Surely you can accept that this is a flaw in the test procedure?

          • neroden

            You don’t understand how the test worked. Read it again. You are flat out wrong. They did it right. You are incorrect.

          • Peter

            They are your words.

          • Ferris B

            When did Peter say that there was a 10% margin of error?

          • Peter

            Actually, I do not say that. But, perhaps you did not see the comma and read, instead, “margin of error within 10%”.

            The 10% refers to the difference in fuel quantities. Volvo: 4.800L; Tesla: 4.460L. Difference = 340mL, which is within 10% of the fuel usage of either car. Hence my comment.

            The energy content of Diesel fuel is about 10 kWh/L, so 340mL has an energy content of about 3.4 kWh, but, after running through a heat engine, the energy into the generator would be about 1 kWh.

          • Peter

            It is a matter of doing the test in a way that it is not open to criticism. It is not a matter of understanding.

        • Mark

          The well to wheel losses are equal – the losses per litre of diesel input for either car are equal. That is the point of measuring how much diesel both cars used travelling that distance.

          Filling/charging to 100% before and after the run, topping up to a known level and discharging back to a known level, or starting empty and adding a measured amount of diesel and seeing how far it goes are all ways to measure the diesel used for a distance. I don’t see how the accuracy would significantly vary. The method they used was like topping up one tank from another, then running a hose from that tank until it was back to the original level to see how much watering got done. You can still measure the total amount of water used from the original tank (and hope for zero losses in this scenario). I don’t see your criticisms of this test as valid (and nothing is ever completely closed to criticism). Given the small sample size however, I wouldn’t call it conclusive in all cases, but it certainly does demonstrate the point that EVs are likely not to have higher emissions regardless of electricity source.

          One reason not to fill to 100% is that the battery balancing circuit may cause unusual fluctuations in the charging pattern that would not cause any real losses on the grid, but likely would on a generator.

          • Peter

            My criticism was that the battery of the Tesla was not recharged after the test, whereas the Volvo was. So, the Volvo fuel added after the run corresponds to that run. But, the battery was charged before the run. What quantity of fuel was used to charge the Tesla battery after the run? That is the figure that should be used.

            I have no concerns that such a test was done. I think it is great demonstration.

            But, doing experiments often involves work to figure out how the test should be done. For example, the vehicle emissions tests allowed a significant car company to game the measurement process. As a result, such tests need to be improved to prevent such behaviour.

            The experiment is changed to improve it and make it more accurate, realistic or precise. It is worth changing this experiment in the light of such comments.

          • Mark

            I just don’t see how it changes the experiment. It’s going to take the same amount of charging to go from 50% -80% before or after the drive. If it uses 80%-50% to do the drive, you can assume 100% if the diesel was used. Given a linear relationship between charge level and energy input at this state of charge, if the car only used from 80%-53% you can calculate what proportion of the diesel was consumed by that driving route.

            Would it be easier to record the amount of charge used, then measure diesel use on recharging? Probably. Does it make a difference to the ability to calculate the diesel usage for that drive? Can’t see how.

            The reason the emissions testing needed changing was because the results were measuring different performance characteristics to those intended. I believe this test measured what it was intended to.

          • Peter

            Charging form 50 to 80% and then discharging from 80 to 50% relies on two end points. Starting from 80%, doing the run and recharging to 80% relies on one end point. The fuel used corresponds to the run and avoids unknown effects.

            Re the emissions test: Sure it measured what was intended, but that was not related to normal use and allowed the car company to game the system.

          • Mark

            Not sure what unknown effects you might be positing here. Would you care to elaborate?

            It may be that due to the vagaries of battery management systems charging to 80% the second time gives a different reading. In that case I’d suggest a larger sample set across more runs would be more beneficial than changing the measurement method.

            I suppose there are unknown effects from air bubbles in the Volvo’s fuel tanks if we really want to get into unknowns.

          • Peter

            Ahh… They would be “unknown unknowns” to use Donald Rumsfelt’s terminology”. :-).

            One should construct experiments to avoid any “known unknowns” – things we know we don’t know. Hence, make the Volvo and Tesla runs as similar as possible, so as to see differences.

          • Mark

            Yes, I believe they drove the same route for this reason.

            What I believe you are essentially putting forward is that if you know you consumed 5 litres of diesel (due to an inherent metering system) that you’d still need to measure that again by topping up a tank to know you’d consumed 5L. Honestly if there are any unknown unknowns all bets are off. For example, if you can’t trust the energy metering of the Tesla, you also couldn’t know if you’d charged it back to the same level.

            The reason you only need to charge the Tesla off the generator once is so that you have a measurement of how much diesel gives you how much charge. That was the other known unknown in this test. I’ve no reason to suspect it varies significantly between charges, but perhaps someone should provide a larger sample set.

