Forums

TL;DR: Conclusion:
Charging a Model 3 at 240V is very efficient, and noticeably more efficient than charging at 120V. You will pay about 12% more to charge at 120V compared with charging at 240V.

I installed a nice little power monitoring device in my Circuit Breaker Box (an IoTaWatt from www.iotawatt.com) that lets me monitor the power used by each individual circuit, and decided to authoritatively answer the question of how efficient the Model 3 LR is at charging on 120V vs. 240V. I've seen all kinds of numbers thrown around, using all kinds of proxies for actual power (Miles gained, battery % gained, etc), but now that I had the ability to measure actual power I figured I could be more accurate.

See below for more data. The forum software won't let me post this all together.

Frank99 | 2019年10月19日

Setup:
Temperature: 75-85F (24-29C) in my garage
Equipment used:
- IoTaWatt in the Circuit breaker box, with dedicated monitors per circuit
- Stats running on my phone, querying the car's API
- Gen 2 Mobile Connector charging at 32A @ 240V, or 12A @ 120V.

Technique:
- Charging started in the evening, stopped in the morning. IoTaWatt and Stats were queried to find how much power was drawn from the house panel, and how much was stored in the battery.
- The 240V outlet and 120V outlet are very short runs from the Circuit Breaker box, minimizing power loss in the wiring. The 240V circuit is about 15' of AWG 6, and the 120V circuit is about about 5' of AWG 12.

Results:
At 240V (Measured average about 252V), Stats showed 40.62 kwh added to the battery. IoTaWatt showed 42.5 kwh pulled from the wall. Efficiency = 40.62/42.5=95.6%.
At 120V (Measured average about 126V), Stats showed 10.15 kwh added to the battery. IoTaWatt showed 11.9 kwh pulled from the wall. Efficiency = 10.15/11.9=85.2%

Conclusion
Charging at 240V is very efficient, and noticeably more efficient than charging at 120V. You will pay about 12% more to charge at 120V compared with charging at 240V.

Example: Driving 10,000 miles per year, at 250 Wh/mile, will take about 2500 kwh from the battery. At 240V, you'll pull (2500/0.956) = 2600 kwh from the wall over the year to recharge the battery. At 120V, you'll pull (2500/.852)=2900 kwh from the wall over the year, which is 12% more power and likely 12% more money.

Limitations:
This test was done in optimal conditions - warm enough to not have to run the battery heater, and in the battery's optimal charging range, while at the same time cool enough to not need to run the A/C to cool the battery (I think...). The numbers are likely to be different if you're charging outside at night in Chicago in winter, or if you're charging outside in the afternoon in Abu Dhabi in the summer. I do not have the ability to measure those efficiencies (well, I could do the equivalent of Abu Dhabi here in Phoenix...).

So, there you have it.

dave | 2019年10月19日

Thank you Frank, useful information, concise data, great work!

Bighorn | 2019年10月19日

Awesome work!
12% sounds kind of meh compared to increased frictional losses of 236%:) And the cost penalty could be compounded above 12% if you cross a per kWh pricing threshold based on kWh volume.

TexasBob | 2019年10月19日

Excellent data. Thanks.

calvin940 | 2019年10月19日

My solar net metering should be installed next week and then my 33PVs go online. I bought the Sense Monitor with the additional solar leads and have it up and running . I am obsessed with watching .y house power usage and figuring out what devices are turning on when and how much power they use. It has machine learning to use the electrical signature to figure out what they are. I am psyched to start seeing the real breakdown of charging usage as well as all the other stuff on my power.

Tesla2018 | 2019年10月19日

Great writeup. Do you know or could you run a test and see if charging at 240v at 36 Amps is more or less efficent than charging at 18 amps in order to icrease range by the same amount?

I only drive about 30 miles. a day so once evey five days I charge from 100 miles up to 250. miles and it takes about 5 hours using 36 Amps to add 250 miles of range. If I set it to 16 it would take longer but I am sleeping and not using the car anyway, so I was wondering if it makes any difference in efficiency.

gballant4570 | 2019年10月19日

Tesla2018, no data to back this up..... but I would think that longer charging time might increase total inefficiencies or losses, assuming that they are fairly constant during charging no matter the charging level. It would be interesting to see that tested though.

Tesla2018 | 2019年10月19日

I heard there were some additional losses due to increased heat from charging at higher Amperages.

Sort of like seeing if a 50 watt bulb uses half as much power as a 100 watt bulb. In that case it uses half that watts But in the case of the car we are drawing amps and not watts and not powering anything except for some type of internal converter that determines if the car iis charging using household or supercharger power.

