Measuring the Total Energy Consumption of Your EV

Measuring the Total Energy Consumption of Your EV

Part 1 – Tezco’s Homebrew Energy Monitoring and Data Logging Device

In the following posts I’m going to describe how I built a data logging system to measure my EV energy consumption. Other posts have described monitoring systems using TED – The Energy Detective; however, TED relies on data transmission through your home’s wiring (aka PCC – Powerline Carrier Communication). PCC has it’s drawbacks including sketchy transmission over long distances, interference from on-line noise, and attenuation when noise filters are on the same line.

Since my garage is detached and several hundred feet away from my main breaker panel, and my home office power is heavily damped due to large UPS’s, I first took at look at the TED forums and read about the difficulties that some users had with TED in similar situations. I also wanted to monitor temperature readings in my garage, so I decided to take a road less traveled.

Next – Data Summaries


Part 2 – Data Summaries

I'll edit this section from time to time to show some of the data that I have collected.

Revised 02-15-13:
Data points on the following graph show the energy consumption (normalized to idle energy loss per 24 hrs) of the car when it sits parked in the garage, referenced against the ambient garage temperature about 8" off the floor (ie at battery pack level). The pink square represented the night the car upgraded to v4.2 (no significant increase in energy use, despite all the bells and whistles that go off during updates). Note that my firmware versions never have included active sleep mode. It appears that there is more loss at colder temperatures (or perhaps it takes more energy to keep the pack at proper charging temp during the charging cycle?). There is a moderate scatter of data points which suggests that the SOC at the end of each top-off cycle may vary somewhat.




Next – Assembly of the Data Logger

Part 3 – Assembly of the Data Logger

My system is built with components that are available online from Onset Computer Corporation, as well as miscellaneous hardware items that can be purchased at Home Depot and your local hardware store. The system uses a data logger which locally records four channels of information. The data sampling rate can be set by the user, and can be as short as every second. The logger holds 43,000 readings (which is about 7.5 days worth of four channel logging, when set to sample once a minute).

The data in the logger can be downloaded via a USB cable to a laptop or a data shuttle, or you can unplug the logger and bring it over to your PC. Since my logger is mounted inside a junction box, I didn’t want to have to turn off the power to open the box to access it or to physically take out the logger so I decided to go with the data shuttle. Initializing the logger does require that you connect the logger and it’s attached sensors to a computer which is running the HOBOware Pro software. Since the license is for one computer only and I wanted to run the software on my PC, I decided to make my whole setup mobile, so I could bring it in to the computer for initialization. Therefore, my device has a plug on one end, and a receptacle on the other. I just unplug the EVSE from the wall, and insert the box in-between.

The HOBO logger’s Li button battery is good for about a year, but the life is shortened at low temperatures. I’ve put a nightlight in the box, and monitor the interior temperature with the 1 foot temperature probe. The other 6 foot probe hangs down out of the box and is held about 6 inches off the garage floor, to approximate the air temperature at the level of the Tesla battery pack.

The main components purchased were:
(see http://www.onsetcomp.com/products/data-loggers/u12-006 )

U12-006 (HOBO U12 4-External Channel Data Logger)
BHW-PRO-CD (HOBOware Pro Software)
Cable 4-20MA (Connects voltage sensor to logger)
CTV-B (0 - 50 Amp split-core AC current sensor)
TMC1-HD (Air/Water/Soil Temp Sensor -- 1' cable)
TMC6-HD (Air/Water/Soil Temp Sensor -- 6' cable)
U-DT-1 (HOBO U-Shuttle)

Disclaimer

I’m not a licensed electrician, so I can’t vouch for the Code compliance of my system. Certain aspects are not up to code; for example, I clamped the voltage monitoring leads into the output receptacle using the same clamps that hold the main power leads. My garage also had an old NEMA 10-50 outlet that I use for welding. This is code compliant when used for 240V with two hot leads and a neutral, but mine was wired with a ground instead of the neutral. Fortunately this works with the GFI needs of the AV EVSE that I use for charging. You might want to use more a modern NEMA 14-50 for example, and might want to purchase a higher amperage split core sensor if you are thinking of charging at more than 50 amps in the future.

Next – Photo of the Parts, Pre Assembly

- - - Updated - - -

Part 4 – Photo of the Parts, Pre Assembly

I purchased a Raintight junction box and drilled out a 2 1/8 inch hole on the side for the outlet receptacle using my drill press and some oil drizzled on for cooling during cutting. The wooden plywood block is used to mount the Data Logger, such that cold from the box wasn’t transmitted to the logger, and the light bulb is used to keep the logger warm during the winter. The four plates with the tabs are flush mount hangers which make it easy to pull the box off the wall for the logger’s initial startup, or re-initialization should the battery fail. (The shuttle can normally power the logger while you replace the battery (which preserves the initialization settings), another plus for buying the shuttle.)



