Charging efficiency revisited

[brucet999]

I have searched TMC for charging efficiency threads but haven't found a simple way to determine how much AC power is drawn per kWh put into the battery and none of the posts used a 30A 240V source.

I am using a 30A dryer circuit and can see on the charging screen that I am charging at 234V, 24A. At my main panel the usage will be greater, since the UMC is seeing only what is available to it at the 10-30 outlet, after line losses.
When not in use, the voltage on the circuit reads 240, so I presume that usage at the meter would be measured at 240V, with the difference being from line losses.
If I set charging units to percent, I can record beginning and ending SOC percentages to calculate kWh received.
From time of charging at 24A 234V and converting to 240V as received at the meter, I can calculate kWh AC consumed by the charger.
Then it is trivial to calculate charging efficiency. Yes?
I should then be able to calculate wall-to-wheel cost.

Am I missing something significant?
Is there a better way?

 

[FlasherZ]

You'll want to measure the voltage at the main panel while charging to determine where the losses are going. For *most* cases, using the 240V circuit starting voltage will be fine - only reasons you'd want to do differently is if you had a very long and potentially undersized run to the transformer, or the transformer is really overloaded; then you'd want to measure voltage at your main panel while charging is taking place, because that will be the actual voltage that would allow you to calculate kWh reflected on your meter from charging time and current. That works for everything but going to 100%, as the current tapers near the end and will mess up your calcs.

The problem with using percentages for SOC is that Tesla's % reporting algorithm may not be linear in nature; the difference between 10-20% might be different than 90-100%.

The API will give you the energy added to the battery, to hundredths of a kWh - so you might want to use VisibleTesla to get that data. The API reports I've added 4.53 kWh to the battery since I've plugged in (chargeState.charge_energy_added). You could also get your exact charging duration from VisibleTesla, and multiply that times your 24A times your voltage under charge at the panel to determine what losses looked like.

The overhead of the UMC is fairly minimal, but is there - when the car says it's charging at 24A, it's not necessarily charging at 24.00A; my car says it's charging at 79A but at the same time the circuit on my panel shows 80... shorter charging sessions won't have enough precision to be really accurate.

If I were attempting to test this for my own use, I'd use the voltage at my main panel under charging load times charging time (hours) * current measured on that circuit (from a CT ammeter). I would then compare that to the car's reported charge_energy_added to determine efficiency. I think that's the best we have to go on right now.

 

[Great Dane]

I found the magic number to be 1.3 multiplier.
what ever the car says it used multiply by 1.3
gets you very close to actual usage

 

[brucet999]

You'll want to measure the voltage at the main panel while charging to determine where the losses are going. For *most* cases, using the 240V circuit starting voltage will be fine - only reasons you'd want to do differently is if you had a very long and potentially undersized run to the transformer, or the transformer is really overloaded; then you'd want to measure voltage at your main panel while charging is taking place, because that will be the actual voltage that would allow you to calculate kWh reflected on your meter from charging time and current. That works for everything but going to 100%, as the current tapers near the end and will mess up your calcs.

The problem with using percentages for SOC is that Tesla's % reporting algorithm may not be linear in nature; the difference between 10-20% might be different than 90-100%.

The API will give you the energy added to the battery, to hundredths of a kWh - so you might want to use VisibleTesla to get that data. The API reports I've added 4.53 kWh to the battery since I've plugged in (chargeState.charge_energy_added). You could also get your exact charging duration from VisibleTesla, and multiply that times your 24A times your voltage under charge at the panel to determine what losses looked like.

The overhead of the UMC is fairly minimal, but is there - when the car says it's charging at 24A, it's not necessarily charging at 24.00A; my car says it's charging at 79A but at the same time the circuit on my panel shows 80... shorter charging sessions won't have enough precision to be really accurate.

If I were attempting to test this for my own use, I'd use the voltage at my main panel under charging load times charging time (hours) * current measured on that circuit (from a CT ammeter). I would then compare that to the car's reported charge_energy_added to determine efficiency. I think that's the best we have to go on right now.

Thanks for the quick and complete answer.

As for checking current under load, the only place I would have access to the wires would be in the panel or a J-box midway from the panel to the outlet, about 40 feet from each end. Presumably the current would be the same as measured anywhere on the circuit even as line losses reduce voltage along the way.

I won't be purchasing an expensive new Ammeter, so my Fluke T5-600 would have to do, but its accuracy is rated only 3% +- 3 counts, whatever that means. My guess is that the Tesla amperage reading might be more accurate than 3%.

Testing panel voltage under load is something I had't thought of. Should be easy though, right across the breaker terminals or maybe the buses if they are accessible on a 50-year old SquareD Homeline panel. My guess is that 80 feet from the panel, the 234V that the Tesla sees will be from line loss on my side of the meter.

Is Visible Tesla available for ver 7 firmware? I thought I read that he had problems getting it to work right after the firmware change.

 

[FlasherZ]

Yes, current will be constant anywhere north of the 14-50 or HPWC.

As for VT, not sure... I don't use it.

 

[brucet999]

Yes, current will be constant anywhere north of the 14-50 or HPWC.

As for VT, not sure... I don't use it.

You get data from the API. Presumably this is something that requires a substantial amount of "geek skill" to use?

 

[FlasherZ]

You get data from the API. Presumably this is something that requires a substantial amount of "geek skill" to use?

The way that I do it, I would say "yes". However, VisibleTesla and Teslalog.com (see TeslaLog.com - Your hosted Tesla Data Logger - Announcement / Support threads for more information) are easier and more user-friendly ways to work with the data.