Setting up Torque to show BMS data

[JejuSoul]

...It's obvious that the unknown data is 16 bits, but still nothing we can relate to...

Elmil, Thanks for the data, but I disagree with your conclusion from it. In terms of unknown1 being a 9 bit value not a 16 bit your data is the same as mine, and the same as SoulEV2016s posted above. Having data from 3 cars is stronger evidence than just from mine.

Firstly though I must shift the naming of the two unknowns in MY2016 cars. My car is a MY2015, I have the single Unknown1 value. I agree with you that the MY2016 cars added a new value before the original one.

002_OBC_Unknown2,Unknown2,2102,(i<8)+j,0,65536,?,794
002_OBC_Unknown1,Unknown1,2102,(k<8)+l,0,65536,?,794

All the data shows that the same higher bits are always set in unknown1. bits 10 and 13 are one. bits 11,12,14, 15 and 16 are zero
In unknown2 bit 13 is one. bits 10, 11,12,14, 15 and 16 are zero.
For me to drop the idea that this is a 9 bit value not 16 bit is for someone to show any value for the top 7 bits other than the above.

For the lower 9 bits. I have watched my car over 2 short charging sessions. The number varies from 245 to 265.
Your numbers posted vary from 174 to 306. Hence my idea that it is peak voltage seems to be wrong. I still think it is a voltage though.
Your quote above shows a single value for each session, not a range. I think if you look again you will see a range.

For now I think these numbers are 9 bit values. So I am adding the following code.
I will keep unknown1 as it is a 16bit integer, but also test the lower 9 bits as the code Maybe_Volts
Will continue to test them until someone shows this idea is false.

,For MY2015 OBCs
002_OBC_Maybe_Volts1,MaybeVolts1,2102,({i:0}*256)+j,0,511,V,794

,For MY2016 OBCs
002_OBC_Maybe_Volts2,MaybeVolts2,2102,({i:0}*256)+j,0,511,V,794
002_OBC_Maybe_Volts1,MaybeVolts1,2102,({k:0}*256)+l,0,511,V,794

 

[SoulEV2016]

Test in progress.

 

[SoulEV2016]

Good, very good (maybe_volt don't work at all on OBC2016).

Right-clic on the picture and then, use "display picture" to view at maximal resolution.

 

[SoulEV2016]

Damn ... heated steering wheel pump more than an heated seat = Kia, what it's that ! Not good ! :lol:
Many measures are an IMPACT value (they decrease over the time because of heat or regulation, except for lights )

Standby : 8A - 13,2v - 105,6w
no fan, no heat, no charger, no lights.

1x Heated seat (full power = 3 plots) : 66w
2x Heated seats (full power = 3 plots) : 119w

Heated parts are regulate according to the setting (see graphic for 2 or 1 plot setting).
No setting on steering wheel, you must shutdown alone if it's too hot.

Heated Steering Wheel : 79w
Assisted steering wheel (on 0 km/h = max torque) = 185w

DRL LED + rear LED = 13w
Night Light = 152w
HighBeam Light = 266w

Night Light + FOG Front = 266w
Night Light + FOG = 323w
Night Light + FOG + HighBeam = 452w

Front FOG only = 132w
FOG only = 158w

1 Window motor = 80w
2 Windows motors = 145w
Carefull, keep pushing the button of window don't stop the power on the motor at the end (power is always max after close is complet)

Inside lights (luxury model) = 27w
Inside + Mirrors lights (sunpanel) = 40w

Defrost rear window and 2x mirror = 264w

Regulate FAN = 6 levels (no heat, no AC) = from 40w to 300w

FULL FAN + HEAT (22°C) = 410w
Carefull, not all heat elements use the DC-DC converter.

AUTO mode + Night lights + 1x Heated Seat and Steering Wheel = 524w
(FULL FAN + HEAT + A/C)

 

[JejuSoul]

Good, very good (maybe_volt don't work at all on OBC2016).

That graph of low voltage power is good, really good.
I don't know why maybe_volt doesn't work for you. One of them is the same code as I am using.

I tested it again on a different charger. I still think it is a voltage. In this test it varied from 178 to 220. The upper bits were the same as before.
In the screenshot below maybe_volt is 212 and ChargeVolts is 219V

I think the 3 OBC temperatures are too high. We should subtract 28.
It makes the three OBC temps in the same range as the 3 VMCU temps and the LDC temp.
Perhaps someone with cold winter temps could test to see if the OBC temp is ever negative.
Hence I am changing these 3 OBC codes.

