Soul Spy?

[JejuSoul]

Have added data for 1 more car, and a new column Accumulative Charge Power

Car         t           Age          km      max det      min det          ACP kWh
T           9.7          ?           ?            0           0             34.9
U           106.1        1           1500         0           0            359.7          
C           180.7        2           3266         4           0.8          657.9 
K           308          6           7000         2.5         0            915
G           484.5        6           5500         8.3         0.8         1471.9
S           1443         6         18,084         9.3         8.3          3040.6   
N           1688         ?         20,024         10.8        9.3         4302.5

Car - A random letter to identify the car
t - Accumulated Operating Time in hours
Age - Age of car in months
km - distance on the Odometer in km
max det  = maximum deterioration of a cell in %
min det  = minimum deterioration of a cell in %
ACP = Accumulative Charge Power in kWh

Some oddities about this data. It seems many of these cars have been stationary with the engine on for long periods of time, because their average speed seems too low. Possibly because the owners have been hacking the OBD while the car is stopped. Also one car has used far more power than the mileage suggests, perhaps that owner had the heater/air con on while doing those many hours of testing.
Alternatively we may have factored some of these figures incorrectly.
But for my car they seem roughly accurate. My average speed (according to the trip meter in the car) is 30km/h not 22.7km/h as these figures suggest.
My average power consumption (according to the meter in the car) is 7.7km/kWh or 7.65km/kWh as these figures suggest.

 

[goev]

The age of N is 10 months

If there is anything you need me to check on my car, just ask!

 

[TyrelHaveman]

Some oddities about this data. It seems many of these cars have been stationary with the engine on for long periods of time, because their average speed seems too low. Possibly because the owners have been hacking the OBD while the car is stopped. Also one car has used far more power than the mileage suggests, perhaps that owner had the heater/air con on while doing those many hours of testing.
Alternatively we may have factored some of these figures incorrectly.
But for my car they seem roughly accurate. My average speed (according to the trip meter in the car) is 30km/h not 22.7km/h as these figures suggest.
My average power consumption (according to the meter in the car) is 7.7km/kWh or 7.65km/kWh as these figures suggest.

It's probably better to use accumulative discharge power than charge power -- since the car starts its life with the battery somewhat charged, but it'll start of reading 0 charged and 0 discharged, but then start driving and show positive discharged; it's a more accurate measurement of how much energy the car has actually used.

Isn't it amazing that Kia's engineers let them get away with calling this "power" instead of "energy" in the UI, by the way? Maybe they don't speak English and it's a poor translation from Korean.

Also, unfortunately I don't know the exact odometer reading on my vehicle when I recorded the data I've posted in the spreadsheet, so any calculation you're doing with km is likely to be incorrect.

Our car has spent many hours parked, but on, without driving, largely due to my experiments but also sitting at an electric car show we ran before we purchased the vehicle (the dealer was showing it there for us). I've also spent some time running the heater while testing in order to watch the infotainment SOC% drop so I could record data from the C_CAN at that instant.

 

[JejuSoul]

The site I like which explains the difference between energy and power is
Do the Math - Using physics and estimation to assess energy, growth, options—by Tom Murphy
He has a page on Useful Energy Relations here - http://physics.ucsd.edu/do-the-math/useful-energy-relations/
"A kilowatt-hour is the amount of energy accumulated by running something at 1000 W for one hour"

I'll have a look at the accumulative discharge figures tomorrow - although I expect them to be very similar to those for accumulative charge.

 

[TyrelHaveman]

Energy (kWh) = Power (kW) * Time (h)

The interesting unit to me is "Ah", which indicates a measurement of electric charge, which on the fundamental level is a measure of how many electrons have been moved (1 Ah ≅ 22,900,000,000,000,000,000,000 electron charges). (We've moved a LOT of electrons in our cars!)

I'll have a look at the accumulative discharge figures tomorrow - although I expect them to be very similar to those for accumulative charge.

On mine it looks like the cumulative discharge energy is always slightly more than cumulative charge energy.

 

[JejuSoul]

Playing with the data above in Excel I created this chart.

Percentage Remaining is calculated by averaging the Maximum and Minimum Deterioration, the subtracting this from 100.
Number of Cycles is calculated by dividing the Cumulative Charge by 27 (the usable amount on the battery)
Note that none of the cars charge a full 27kWh every day. It varies from once every 2 to once every 5 days.

This method is not intended to be exact. Just playing with the data to see if it gives reasonable results.
If it were accurate the majority of the cars in this sample would qualify for a battery replacement in 7 years.