Next – Photo, Partial Assembly

- - - Updated - - -

Part 5 – Partial Assembly

The major parts are now installed. Voltage sensor is on the left, current clamp on one leg of the mains, and heater light bulb on the right.



Next – Photo of Assembled Device

- - - Updated - - -

Part 6 – Photo of Assembled Device

The Logger is installed. Exiting the box thru a grommet bushing behind the wood block is a the USB cable (to be used when I plug in the data shuttle), the long temperature probe, and a lamp cord that connects to the nightlight. The cord is plugged into a nearby 120V receptacle in the winter to keep the logger toasty. Might have to move up to a 7 watt bulb if this cold weather continues….



Next – Initial Plot

- - - Updated - - -

Part 7 – Initial Plot

Here is my first data. Downloaded from the logger to the shuttle, sent over to the home PC via sneaker net and uploaded to the HOBOware Pro software package. This software includes some statistical subsets that can, for example, analyze the 10 hour window I have drawn between the vertical black lines which shows the mean temperature near the garage floor was 28.5 degrees (Solid green line). (You can see a few squiggles in that line: Negative blip when I went in and out of the door, and positive blips when the heat turns on (set to come on at 34 degrees, but stays colder at the floor.)

The dashed green line is the temp inside the box, which fell from 70 degrees and finally stabilized around 40 degrees overnight. The solid black line measures the amps sent to the EVSE and Model S, and the blue line, the voltage. I had the S unplugged for a couple of days, so the 2-hour charge shows just how much the “vampire” load is (still waiting for the 4.0 or 4.1 or 4.2 update, so currently no display sleep mode in my S).

ps Has anybody measured the power factor of the charging system, so I can calculate the actual kW?

EchoDelta said:

Wait- vampire load is a daily 4.7 kWh ... is that with the car plugged in in a non-charging state... (Junction to car loss, in which case it would be a % not kWh/day) ... or of the charger without a car plugged in?

Its high. That's like 25% of my daily household electricity use. Would you mind clarifying what is plugged in and in what state as you accumulate towards that reading?...

Click to expand...

I have an AV EVSE that I use for charging my Leaf and the S, so only one of the two is "plugged in" at a time. I don't have my data here at work, but if I remember correctly, the S completed it's standard (85%) charge in the morning. It was then un-plugged and sat in the garage (around 30 F) for about 9 hours, at which time I plugged it back in and a new top-off charge began. Since I haven't received the software upgrade that allows sleep mode, my electronics are on 24/7. Also unknown is whether any energy went towards battery heating. (The Leaf supposedly doesn't run the battery heater until the pack reaches -4 F, but I'm not sure that we know under what conditions the Model S heater comes on when the car is idle.) In any event, I calculated the kWh of the second charge and then normalized it out to 24 hours.

I took a 70 mile trip Sunday and will post the charging plot and results when I get a chance to double check my figures. It looked like I was getting a hair over 2 miles per kWh off the wall, but it was only about 5 F, so the heater draw was high. If I hadn't backed out the "vampire load" losses, the mileage would have been even worse.

The Leaf has very little "vampire load" losses, but is getting a similar mpkWh (which is about half of what I get in the spring or fall).

Here is a HOBOware plot of Monday's energy usage. Note that I have set the measuring calipers to encompass the Tesla charging peak, and I can see the duration of the charge (8hrs, 40 min), and from the stats at the left, the mean voltage (233.98) and the mean current (S was set to limit to a 20 amp draw; actual draw was 19.633). The charge followed a 70 mile trip Sunday evening.

Previously I estimated idle power losses about at 5.8 kWh per 24 hours. Since the drive time was 1.25 hours, the remaining 23.75 hours that the car sat in the garage were subject to idle power loss (5.34 kWh). The drive required the rest of the 40.73 kWh charging input (ie 35.43 kWh), which means that the car got 1.98 mpoowkWh (miles per out-of-wall kilowatt hours) if my spreasheet formulas are correct. (Ambient temp probably averaged 10 F during the drive.) The ambient garage temp was probably around 30 F while the car was in the garage on Sunday, which is similar to the conditions that existed when I calculated the idle power losses.

In comparison, the Leaf got about 2.12 mpoowkWh while driving a shorter distance and at lower speeds, which is about 1/2 of it's typical spring and fall mileage.

pbrulott said:

Teczo, thanks for this precious info.

If I understand well, at 10F driving, your wall to wheel consumption was 1.98 miles per ouf of wall kWh and that would compare to 3.27 miles per battery kWh Rate EPA. A 65% factor that would be split between 1) charging efficiency loss and 2) cold driving impact onto the wh/mile. The other thing is that at 30F, you saw 5.34 kWh from the wall loss due to battery heating.