Edit: I was wrong. Do not subtract 28. Instead these numbers are signed. Thanks Elmil
Code below is corrected.

002_OBC Temp 1,OBC Temp 1,2102,Signed(S),-40,120,C,794
002_OBC Temp 2,OBC Temp 2,2102,Signed(T),-40,120,C,794
002_OBC Temp 3,OBC Temp 3,2102,Signed(U),-40,120,C,794

 

[Elmil]

All the data shows that the same higher bits are always set in unknown1. bits 10 and 13 are one. bits 11,12,14, 15 and 16 are zero
In unknown2 bit 13 is one. bits 10, 11,12,14, 15 and 16 are zero.
For me to drop the idea that this is a 9 bit value not 16 bit is for someone to show any value for the top 7 bits other than the above.

I should have backed this with more data, sorry about that.

When I did 30000km service this spring, they let me borrow a MY2016 demo car, and I still have the log files from that one.. :roll:
One interesting sequence is when I started the L2 charging, the logger happened to catch a moment when the A/C power was on, and the charging was just about to start.

The "unknown1" value showed 0x0DF2, and the output voltage showed 0x0DF9.
This just can't be a coincidence, and it suggests that the "unknown1" is a DC voltage in the same scale as the output voltage.

When the charge had started, the "unknown1" was jumping around at about 0x1280-0x1290. The "unknown2" did the same at 0x0FE8-0x1016. So, "unknown2" also uses the whole 16-bit word.

Translating the hex values to "HV battery" voltages, gives a working range for the "unknown1" of around 470-490V DC. This seems a bit odd, but here is my theory:

The unknown values (both) are not real measured voltages, but calculated reference values (from the converter's control loop?), perhaps only used for diagnostics. Interestingly enough, the "unknown1" might actually indicate that this OBC, in theory, could be able to charge a higher voltage battery. The "unknown1" could for example be the open circuit voltage for the converter.

Just speculations...

 

[Elmil]

I think the 3 OBC temperatures are too high. We should subtract 28.
It makes the three OBC temps in the same range as the 3 VMCU temps and the LDC temp.
Perhaps someone with cold winter temps could test to see if the OBC temp is ever negative.
Hence I am changing these 3 OBC codes.

002_OBC Temp 1,OBC Temp 1,2102,s-28,0,255,C,794
002_OBC Temp 2,OBC Temp 2,2102,t-28,0,255,C,794
002_OBC Temp 3,OBC Temp 3,2102,u-28,0,255,C,794

No, they are correct, and they do get negative.

 

[JejuSoul]

...The "unknown1" value showed 0x0DF2, and the output voltage showed 0x0DF9.
This just can't be a coincidence, and it suggests that the "unknown1" is a DC voltage in the same scale as the output voltage.
.. perhaps only used for diagnostics...
Just speculations...

I agree we are speculating. And I agree Kia has probably added an extra diagnostic value to the later OBCs.
The reason - 10% of the original OBCs were failing and Kia didn't know why.
There is a further difficulty to our testing. We are not testing the same thing.
You are probably using a V10 OBC. Your lease car was a V12
I am using V11. SoulEV2016 is using V12.
And the version we really want to test is V14 which none of us have.

Elmil you have evidence for the MY2016 that unknown1 may be a DC voltage.
Do you have any evidence for the MY2015?
I still think it is worth checking on my car to see if unknown1 is an AC voltage. I want to see what value I get if I charge on 110V
I am looking at the range of values I am seeing on 220V AC - unknown1 is 220 +/- 50
If they were the last 9 bits of a larger number the range should be distributed between 0 and 511

Elmil: can you explain this comment - "No, they are correct, and they do get negative."
If the original version is correct, we do not subtract anything - they are positive integers - they cannot be negative. ???

 

[SoulEV2016]

Update :

CSV v8 :
- add LDC v2 file (Elmil + JejuSoul) = values of the DC-DC converter (12v supply from 400v battery)
- modify OBC2015 & OBC2016 (JejuSoul) = maybe_volt values, in development.