 

[JejuSoul]

Have corrected the wording for Cumulative Charge and added Cumulative Discharge

Car         t           Age          km      max det      min det          CC kWh        CD kWh
T           9.7          ?           ?            0           0             34.9          16.1
U           106.1        1           1500         0           0            359.7         361.2      
C           180.7        2           3266         4           0.8          657.9         689.0
K           308          6           7000         2.5         0            915           1414.8
G           484.5        6           5500         8.3         0.8         1471.9         1465.8
S           1443         6         18,084         9.3         8.3          3040.6        2940.3    
N           1688        10         20,024         10.8        9.3         4302.5         4294.9 

Car - A random letter to identify the car
t - Accumulated Operating Time in hours
Age - Age of car in months
km - distance on the Odometer in km
max det  = maximum deterioration of a cell in %
CC = Cumulative Charge in kWh
CD = Cumulative Discharge in kWh

I don't understand the figures for Cumulative Discharge. Why are some smaller others larger than the Cumulative Charge? Why is one Cumulative Discharge (car K) totally different from its Cumulative Charge. That car is mine. The OBD readout for that line is 7EC 26 00 03 93 C0 00 37 94 - byte4 the C0 is different to all other measurements taken so far. Everyone else has seen 00.

 

[2016Electric]

Playing with the data above in Excel I created this chart.

Percentage Remaining is calculated by averaging the Maximum and Minimum Deterioration, the subtracting this from 100.
Number of Cycles is calculated by dividing the Cumulative Charge by 27 (the usable amount on the battery)
Note that none of the cars charge a full 27kWh every day. It varies from once every 2 to once every 5 days.

This method is not intended to be exact. Just playing with the data to see if it gives reasonable results.
If it were accurate the majority of the cars in this sample would qualify for a battery replacement in 7 years.

Umm.. I hope 100 represents 1000 etc? Something can't be right. That deterioration is far to high.. and likely the relationship isn't linear. At that rate, people will be down to 70% after 300 cycles. At around 150 km per total discharge, That's 70% after only 45000 km. I must be interpreting your information incorrectly.

 

[TyrelHaveman]

I don't understand the figures for Cumulative Discharge. Why are some smaller others larger than the Cumulative Charge? Why is one Cumulative Discharge (car K) totally different from its Cumulative Charge. That car is mine. The OBD readout for that line is 7EC 26 00 03 93 C0 00 37 94 - byte4 the C0 is different to all other measurements taken so far. Everyone else has seen 00.

You also need to look at the last byte of line 25, it's the first byte of the cumulative charge.

 

[JejuSoul]

Some more detail about the Cumulative Charge / Discharge oddity on my car.
The OBD readout from two weeks go is :
7EC 25 59 EE 50 00 09 32 00
7EC 26 00 03 93 C0 00 37 94

And the OBD readout today
7EC 25 B0 92 00 00 B0 17 00
7EC 26 00 3F 9F 00 00 3D 91

In the first set of data the two values Cumulative Charge / Discharge are far apart which seems very strange, and
byte4 the C0 is different to all other measurements taken so far. Everyone else has seen 00. I suspect this is an error flag.

In the second set of data the two values Cumulative Charge / Discharge are close together which seems normal, and
byte4 is zero. It seems the first value for Cumulative Charge was an error.

 

[SiLiZiUMM]

Hello all,

I just went to the repair shop to visit my real car (the titanium one), to see the main battery health (it's been parked there for 2 months now). The main battery got 2% drop from 80% to 78%. The 12v battery was dead on arrival (we used a portable battery to power the car). I added 2101/2105 readings in the sheet! There were no trouble codes.

We'll have to wait at least 2-3 more weeks, there is a missing wire harness for brake lights. The accident happened on August 11, 2015...

Data from the car (using my parser):

org.hexpresso.soulevspy.util.BatteryManagementSystemParser$Data@b0c2a93[
accumulativeChargeCurrent=833.6
accumulativeChargePower=311.90000000000003
accumulativeDischargeCurrent=860.9000000000001
accumulativeDischargePower=315.40000000000003
accumulativeOperatingTime=706166
airbagHwireDuty=80
auxiliaryBatteryVoltage=11.9
availableChargePower=68.61
availableDischargePower=90.0
batteryCellVoltage={3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.94,3.96,3.96,3.96,3.94,3.96,3.94,3.96,3.96,3.96,3.96,3.96,3.96,3.94,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96,3.96}
batteryDcVoltage=380.40000000000003
batteryInletTemperature=6
batteryMaxTemperature=6
batteryMinTemperature=6
batteryModuleTemperature={7,6,6,6,6,6,6,7}
driveMotorSpeed=0
heat1Temperature=6
heat2Temperature=6
maxCellVoltage=3.96
maxCellVoltageNo=54
maxDeterioration=1.0
maxDeteriorationCellNo=26
minCellVoltage=3.94
minCellVoltageNo=67
minDeterioration=0.0
minDeteriorationCellNo=3
stateOfCharge=75.0
stateOfChargeDisplay=78.5
]

