[IonicEV]

Another "rumor" is that 90% is usable and total capacity is 31 kWh. I hope this issue can be settled once and for all by these Torque readings.

I suspected it all alone. Ionic for sure behaves quite different(better) from other EVs while charging and when getting close to empty(I ran demo one to 5 miles left with no panic or cliff hanger like in Leaf; or Bolt which just shuts down with less than 25 miles left). It is easy to figure the upper buffer, but lower is not that easy as you need to go very low to figure it out. I would not be surprised if lower buffer even bigger than upper. Battery in Ionic behaves like bigger battery, may be even bigger than 31kWh, usually capacity is even number, so it would not be a stretch to assume 32kWh total capacity.
I setup Torque Pro relatively easy on Mac with Android File Transfer app. Just created (missing) "extendedpids" folder under ".Torque" folder, then dragged and dropped files to corresponding folders under .Torque folder on the phone. The rest is easy, but I had to battle with few OBD II Bluetooth adapters either not compatible with Torque Pro (BTLE LELink 2 and Veepeak - could not even pair, lol, it was WiFi). Finally, some no name ELM327 adapter paired and Torque Pro started to receive the data. I ordered recommended "Konnwei KW902" to stay away from any troubles in the future.

BTW, I think it should be relatively easy to determine how accurate SOC reported by OBD. Li-Ion batteries has very well documented voltages-to-SOC.

 

[IonicEV]

SOC BMS does not relate to the total capacity. It is another way to measure existing usable capacity. It will always vary between 2% and 95%. Hence it cannot be used to measure either deterioration or total capacity. If someone can show me I'm wrong I'd be pleased. It would be nice to finally have an answer to the question - what is the total capacity?

I think the only way to get closer to real capacity figure, by taking voltage measurement and it should very accurately point to SOC of the battery regardless of capacity. We know 28 kWh is usable. It should be possible to find those top and bottom buffers when we know the real SOC (not BMS nor Display). The only thing I would caution about: the voltage value-to-SOC during charge and discharge cycle could be quite different, so the best way is to stick to discharge cycle(no charger connected) voltage as it excludes any charger interference.

 

[IonicEV]

Here is some readings I took on slightly discharged battery:

BMS SOC: 89.5%
Cell Voltage: 4.04V

It is known Li-Ion cell max voltage is 4.2V at 100%. So at least SOC and voltage do not match with SOC higher than voltage suggests. Have to dig more info and take readings at full charge.


Below is some info on Voltage vs. SOC of Li-Ion 18650 cell from other forum:
SilverFox
Li-Ion State of Charge and Voltage Measurements

There has been a lot of discussion on how to figure out the state of charge on Li-Ion cells by measuring their resting voltage.

I picked up some information on high current draws that gives the following values:

4.2V – 100%
4.1V – 87%
4.0V – 75%
3.9V – 55%
3.8V – 30%
3.5V – 0%

Please note that resting voltage means the cell has stabilized at room temperature and the voltage has also stabilized.

I decided to check a brand new 18650 cell at a defined current draw. This cell is a Pila 600P rated at 2200 mAh. The test current was 2 amps with a low voltage cut off of 2.8 volts.

At a 2 amp current draw, this is what I observed:

4.20 volts – 100%
3.97 volts – 80%
3.85 volts – 60%
3.77 volts – 40%
3.72 volts – 20%
3.58 volts – 0%

This cell tested at 2000 mAh capacity at 2 amps. I ran 2 amps for 400 mAh, then stopped the test to let the cell and voltage stabilize. I then continued to do this 5 times to come up with the values listed.

 

[IonicEV]

Some interesting EV battery info

Yesterday, an opportunity presented itself to take measurements of the Ionic battery in turtle mode (4 miles range left) and then fully charged. Also, the good news that it delivered all reported 137 miles after last charge to 96%(Display). It also demonstrated the outstanding experience while handling low battery conditions all the way to 4 miles, including the turtle mode when remaining range flipped from 5 to 4 miles left without any hint of panic or freakiness (--- fits on Leaf), as it simply re-mapped accelerator sensitivity to reduce the risk of power surges that could trigger voltage drop below safety thresholds. That is my friends very, very rare occurrence in EV world. And, it is not incidental as later data capture suggested the reasons for this graceful handling.

I own home charger JuceBox 40 Pro and it is smart enough to figure out exact amount of charge delivered to the vehicle battery during charging sessions with nice graphing and time info (it knows its own efficiency and specific vehicle charger efficiency). The fact it is a pro version worked in my advantage as it stopped charging once it delivered 28kWh. This offered yet another opportunity to capture the battery state at 91.5%(BMS)/96%(Display) SOC. Later I have edited charger vehicle info and increased total capacity to 30kWh, so charger delivered more to 95%(BMS)/100%(Display) SOC. It will help with approximations later.