Kia_Soul_EV_Battery_Cell_data.csv - Kia_Soul_EV_BMS_data.csv
Kia_Soul_EV_Extra_data_v3.csv - Kia_Soul_EV_Extra_gauges_v2.csv
Kia_Soul_EV_LDC_v2_data.csv - Kia_Soul_EV_OBC2015_v4_data.csv
Kia_Soul_EV_OBC2016_v3_data.csv - Kia_Soul_EV_TPMS_data.csv
Kia_Soul_EV_VMCU_data.csv

Mirror (All-in-ZIP) = http://www.4shared.com/zip/2rlgqn3Cba/Kia_Soul_EV_2014-2016_Torque_C.html



DASH v80 :
- delete Torque CMD and EST (in the Eco-Drive panel)
- replace by AUX Power & DC-DC Voltage (in the Eco-Drive panel) = can restrict the car consumption (heat-A/C) according to the target range
- add 3x OBC Temp & DC-DC Temp are in the CHECK panel (with round indicators of the drive-selector)

http://up.sur-la-toile.com/i1ixd
Mirror (ZIP) = http://www.4shared.com/zip/YlUS_1GOba/all_v80dash.html

 

[Elmil]

I agree we are speculating. And I agree Kia has probably added an extra diagnostic value to the later OBCs.
The reason - 10% of the original OBCs were failing and Kia didn't know why.
There is a further difficulty to our testing. We are not testing the same thing.
You are probably using a V10 OBC. Your lease car was a V12
I am using V11. SoulEV2016 is using V12.
And the version we really want to test is V14 which none of us have.

This should be at the OBC fail thread, but still, I compared my labels to those in that thread, and I noticed that the numbers probably being P/N are different on my OBC. The label on the back has these numbers:
36400-0E240
E010010401
MANDO V10.1E 15A01

All others I have seen, also newer versions, are:
36400-0E040
E010010400
And on "ricardar": MANDO V10.1R 14H11
So the version V10.1R came before V10.1E??
Also, the label on the Junktion box says OBC:V11.1E on my car...

Elmil you have evidence for the MY2016 that unknown1 may be a DC voltage.
Do you have any evidence for the MY2015?

Not yet found a similar startup state on my car as on the leas, but the data is very much the same during charging. Will try to catch it.

Elmil: can you explain this comment - "No, they are correct, and they do get negative."
If the original version is correct, we do not subtract anything - they are positive integers - they cannot be negative. ???

The data format is the same as for the battery modules. Signed byte (0x01= +1°C, 0xFF= -1°C)

 

[SoulEV2016]

usually, the "v" is for the board revision (electronic stuff, mosfet modification or mecanical correction).
and the followed letter (R) is the flash program revision (to adapt the offset of components for example).
and then, the rest is the Warehouse storage or World assignation of the Mando builder.

usually, you can have :
- v10.3E
- v11.8A
- v12.1D

... and i don't like when they change the MAIN board ... because is not normal for a charger.
flash program is OK because components evolve over the time (offset and protocol tweaks to answer/ask to the canbus protocol).

Bonus picture from my point-of-view. :mrgreen:

 

[Elmil]

Elmil you have evidence for the MY2016 that unknown1 may be a DC voltage.
Do you have any evidence for the MY2015?

Yes, just found it in an old log. "unknown1" must be the same in both MY2015 and MY2016.

 

[JejuSoul]

Thanks Elmil.
So the meaning for codes unknown1 and unknown2 are still unknown!
I will continue tracking it, because I assume it has some kind of diagnostic purpose.
But I won't add these codes to the files we release until we know what it is for.

The OBC temperature codes are known. Can someone who lives where it is cold, test that this works for negative numbers.

002_OBC Temp 1,OBC Temp 1,2102,Signed(S),-40,120,C,794
002_OBC Temp 2,OBC Temp 2,2102,Signed(T),-40,120,C,794
002_OBC Temp 3,OBC Temp 3,2102,Signed(U),-40,120,C,794

Have updated the first post in page 1 of this thread with the latest codes. (missing out the unknowns!)

 

[SiLiZiUMM]

Above : before
Below : after

It works well!

 

[SoulEV2016]

update v9 is planned :

- suppress 2 unknow values
- correction of OBC Temp



REPORTED at the end of the week (2016-12-25).

 

[JejuSoul]

Thanks for the picture. We are +19C today. Not usually this warm in winter though, but never as cold as Quebec.

A couple of requests. Can you post your latest deterioration numbers. I have your car from October '15. One of the first to be measured.
Also can you post your original OBC version. If you pop the plastic cover over the top of the junction box / EPCU you will see a sticker on the top. Your car is only a few weeks older than mine. But yours is a 2016 whereas mine is 2015. I have v11 of the OBC, which no one else seems to have.

 

[SiLiZiUMM]

Thanks for the picture. We are +19C today. Not usually this warm in winter though, but never as cold as Quebec.