Bighorn | 2019年10月19日

@gallant
Quite possibly, but if they are fixed losses, they’re going to happen whether you’re charging or not. Those losses are disproportionately penalizing the slower technique as those phantom losses are also going to affect the faster charged car, but aren’t being quantified.

Bighorn | 2019年10月19日

@gballant

Frank99 | 2019年10月19日

Tesla2018 -
There will be more wire losses at higher amperage, but I expect those to be small. Conversely, there'll be more "overhead" to run the computer longer at a lower amperage charge, and I expect that to be a bigger effect. So, I think that if I dropped the 240V charge current down to, say, 16A, I'm pretty convinced that the efficiency would go down.

bjrosen | 2019年10月19日

Frank99@ The 240 vs 120 wasn't exactly apples to apples, you added 40.6KWh at 240 but only 10.15KWh at 120, that might have some effect on the efficiency because you covered a much wider range of charge levels when doing 240. Would you please do a few more measurements,

240 @32 for 10KWh starting at the same charge level as you did with 120.
240@16 or 10KWh starting at the same charge level as you did with 120.
I'm assuming that you haven't wired your wall connector for 48 or you would have done your test at 48A instead of 32A, but if you do have 48A capability them do the same test at 48A.

Thanks,

bjrosen | 2019年10月19日

I just want to add a thank you for doing these measurements, I've been assuming a 90% efficiency for Level 2 charging, it's nice to see that it's 95%. It only amounts to about 1.25 cent/KWh for me or .3 cents per mile but it's the principle of the thing.

Frank99 | 2019年10月19日

Nah, not going to do any more measurements. In the first place, I have no way to experimentally charge precisely the same number of kwh at precisely the same battery SOC without babysitting the process, and in the second place because in both experiments, the power pulled from the wall is rock-solid overnight- there was no tapering in the charge. In addition, in both experiments the battery was near 200 miles range at the end of charging. Now, there may be a difference in charging efficiency at different SOCs, but I expect that to be much smaller than the other variables in this test.

And I only have the Mobile Connector, so 32A is the max I can charge at. If you send me your Wall Connector, I'm willing to do one more run....

Earl and Nagin ... | 2019年10月19日

@Frank99,
Great work! Thanks for investing and doing the work to contribute to our global knowledge.

gballant4570 | 2019年10月19日

My thoughts above were mainly considering level 2 charging - probably should have stated that. But if for example 2A is consumed as a loss during level 2 charging, the more hours spent charging the more 2A/hrs are consumed. Therefore, charging at a higher level 2 rate would be more efficient. Like I said, no data - just logic.

mrburke | 2019年10月19日

@Frank99 - Any chance you measured the power use of just have the Mobile connector plugged in, but not attached to the car ?

BTW - Excellent work

Frank99 | 2019年10月19日

As a matter of fact, I did. Looks like about 3 watts plugged in, regardless of whether it's connected to the car or not.

mrburke | 2019年10月19日

@Frank 99- Very thorough work. Thanks

bheck11 | 2019年10月19日

@Frank99 - thanks for such a thorough job!

Sure, there are some variables not fully controlled, but this is not meant for publication in a journal. Knowing that the difference is around 12%, give or take a couple, is excellent info.

@Frank99,

Did you try 12A/240v set up? Not sure efficiency is depending on current or voltage, or both.

Frank99 | 2019年10月19日

That test I might try - although it'll likely be 240V/6A to match the power of a 120V/12A charge (if the car will let me set the charge rate that low). Give me a day or two until the battery gets a bit lower...

My guess is that the efficiency will be nearly the same at 240V/6A as 120V/12A.

kevin_rf | 2019年10月19日

It is worth pointing out someone over in Tesla Motors clubs did a similar test with different currents on the Model S and the losses where not constant. It was more a curve with peak efficiency for the Model S being closer to 30ish amps and trailing off a few percent on each side as current was cranked up and down.

While your results are inline with what others have reported on the Model S, it wouldn't hurt to get a few more data points below 32 amps. Say 24a, 16a, and 12a. Those correspond to 30a, 20a, and 15a breakers in the breaker box.

Frank99 | 2019年10月19日

That's a really interesting read, thanks!
I don't know if I have enough interest to do a comprehensive test like that, but maybe I'll do a few points.

There is no practical use case for 6A/240v, or even 12A/240v. In other words, nobody would charge the car like that.

Some interesting testing points would be 30A/240v which I use to lower the cable heat by ~10% and yet achieve reasonable charging speed, or 16A/240v, which is half of the max charge speed.

Bighorn | 2019年10月19日

I was getting 16A/240 off a welding circuit before I spied a spare 10-30 in my garage.

kevin_rf | 2019年10月20日

TeslaDvr, a NEMA 6-15 plug will provide 12a 240v. For people that do not have the ability to ad a 30a (NEMA 14-30, 6-30) or 50a breaker (NEMA 14-50, 6-50) the NEMA 6-15 and NEMA 6-20 (16a charging) are an option.