I took screen captures from JuceBox app. It is interesting to compare it to Leaf 24kWh.

Let's start with pictures first:

1. Leaf 15 SV 24 kWh:

Notice how long is the final 10% it takes and how erratic charging current becomes (fuzzy graph) and then 3 distinctive pops most likely attributed to active balancing attempts during charging.


2. Ionic phase 1 - 91.5%(BMS)/96%(Display) SOC

3. Ionic phase 2 - 95%(BMS)/100%(Display) SOC

Notice how quick is the finishing taper - most likely no active balancing takes place as it is rather quick and very steady current steps all the way to end of charging. Presence of those steps indicates the fact - it is enforced by the charger, not by battery getting close to full. This again suggests the battery capacity is bigger vs. disclosed figures. Probably by large number that one might guess. We can also compare to initial Leaf current tapper suggesting it is indeed battery induced (pattern close to analog noise, not well defined steps). Leaf' 8-10% top buffer would put usable capacity to 21.6-21.8kWh (observable/confirmed by many owners) out of 24kWh battery.

Also you might noticed the total charge delivered to the battery was 28+1.3 = 29.3 kWh. With an additional 4.5%(Display) SOC remaining - it would put total usable Ionic EV battery capacity to: 29.3+1.26 = 30.58 kWh. And this is below freezing temperatures. During summer it would most likely get over 31kWh usable capacity. So at this point we could settle on conclusion Ionic EV has 31 kWh usable for driving.

The remaining question is what is real battery capacity. The only reason we might want to know is to assess the durability of those usable figures over time. More excess capacity the better chances of not experiencing usable capacity degradation in the future.

I will continue later when I finished massaging the captured data, hopefully leading to a closer estimate of the real capacity of Ionic EV battery.

...

 

[IonicEV]

Just publishing raw figures obtained at turtle and following with 100% charge:

Turtle:
Battery Voltage: 320.7V
Average Cell Voltage: 3.341V
SOC BMS/Display: 5.5%/4.5%

Charge +28kWh:
Battery Voltage: 392.6V
Average Cell Voltage: 4.09V
SOC BMS/Display: 91.5%/96%

Charge +1.3kWh:
Battery Voltage: 396.8V
Average Cell Voltage: 4.133V
SOC BMS/Display: 95%/100%


Feel free to interpret those figures. I will get back to it a bit later.

 

[IonicEV]

Finally, I was able to get near EA chargers with 51% battery after driving almost 80 miles on highways.

So lets compare what happened on 50kW free charger and EA $$ charger:

*** Public Free DC 50kWh charger
75% @ 392V*120A=47kW
78% @ 395V*109A=43kW
81% @ 394V*88A=35kW
86% @ 395V*59A=23kW
90% @ 398V*59A=23kW
92% @ 399V*45A=18kW

*** EA 150kW $$ charger. Total paid: $5.19, Energy delivered: 10.57 kWh, Max charging rate: 58.96 kW, Time: 00:13:00
53% @ 55 kW
62% @ 57kW
67% @ 58kW
77% @ 44kW
79% @ 36kW
82% @ 27kW
83% ended

According to evbox.com:hyundai-ioniq-electric: The max. charging capacity this car can handle is 70 kW.

So there could be an improvement in the warm weather to reach max charging rate of 70 kW.

What is interesting from the EA session? I got 32% battery charge boost with 10.57kWh, I used heater for less than 10 mins at 39F ambient temperature and car was 4F below target temp, so effort was minimal to keep it warm (1kWh*10/60=0.17kWh). You can observe the heater draw with rate boost from 57kW to 58 kW between 62 to 67 % SOC. So, what is total battery capacity usable really? (10.57-.17)/(32)*100= 32.5kWh. Not far away from my 32kWh usable battery capacity estimates based on what my car delivers and my own cell voltage vs. SOC research on Ionic EV with Torque Pro app.

It is good to have EA charger when you are in emergency, but charging cost would be more than buying gas and driving ICE car, unless your car can pull 150kW, or at list > 100kW to break even. EA charging is cheaper than EVgo and no membership fees. I am spoiled with free L2 chargers at work and other places I visit. I do not even bother to use free DC chargers 24-50 kW available around.

I wonder if anybody was able to get close to 70kW charging rate?

 

[IonicEV]

Just publishing raw figures obtained at turtle and following with 100% charge:

Turtle:
Battery Voltage: 320.7V
Average Cell Voltage: 3.341V
SOC BMS/Display: 5.5%/4.5%

Charge +28kWh:
Battery Voltage: 392.6V
Average Cell Voltage: 4.09V
SOC BMS/Display: 91.5%/96%

Charge +1.3kWh:
Battery Voltage: 396.8V
Average Cell Voltage: 4.133V
SOC BMS/Display: 95%/100%


Feel free to interpret those figures. I will get back to it a bit later.