Yesterday morning was very cold (colder than usual), -24C (-32 with wind chill)

Can you post your latest deterioration numbers. I have your car from October '15. One of the first to be measured.

Here's the raw data from last week:

>2101
7EA 10 0E 61 01 F0 00 00 00 
SEARCHING...
7EC 10 3D 61 01 FF FF FF FF 
7EA 10 0E 61 01 F0 00 00 00 
7EC 21 66 23 28 23 28 00 00 
7EA 21 FC 05 02 03 00 00 00 
7EC 22 00 0D F5 03 02 03 02 
7EA 22 00 00 00 00 00 00 00 
7EC 23 02 02 01 00 03 BA 02 
7EC 24 B9 43 00 00 78 00 01 
7EC 25 EE 01 00 02 02 46 00 
7EC 26 00 B9 32 00 00 BA C1 
7EC 27 00 52 06 28 41 00 02 
7EC 28 00 00 00 00 03 E8 00 

>2105
7EA 03 7F 21 12 
7EC 10 2C 61 05 FF FF FF FF 
7EC 21 00 00 00 00 00 02 02 
7EC 22 03 00 00 00 00 23 28 
7EC 23 23 28 00 01 55 02 02 
7EC 24 00 5A 1A 00 38 40 6A
7EC 25 00 00 00 00 00 00 00 
7EC 26 00 00 00 00 00 00 00

From Tyrel's Google spreadsheet, you'll have to use the "Pierre-Étienne #5" result in the BMC_Compare tab, as the others are from the other Soul EV I had as a loan. This result is from my own car.

Deterioration is now between 5.6% (cell #64) and 9% (cell #26).

Historical data from February 2016, just after they changed our OBC:

2101
7EC 10 3D 61 01 FF FF FF FF 
7EA 10 0E 61 01 F0 00 00 00 
7EC 21 B3 14 B9 23 28 00 00 
7EA 21 FC 05 FD 02 00 00 00 
7EC 22 00 0F 2B 0B 0A 0B 0A 
7EA 22 00 00 00 00 00 00 00 
7EC 23 0A 0A 0A 00 0C CA 5A 
7EC 24 CA 3D 00 00 7B 00 00 
7EC 25 90 B7 00 00 96 B2 00 
7EC 26 00 36 36 00 00 36 2B 
7EC 27 00 1E A4 4D 41 00 03 
7EC 28 00 00 00 00 03 E8 00 

>2105
7EA 03 7F 21 12 
7EC 10 2C 61 05 FF FF FF FF 
7EC 21 00 00 00 00 00 09 0A 
7EC 22 0B 00 00 00 00 14 B9 
7EC 23 23 28 00 01 50 0A 0A 
7EC 24 00 19 15 00 0A 03 BC
7EC 25 00 00 00 00 00 00 00 
7EC 26 00 00 00 00 00 00 00

Also can you post your original OBC version. If you pop the plastic cover over the top of the junction box / EPCU you will see a sticker on the top. Your car is only a few weeks older than mine. But yours is a 2016 whereas mine is 2015. I have v11 of the OBC, which no one else seems to have.

As I said it was replaced in February this year, I don't have the original make/version of it unfortunately. I can check the new one tomorrow if it could help.

What's your Soul EV # (the last digits of the VIN)? Mine is #4683.



EDIT : added 2101 data

 

[SiLiZiUMM]

SoulEV2016, I mirrorred your CSV files (v1 to v8) in a GitHub repo, available here:

https://github.com/pemessier/KiaSoulEV_Torque

I find it more convenient to manage files like this using source control vs having links on a forum. Also, 4shared is quite hard to use IMO :mrgreen:

I removed the "_vX" tokens from the filenames in the different versions in order to be able to diff the content of the files.

I'll add your dashboards as well. I'll add mine (for Nexus 5X) when it's ready (work in progress)!

Thanks for all this!

 

[JejuSoul]

SiLiZiUMM: Thanks for the data. Could you post the 2101 data as well. What's the odometer reading?

I had already started a github repo for the Torque Pro codes - https://github.com/JejuSoul/OBD-PID-for-HMC-EV
I am also using the zip file and the Dash displays built by SoulEV2016.
The intention is to create a plugin that will automatically install / update the codes onto the phone.

My VIN is #4294 - So it is only a few weeks earlier than yours. (I guess 2 or 3 weeks)
The sticker with your original OBC version number will be on the top of the EPCU. Like mine below.
Your new OBC sticker is at the back, on the new OBC.

 

[SoulEV2016]