This might be due to limits in the breaker box or converting a 120v plug to 240v.

GrumpyinAZ | 2019年10月20日

I use the Tesla Wall Charger. It's initial draw is 44A and I would be curious to see how long that lasts before it drops. Just need the data recorder you're using....

Frank99 | 2019年10月23日

So I took some more data. Adding it to what I had before:
V/Amps Efficiency
120/12 85.4%
240/06 90.5%
240/12 93.6%
240/24 93.9%
240/32 95.6%

It's interesting (to me, but I'm weird) that there's a significant difference between 120V/12A and 240V/6A. 240V/12A is also suspiciously high - if it were instead 91.6% efficient, then there'd be a dead straight line if you graphed 240V charge current against charging efficiency for this data. One of these days, if I'm bored, maybe I'll re-run that test to see if I made a mistake; otherwise, I'll just assume that there's some step function in the charger efficiency based on current draw.

My conclusion: 240V charging is more efficient than 120V charging, but almost any rational charging current at 240V gives a very efficient charge.

Bighorn | 2019年10月23日

Isn’t the higher efficiency of high voltage charging the point of Porsche going to 800V? Plus a thinner wire requirement.

andy.connor.e | 2019年10月23日

How is the lower current less efficient? I would expect the opposite.

FISHEV | 2019年10月23日

"Stats showed 40.62 kwh added to the battery"

Where does this number come from? The self report of the car via StatsApp? We don't know the actual capacity of the battery and there have been some software changes lately that have had a significant effect on the Tesla's self reported numbers. 6% capacity "loss' per reliable testing.

The percentages seem about right overall.

https://batteryuniversity.com/learn/article/charging_lithium_ion_batteries

Bighorn | 2019年10月23日

Says the guy who doesn't know what a kWh is. Avoid troll derailment.

majassow | 2019年10月23日

@andy: During charging, the car is awake, consuming the same constant power which does not go into the battery, but will be noted at the meter. This fixed cost is independent of the charge current, so with lower amp charging, this is a bigger component. Note that the overhead for all these scenarios is in the 140-350W range, which is about equivalent to the roughly 1 mile per hour vampire loss when the car is left awake.

majassow | 2019年10月23日

My guess: The variance noted between the scenarios has to do with temp of the battery when charging. Lower temp batteries may require a bit of conditioning for optimal charging...

Frank99 | 2019年10月23日

Bighorn -
Ignore what their marketers say is the reason for 800V; it's not efficiency (I'm showing 95% efficiency in charging; how much higher can you get?). It is all about the smaller wire, both for the charging cable as well as the cables in the car. Tesla originally went with 400V batteries because they could buy 600V rated IGBT's to put in the charger and motor inverter; 1000V rated parts were difficult to source and expensive. They stayed at 400V for the Model 3 because of the legacy of 400V chargers and design knowledge from the Roadster and Model S.
Since their original design decision, 1000V rated parts have become available and affordable; that's why Porsche and others have been able to go to 800V batteries. It really wouldn't surprise me if the next generation of Superchargers had an 800V capability that'll get revealed when Tesla eventually moves to 800V also (perhaps with the Roadster? Maybe the Semi?).

andy -
IMHO, there are two factors at work here:
1. Resistive losses. Losses in the cabling and electronics due to current flow are a function of a square law - run twice as much current, you have four times the power loss. Conversely, halving current reduces power losses to one-quarter. This should improve efficiency at lower current.
My belief is that resistive losses are actually minimal in my setup. The circuits in the wall are very, very short with well-sized wiring, so there will be low losses there. The wiring in the car is designed for 300A Supercharging, so I wouldn't expect any losses there. The UMC itself has some power loss (it gets warm when charging at 32A), but 24 feet of wiring is pretty short, so there shouldn't be a lot of power loss there.
2. Overhead losses. I believe that there's a constant amount of power being drawn by the charging circuitry, regardless of how much current is flowing. There's a computer system responsible for charging that takes power, there's a measurement system that checks the voltage on all 96 groups of cells within the battery, there's a circuit that's doing the 240VAC->400VDC conversion, etc. This overhead will reduce efficiency at lower currents, because as the charging current gets lower the overhead becomes a greater percentage of the total amount of power drawn from the wall. Said a different way, if you charge at 1kw, you have to charge for 75 hours and pay for the overhead for 75 hours. If you charge at 10 kw, you only have to charge for 7.5 hours and only pay for the overhead for 7.5 hours.