          • The flaw is that the diesel consumed was measured prior to the Tesla travelling the course. The actual energy consumed by the Tesla had no bearing on the reported result. If the Tesla had traveled 10km more or 10km less it would not effect the amount of diesel used. The reported result could only be correct if the experimenters were able to know how much energy the Tesla would consume prior to it travelling the course. Because they could only estimate this they are reporting an estimate rather than a test result.

          • Mark

            Of course the diesel in the Tesla was consumed prior – it was put in the battery. That energy was then drawn out of the battery. The conversion rate from litres of diesel to kWh of charge was measured. In the final step the amount of charge added by the diesel can be multiplied by the amount consumed to run the course and gives a calculation of the diesel used by the Tesla to run the course. Its not an estimate, it’s a calculation. If the Tesla ran an extra 10km, or 10km less, that would have reflected in the amount of charge used, and thus changed the result of the calculation. I’m not sure they ever said they used the total amount of diesel put into generator as the figure for the Tesla’s consumption. The article notes the last step involved “calculations and cross-checking”.
            It’s pretty easy maths – although I’d prefer they provided the numbers. 18kwh added with say 5L of diesel is 3.6 kWh/L. If it takes 17.5kwh to run the course, then it took 17.5/3.6=4.861L of diesel.

          • Yes, that would make sense but nowhere was it said “The conversion rate from litres of diesel to kWh of charge was measured”.

            Assuming it was done that way then there is some inaccuracy due to the state of charge of a lithium pack being only accurately knowable at full charge but that is a far milder criticism. My colleague with the Tesla told me they are advised to occasionally take the vehicle to full charge to keep the battery meter accurate.

          • Mark

            I’d unreservedly agree with that criticism if this was a journal article intended to be a rigorous description of the experiment, rather than a summary of a quick test. As I said, I’d prefer their numbers to be described, and maybe they did document all that just not in the youtube video or this article. For an informal test, and filling in the blanks of the methodology, I think it’s qualitatively and quantitatively reasonably accurate.

            Yes, the vagaries of battery balancing are going to cause some inaccuracy – but I don’t think it quite works like you describe – you don’t more accurately know the amount of energy stored at full charge necessarily. However state of charge can be defined by voltage, rather than stored energy, and you can measure voltage at any charge level accurately enough.

            The state of charge in sense of total stored energy can only ever be estimated from a number of measurements – a truly accurate measurement has to affect the state of charge. The difference between a part and full charge in a Tesla (and most EVs that I’m aware of), is that the voltages of individual cells are only “balanced” to a specific voltage as the charge tapers off – cells that are already full are bypassed to stop their voltage increasing until the other cells are all at the same voltage. This allows you to have known voltage point on all cells, and doesn’t happen if you don’t fully charge. (Some packs are instead designed to balance the cells at their discharged voltage such that when a cell reaches it’s nominal minimum they all do, but when fully charged the voltages differ, rather than differing more at the discharged end). All of this is unnecessary if the cells are identical, but that doesn’t happen in practice.

            However you still don’t know how much energy is in the battery at either point. This has to be estimated from running the battery down and seeing how quickly the voltage drops, accounting for temperature, current demand and the charge/voltage curve of the battery and measuring how much energy was extracted in the meantime. Current battery management is a lot better at estimated charge levels by monitoring the voltage across each parallel cell module (which is then in series to form the whole battery), whereas it was previously estimated on the total pack voltage.

            Overall it’s not 100% accurate, but I’d say modern battery systems get pretty close, especially as they factor in the measurements from the previous discharge to fine tune the weighting factors of all the variables. I’m not sure if the level of error on the battery energy management would be within the same order of magnitude as the error margin on measuring out the diesel, but I’d peg it at close enough to see that an EV charged from a generator can get better fuel economy.

            I did once read that the RAV4 EV, which could be leased with an optional range extender trailer, actually got better mileage running from the range extender than the petrol RAV4. So this result with the Tesla isn’t completely surprising. I can’t find any mileage figures for the range extender RAV4 EV combo now, but I found a bit of a description of it.

          • neroden

            You’re just plain wrong, Ken.

            The only possible criticism would be if you believe that Tesla’s (highly accurate) battery meter was recording the charge level differently before and after the trip. It started at one charge level (before the genset was turned on), and it finished at that charge level (after the trip). As long as those reported charge levels both represent the same actual charge level, *this test is completely 100% accurate*.

            If they didn’t represent the same charge level (which they do), there would be no way whatsoever to do any test, so this is the best you can possibly do.

          • The neroden interpretation that the methodology was “It started at one charge level (before the genset was turned on), and it finished at that charge level (after the trip)” is one possible interpretation of the test description. However it clashes with extra information provided by Ferris B, who seems to have been associated with the testing. Ferris B says “the energy consumed since last charge and/or last trip in tenths of a kwh” was measured and that wouldn’t be relevant to the neroden methodology. Ferris B also says “the energy consumed was slightly less than the energy recorded going in to the battery” and this contradicts the neroden methodology as well. There are also problems with the neroden methodology in that it is quite hard to know the state of charge of a lithium battery, other than at full charge, with much accuracy. See the publication State-of-charge and capacity estimation of lithium-ion battery
            using a new open-circuit voltage versus state-of-charge available from to get some insight into the problem.