 

[JejuSoul]

Something can't be right. That deterioration is far to high.. and likely the relationship isn't linear. At that rate, people will be down to 70% after 300 cycles. At around 150 km per total discharge, That's 70% after only 45000 km. I must be interpreting your information incorrectly.

The dataset we have is only 6 vehicles. And they are a very unusual set. But it is a mixed set. Some have very high deterioration rates and some have very low. Overall it does seem to show a much faster deterioration than we had hoped for. Is it valid, I don't know. The data only becomes useful to make comparisons and predictions when we have a greater sample, and a longer time scale to see how the data matches reality.

...likely the relationship isn't linear. I assume it is linear. What kind do you expect?

There is a data set for the 1eaf EV that is freely available. see http://www.pluginamerica.org/surveys/batteries/leaf/results.php
The deterioration rates of the 1eaf are certainly not linear with use. I downloaded the data and experimented with the same kind of graphs.
The primary result for the 1eaf is - The battery degrades faster in a warmer climate.

Oddly in our tiny dataset of 6 vehicles the reverse is true.

I don't like the way this website converts some words automatically into URLs, so I changed the name of Nissan's EV to 1eaf.
I know that is how this website makes money but I find it distracting in my posts.

 

[TyrelHaveman]

The primary result for the 1eaf is - The battery degrades faster in a warmer climate.

Yes, this is very well known. I live in a mild climate and 2013+ LEĄFs especially are doing extremely well. I don't know of anyone who has less than 88% original capacity, and some of these cars are nearly 3 years old now. But you can find many of them from Los Angeles or San Diego, where it's much warmer, which easily have 30% or more loss.

In my climate, most people with 2011-12 LEĄFs lose the first capacity bar (which represents 15% loss) around 35,000 miles / 60,000 km. 2013+ LEĄFs are seeing much less. There's a 2013 LEĄF taxi in England, which has a similar climate to here, which hit 100,000 miles with only 10% degradation. Screenshot from LEĄF Spy here. And that one gets quick charged several times per day!

Since the Soul EV blows climate control air through the battery, but the LEĄF battery is completely sealed, I would expect that climate will have less of an impact on a Soul EV than a LEĄF. However, since the climate control isn't always on, outdoor climate will still have an effect.

By the way, I think a discussion about battery degradation should go in a separate thread, so we can focus on the CAN bus specific stuff itself here, if possible.

 

[Elmil]

Tyrel: I had a look at your latest BMS data in the spreadsheet (BMS_compare, T#8), and I think you had your L2 charger plugged in at that time, though not charging. This shows by the (unknown) byte at position (2101: 21 99 21 2D 23 28 20. I have noticed 3 bit flags in this position that are set in the following conditions:
bit7=charging
bit6=L3 (CHAdeMO) plugged in
bit5=L2 (A/C) plugged in

Also, bit0 and bit1 always seem to be set in the "running" state, but I have no idea what they mean..

 

[SiLiZiUMM]

Elmil : This would make sense. Did you data while doing L3 charging to confirm byte 6? I'll visit my sister this weekend and will use a lot of L3 chargers, I can check at that moment!

 

[Elmil]

Elmil : This would make sense. Did you data while doing L3 charging to confirm byte 6? I'll visit my sister this weekend and will use a lot of L3 chargers, I can check at that moment!

Yes I did. This byte is C0 when I charge CHAdeMO and 40 when charging has stopped but connector still plugged in; A0 when charging L2 and 20 w/o charging.

Another thing that bothers me: The battery amperage in 2101: (21.7+22.1) does not scale correctly when charging (negative values) in my opinion.
When I'm charging using the original 2 kW EVSE cable, then at the wall I read around 1.9-2.0 kW, but the battery voltage * amperage from OBD wouldn't be more than 1.1 - 1.3kW. With an efficiency of 85%, the expected amps to the battery would be about 4.8A at 350V. I see amps between 3.0-3.5A around 350-360V. Will try to do a comparison next time I use a CHAdeMO that displays the charging current and see if the same mismatch can be seen.