If we assume the linear BMS SOC between 91-100% and 4.2V highest voltage of Li-ion cell for quick estimate, we would arrive to the following conclusion:

1. BMS SOC accuracy?
4.133V - 4.09V = 0.043V = 3.5%, SOC = 1.3kWh => 0.0123V = 1% = 0.3714kWh
4.2V - 4.133V = 0.067V => 5.4% SOC => proves BMS SOC is accurate at the top.

2. What is total battery capacity?
The problem is we still do not know what is the total capacity of the battery is. It is clear though that BMS SOC is
accurate.

a. (1.3kWh) We could derive total capacity from the the fact 1% of capacity = 0.3714kWh. So total
capacity should be around: 37.14 kWh. Hm, there was some leaked info about 202? Ionic EV would have
38.5kWh. Is it possible they just were referring to the real capacity of the current EV battery?

b. (28kWh) 86% = 28kWh => would arrive at 32.56kWh.

3. Safety buffers?
The safe buffer is 5% at the top and not sure about bottom buffer, the 3.341V @ 5.5% BMS SOC is also strange
as it is a bit too low for Li-Ion cell in general.

All I know the nominal usable is ~32kWh at low temp. I have to wait till summer to figure it a bit better (I am expecting it to grow to 34kWh at 80F).

Interesting the fact at high SOC Display is more optimistic and at low SOC it flips around and becomes more pessimistic.

 

[IonicEV]

Found some info on minimal discharge voltage for Li-ion far all types of cells: 2.8–3.0V.

Assuming 3.0V is 0% SOC:

3.341V - 3.0V = 0.341V = 5.5% => ??? not enough data to approximate. I would need to take a few measurements to determine it. The V-to-SOC is similar to what is happening at top end. So we could try to use same figures from the top to apply to the bottom end
0.0123V = 1% = 0.3714kWh => 0.341V delta = 27.72% SOC = 10kWh. Something is wrong, but it looks like the bottom buffer is way bigger that top one for sure.

Lets calculate it different way: if we take into account Temperature compensation @ 32F, then 3.341V would put SOC = 15%, so BOM SOC 5.5% is not a true figure.
Knowing approximately usable capacity ~ 32kWh, then the bottom buffer = 4.8kWh - 1.76kWh(5.5% BMS) = 3kWh or close to ~ 10% SOC

And finally, we could deduct approximate total battery capacity: 3 + 32 + 1.6 = 35.6kWh.

Note: all my readings were around freezing point 32F, but if I recall correctly battery was ~48F. Total battery capacity decreases by ~3% at this temp. Also, I should have used 2.8V as it is true 0%. 3.0V I used just to be on the safe side and reduce any error in assumptions and I used lowered values of usable capacity to calculate buffers as we do not know the real capacity. The end result actually aligned very close with 37.14kWh calculated from 1.3kWh data sample.

Please, comment if you see any errors or wrong assumptions.

 

[IonicEV]

I would need more data to work on the bottom buffer as temp is critical to determine SOC and I did not capture enough samples to model.

If we assume the battery temp was 48F - it would put SOC at 91%, so the BMS SOH would be 9 - 5.5 = 4% below real capacity. Also, the total capacity is reduced by 2% at 48F, I think we could round it up to 5% for bottom buffer.

So if we assume 5% buffers on top and bottom, the real capacity would be: 1.6 + 32 +1.6 = 35.2kWh.

34kWh figure should not have much of a doubt at this point.

For example Tesla has ~ 2% buffer (top+bottom) total on 100kWh battery. This is why Tesla offers a SOC limiter option to prolong battery life.

 

[IonicEV]

I have experimentally changed some things, mainly increasing timeouts, to make the SoulEVSpy initialization sequence more like EVNotify. There is a debug build here: http://spjeldager.dk/SoulEVSpy-0.1.3-2913-debug.apk
If you test it in the Ioniq, could you let me know how it goes, and mail me .txt and .csv file examples at kia@spjeldager.dk ?

Initially it shown some car info (unknown car), bms info: %, Voltages, Temp, GPS Info; but never shown anything in cell map, DC-DC. After I switched to Torque Pro and EVNotify and then quit those and reactivated SoulEVSpy, it stopped receiving any data at all.

 

[IonicEV]

Thanks for giving it a try. Would you mind emailing the soulspy.log.2019XXXX.txt files recorded during the two attempts

On my latest attempt I got cell map to show up, so there is some progress. It took a bit to get it populated.

Interesting fact from it: The temp difference causes cell voltage variations, most likely relative to the physical locations of the cells inside the battery. This is good visualization and perfectly explains why the voltage differential exists between cells. Oh, the battery is perfect, kudos to LG & Hyundai!