            Mark interpreted the methodology description as “The conversion rate from litres of diesel to kWh of charge was measured” which is different to the neroden interpretation and mine and to which I responded “that would make sense”.

            My initial interpretation of the methodology seems to be, after some discussion here, likely incorrect though it is consistent with the test description, like all the others. When I initially tried to clarify by asking “How could you know that “Enough energy was added to the car’s battery pack to cover the equivalent of at least 100km” the response that “The Tesla records the energy consumed since last charge and/or last trip in tenths of a kwh” didn’t answer the question so I stuck with my interpretation at that stage.

            Now I’m mostly confused but given that the neroden methodology clashes with the insider Ferris B statements, I suspect neroden is confused too without having realised it yet.

          • neroden

            No. Both of them rely on one end point.

            Yes, the Volvo is starting at the top of the “gauge” and ending at the top of the “gauge”. The Tesla is starting at the bottom of the “gauge” (53%) and ending at the bottom of the “gauge” (53%). This makes no difference whatsoever. Either way you are restored to the starting level.

  • JFK

    With regard to life cycle emissions the Model S is not the cleanest EV but it comes in very close, and most certainly not “filthy.” The Model 3 is comparable to the Leaf and other similarly sized EV’s. The Prius and other hybrids do quite well.

  • neroden

    THANK YOU. This is going to be my new go-to article to debunk that old canard.

    This test was set up to disadvantage the EV in every way possible — use the heaviest, least efficient EV, the lightest, most efficient diesel car, and then power the EV from a small diesel *generator* which isn’t that efficient either. And there you are — the EV uses less energy.

    • Maybe the Tesla consumed less energy but it probably didn’t. The Tesla energy consumption was an estimate made prior to testing so the claim relies on the accuracy of this estimate rather than a test result. The reason it probably didn’t is that besides being heavier, the Tesla has to round trip all energy through the battery costing something like 18% and that is a big inefficiency to overcome by regenerative braking and operating the diesel generator at peak efficiency.

      Maybe the Tesla consumed less energy but an estimate of energy consumed by the Tesla is poor evidence to support such a bold claim.

      • Ferris B

        I can see how you view this, being uncomfortable with a test that negates your long held belief is understandable, let me comfort you with the fact that in a few short years this will all be forgotten, renewable energy will dominate the power grid and also the charging of the Australian electric car fleet.

        • I expect that “renewable energy will dominate the power grid and also the charging of the Australian electric car fleet.” I’m also thrilled when my colleague with a Tesla allows me to drive it. Unfortunately I can’t afford one for myself but I can afford an electric bicycle, which is my favourite form of transport.

          But how is that relevant to whether the test has a flawed methodology? In this comparison an ICE provided the energy for both vehicles so it is actually a test of whether round tripping energy through a battery is more efficient than direct drive. Maybe it is, but that is unlikely, so the claim should be supported by measuring the fuel used by the Tesla to complete the course rather than estimating it.

          • neroden

            This test has a 100% accurate methodology. Why are you throwing lying smears at it? Cut it out or I’ll have to assume you’re a fossil fuel troll.

      • neroden

        OK, so basically, what you wrote is just complete tosh. Nonsense. Go read the methodology again. You are wrong, and that’s all there is to say about it. You are flat out wrong about the methodology.

        There’s no two ways about it: The Tesla started and ended its trip with the same charge level. The Volvo started and ended its trip with the same fuel level. The Tesla used less diesel to make the same trip.

        So everything you wrote is purest lies. Smears. Dishonesty.

        If you’re an honest person, if this is an honest mistake, withdraw your dishonest comment and apologize.

  • Barry Alternative Fact Covfefe

    Thats incredible!

  • Initially, this appeared to be a beautiful test of which is more efficient, a diesel generator or a Diesel engine propelling a vehicle, and whether the difference in efficiency between the two is sufficient to offset the inefficiencies in generating, storing, and then using electrical energy to propel a vehicle of similar capabilities.

    However, one crucial piece of data was not reported, namely the state of the battery after the drive was completed, that is, how much kWh would be needed to recharge it to its initial charge state. Using that and the fact that 4.46L of diesel provided 18kWh of energy to the battery, we could say how much diesel fuel would be needed to recharge the battery, if it was too inconvenient to actually do the recharging and measure the fuel use.

    I can accept the result that the fixed diesel generator was more efficient than the moving Diesel engine, and that the difference was more or less similar to the energy losses in the Tesla, but not much more can be said, in my opinion.