 

[TyrelHaveman]

Tyrel: I had a look at your latest BMS data in the spreadsheet (BMS_compare, T#8), and I think you had your L2 charger plugged in at that time, though not charging. This shows by the (unknown) byte at position (2101: 21 99 21 2D 23 28 20. I have noticed 3 bit flags in this position that are set in the following conditions:
bit7=charging
bit6=L3 (CHAdeMO) plugged in
bit5=L2 (A/C) plugged in

Also, bit0 and bit1 always seem to be set in the "running" state, but I have no idea what they mean..

You are correct, I did have the J-1772 plug attached but not charging (timer set).

 

[TyrelHaveman]

Another thing that bothers me: The battery amperage in 2101: (21.7+22.1) does not scale correctly when charging (negative values) in my opinion.
When I'm charging using the original 2 kW EVSE cable, then at the wall I read around 1.9-2.0 kW, but the battery voltage * amperage from OBD wouldn't be more than 1.1 - 1.3kW. With an efficiency of 85%, the expected amps to the battery would be about 4.8A at 350V. I see amps between 3.0-3.5A around 350-360V. Will try to do a comparison next time I use a CHAdeMO that displays the charging current and see if the same mismatch can be seen.

You get a 2 kW cable? In the US, the cable we get can plug in to a 15 A x 110 V outlet, and using the 80% rule for constant current, cannot legally draw more than 12 A, which at 110 V would be just 1.32 kW.

I have a power monitoring device in my home and have the EVSE we use for the Soul separately monitors, so I can check that while looking at the reported amps to verify as well. The Soul draws 6.8 kW usually, at 240 V, so ~28 amps.

I can also check with CHAdeMO but the one in my town doesn't display amperage so I wouldn't be able to compare; there's one about 60 miles away that does display amperage, so the next time I'm down there I can check it as well.

I'll add info to the spreadsheet based on what you've found, btw.

 

[SiLiZiUMM]

I'll check that this weekend, here all CHAdeMO chargers do display the charging stats. From an older pic I took, I can read 122A @ 390V

 

[TyrelHaveman]

Okay, I'm in the car now, charging on L2, and looking at the data.

My power monitor says I'm currently at 28.8 A * 246 V = 7.1 kW

The data from the car:

21 3F 23 28 23 28 A3 FF
22 5A 0D A6 0F 0E 0F 0E

These are showing amperage = FF22 = -166 = -16.6 A, and voltage = 0DA6 = 3494 = 349.4 V, so power = 5.8 kW.

I have a couple thoughts on the discrepancy. First, the car was reporting 0.1 A usage before it started charging, so it is probably pulling in a little more than would be apparent here. In addition, there are going to be line losses between where I'm measuring the input and where the car is getting it.

I started wondering if I could find out what voltage and amperage are actually coming into the car, as my power monitor is some distance from the EVSE itself, and there are line losses between there and the EVSE, within the EVSE itself, and in the J-1772 cord, etc.

Yesterday I'd found out that you can get data from the on-board charging (OBC) by querying command 21 02 from the OBC, 79C. This is what I get while charging now:

79C 10 17 61 02 E8 03 1F 00
79C 21 01 00 09 75 10 6C 12
79C 22 8B 0D C2 01 1F 00 B0
79C 23 2B 24 32 00 00 00 00

The actual data within these messages is:

E8 03 1F 00 01 00 09 75 10 6C 12 8B 0D C2 01 1F 00 B0 2B 24 32 00 00

Compared to the data when I was not charging:

E8 03 1F 00 01 00 00 00 00 00 00 00 00 00 00 00 00 00 10 12 0F 00 00

There's a lot going on here, but likely the voltage is in there somewhere. As I mentioned previously I'm at 246 V right now; probably somewhat lower than that by the time it gets into the car.

Several minutes later, still charging at about the same rate, these are the data:

E8 03 1F 00 01 00 09 91 10 6B 12 8D 0D C9 01 20 00 B2 2B 29 2F 00 00

I believe voltage is "09 91" and "09 75" -- these are 2449 (244.9 V) and 2421 (242.1 V) respectively, which make perfect sense.

Amps is probably the "01 20" and "01 1F" -- 288 (28.8 A) and 287 (28.7 A) respectively.

So, the car is getting 244.9 V * 28.8 A = 7.05 kW, and putting about 5.83 kW in, which is an efficiency of 82.7%. Not great, but not terrible.