If I were to suggest that this "overhead" was, for example, 90W, then with some super-secret calculations, I can see that the 240V efficiencies all become about the same at 96%. 90W is a very realistic overhead power level for the circuits I discussed, so I'm thinking that this is some pretty strong evidence for this opinion.

Frank99 | 2019年10月23日

Fishy -
The Stats app queries the car to get the energy stored in the battery from the last charging session in kwh, not in battery percentage. As a result, the calculations I'm doing have nothing, absolutely nothing, to do with the actual capacity of the battery or any changes to such reporting that have been done. As a result, you're once again wrong.

andy.connor.e | 2019年10月23日

@majassow

That makes sense if the fixed energy usage is that high during charging. I didnt realize there was so much, i'll probably have a 50A outlet put in when that time comes.

FISHEV | 2019年10月23日

"Says the guy who doesn't know what a kWh is."@BigHorn

Be happy to explain kWh to you but the question was how to measure the the kWh capacity of the batter in order to come up with efficiency numbers. We depend totally on Tesla to tell us that and we have had some drastic changes in the capacity (6% change per Bjorn's reliable tests).

The percentages look about right for line and equipment losses and the 6% change applied to the efficiency numbers is not a big difference.

But Tesla has never stated exactly what the Model 3 battery capacity is and it does add a level of variability to the efficiency numbers.

andy.connor.e | 2019年10月23日

All you're doing is manufacturing doubt.

Frank99 | 2019年10月23日

Fishy -
Any changes to the reported capacity of the battery as a result of a firmware update aren't applicable here - all my measurements have been done in the last week with no intervening updates. So, still wrong.

majassow | 2019年10月23日

@andy: 250W isn't so much when you think about how much power a typical desktop computer consumes. But with a fixed overhead of ~200ish watts, the faster you charge, the less time the car is awake: and the more efficient your charging.

Frank99 | 2019年10月23日

majassow -
Temperature has been very stable here in Phoenix over the last week - high 80's to mid 90's, so my garage has been very stable around 80. Charging generally started about 1 AM, so the car had been parked for hours giving the battery a chance to equilibrate to garage temperature, so there should be no changes due to battery/environmental temp. Good guess, though.

Interesting observation about vampire losses - 1 mph would be about a 250w continuous power draw. I can't justify that with the numbers I'm seeing (see my long post that I was writing as you posted yours), but I came up with 90w. It's not clear to me that the numbers should correlate - after all, the computer(s) responsible for charging, and the charging circuitry, aren't likely in use while the car is sitting in the garage - but it's interesting that they're in the same ballpark. For example, Stats tell me that my car's average vampire drain is about 0.58 mph - which equates to about 145w, suspiciously close to my calculated 90w.

andy.connor.e | 2019年10月23日

You probably get back that much if you upsized your wires.

FISHEV | 2019年10月23日

"Any changes to the reported capacity of the battery as a result of a firmware update aren't applicable here - all my measurements have been done in the last week with no intervening updates."@Frank99

Issue is accuracy of the number of kWh the battery stored. Unlike your kw into the car, we can't really measure the battery capacity which is half of the efficiency equation.

I'd guess the number is close enough but it is a "soft" number and we have seen recent evidence that the variation can be at least 6%. It could go both ways and decrease or increase efficiency.

Frank99 | 2019年10月23日

Fishy -
Please show your work, i.e. the equations that require "the battery capacity which is half of the efficiency equation".

In my work, I define efficiency as (power stored in the battery) / (power drawn from the wall). There is no term that requires the battery capacity. In smaller words, this is the percent of the power that the utility will charge you for that actually gets stored in the battery.

tucsonsims | 2019年10月23日

Great data!

"It's interesting (to me, but I'm weird) that there's a significant difference between 120V/12A and 240V/6A."

@Frank99 - I would surmise that this step function increase in efficiency at 240V is due to the nature of the boost converter topology in the onboard charger. A boost converter has to work harder (more losses) when boosting 120V rms to 400V DC vs 240V rms to 400V.

FISHEV | 2019年10月23日

"the equations that require "the battery capacity which is half of the efficiency equation".@Frank99

I think that was my question. How did you measure the "Stats showed 40.62 kwh added to the battery"?

Per Tesla StatsApp, my battery capacity, my Rated Range varies wildly between 302 to 320 since the 2019.x.x update.

WEST TEX EV | 2019年10月23日

+1 tucsonsims ... “nature of the boost converter”

Frank99 | 2019年10月23日

tucsonsims -
I'm not an expert on off-line switching converters, but it seems to me that for a desired output power level, 120V means that the duty cycle is going to be twice the duty cycle of 240V. But, the same could be said for doubling the output power level at the same voltage, and I don't see that in the data. That being said, I'm likely thinking only of low-power designs, and not something that can handle 48A at 240V...