Some posts on the charging efficiency topic reported 55% losses on 110V and 30% on 240V. TM Product Specialist are telling me that charging current has no impact on charging efficiency... So what I'd like to know is the charging efficiency piece. Would you be able to isolate #1 vs #2 knowing your wh/mile during the drive?

Click to expand...

I think that most of my idle power losses are due to the drain from the electronics being powered up 24/7 (I don't have the newer software that allows sleep). I don't know if any electrons are used to heat the battery at my ambient temps; I don't Tesla has given us much information as to when battery heating clicks on when the car is idle, or when moving.

As you have also pointed out, the inefficiency in the charger itself cause the idle power loss calculation result to be higher than it actually is, since the charger inefficiencies can't really be separated from the idle power losses.

The percentage of wall power that goes to the wheels vs idle power losses ("vampire load") would be different every day, depending on what distances you drive. Today the car is sitting in the garage, so 100% goes to idle loss (and charger inefficiency).

In my calculations I've included the power that goes to the electronics when the car is moving as a necessary evil to get the wheels moving, so even though it doesn't really go directly towards propelling the car, I can't really call it an idle power load. If you subtracted that energy, the mpoowkWh would be higher.

I'm not sure about the statement that charging efficiency is the same at all charging currents. My Leaf takes much longer to charge on 120V -- you'd expect about twice as long, but it is more like 2 times plus an extra 1/3. This is due to the fixed energy requirements for the accessory systems that run during charging (coolant pumps, electronics etc). Add twice the current, and all of that second addition goes towards charging.

I've got my charge dialed back to 20 amps because there were some posts on the 12V threads that report that some service managers recommend long charging periods to help keep the 12V charged up. More recently there have been some indications that it may be overcharging, rather than undercharging the 12V, that is causing problems, so perhaps I'm doing the wrong thing.

I'm going to try to track the idle power losses for a few days to see how much they vary. (For example, the SOC may not be exactly the same each day, which could change charge times.) Unfortunately the temperature in the garage varies a bit, which is another variable that could make a difference.

Once I know how consistent the idle power loss is, then I can try varying the charging amps to see if efficiency changes. I think it would be harder to do using the energy calcs from driving -- my foot some days is very gentle on the accelerator, but other days I just get an itch to enjoy the performance!

tzejohn said:

....I did do a meter reading between 5pm one day and 8am the next day, where the battery was already fully charged, and I did not see any increase in kWh consumed on the circuit.

Click to expand...

You would have to initiate a new charge to see what has been lost in the interim. Since I'm switching my J1772 back and forth between my S and the Leaf daily, I get a fresh charge cycle whenever I plug the S back in. You do need to know when the prior charge completed and then when the current top-off charge started so you can figure how many hours were inbetween. If that interval was 12 hours for example and the kWh for the top-off charge was 12 kWh, then you would have an idle power loss of 12 kWh every 12 hours (or 1 kWh each hour, or 24 kWh for each 24 hour day).

- - - Updated - - -

tzejohn said:

I don't have as much detailed information. My meter is only reporting total kWh consumed by the circuit. I regularly do a 78 mi roundtrip which, according to the car reading, consistently use about 29kWh. Reading the meter on my circuit, that is showing 38kWh. This is on a daily basis. Does that extra 8kWh seem high to anyone?

Click to expand...

(Forced to split up this post so I can see my Post Ticker go to 100!) (Of course it's a very long way to catch up to vfx's 12,777 posts!!!!) Edit: Dang, forgot that a second post gets added to the bottom of the first! 100 is still around the corner....

The 8 kWh difference includes the idle power losses as well as the charger inefficiency loss. Haven't figured out a good way to separate the two apart, but the more you drive, the longer you will need to charge, and the larger the power loss from charger inefficiency becomes. True idle power loss might be 4 kWh plus another 2 kWh due to charger inefficiency if no driving that day, but in your case, the loss in the charger is going to be larger.

Here is a summary of the data I've collected this week

Data points on the following graph show the energy consumption (normalized to idle energy loss per 24 hrs) of the car when it sits parked in the garage, referenced against the ambient garage temperature about 8" off the floor (ie at battery pack level). Note that my software version is still 1.15.14, so sleep mode does not yet exist for my S. The highest energy loss was at the lowest ambient temperature, but it seems much higher that the other data points. This data was obtained from a 9 hour idle period; the other data points are derived from 24 hr idle periods. Possibilities for the non-linearity of the graph include a trigger for battery heating exists below 30° F, idle power loss isn't linear, there is a variation in the SOC from charge to charge, or some other factor.

I've kept track of the kWh from my monthly power bill for several years now. My consumption seems to be quite variable. I haven't really been able to notice the difference before and after I got the Leaf last Jan, but it doesn't use as much power as the S.

A few more data points on the graph...

A bit more data; adjusted scale on X-axis.