LEAF CANbus decoding. (Open discussion)

[JeremyW]

I spent some time last night trying to find a message indicating the number of regen bubbles to help shed some light on the regen reduction riddle. I failed to find anything that directly correlated 1:1 with regen bubbles but I think I found something better. CarCan msgId:260 d:1 appears to contain available regen. I put a watchpoint on it in my CANary for my drive this morning and saw that regen bubbles appear when this message reaches 6, 12, & 18 (I presume I would have gotten the last bubble if I was able to reach 24). There is hysteresis so the bubbles only disappear when it falls to the threshold below that (the 2nd bubble comes on with 12 and off with 6; similarly, the 3rd comes on with 18 and off with 12). Should be illuminating to plot this along with speed, SOC, and temperature throughout the last year. I hope to do that tonight.

On the drive home I was able to confirm that, indeed, the final (4th) regen bubble lights when 260:1 reaches 24. Furthermore, I was able to get the message to go all the way to 32 - equivalent to 1 1/3 more bubbles which the dash was unable to display. Would be interesting to see how high it can go for folks living near hills (requires long durations of regen to build up the regen bubbles).

Very cool. Seen something similar on the EV bus? Greg should add it to the LeafDD.

 

[TickTock]

Here are three plots of the exact same route to work in the morning on three separate dates. Blue is speed and red is available regen (bubbles times 6). All starting at full charge (~95%). SW update was on July 10 (just before the middle plot). Top of graph is 90mph for speed and 40 for the available regen. Each 6 steps in the available regen equates to one regen bubble on the dash. There is a clear trend of reduced regen while at speed going into fall and winter.

March 29, 2012, Pack Temp=31C:

August 08, 2013, Pack Temp=29C:

December 11, 2013, Pack Temp=22C:

 

[TickTock]

Two more. Here is the drive home on August 8, 2013. Pack Temperature was 37C. I reached 40 regen units (equivalent to almost 7 bubbles)! and 42kW max regen (post sw update but higher temperature and lower SOC than in the morning):

[Edit: added another graph]
Here's the drive home on December 11, 2013. Pack Temperature was 26C and I never got more than 12kW regen. However, unlike the August return trip, I charged at work so am starting again with a full charge and ending at 49% SOC.

 

[JeremyW]

Alright the first one to figure out how to manually turn on the pack heaters wins! :twisted:

From Page 344 of the "EVB" portion of the 2012 service manual....
With CONSULT
1. Select “ACTIVE TEST” of “HV BAT.”
2. Select “HEATER RELAY UNIT.” Touch “START."

It might just cycle the heater relay, or there might be a way to keep it on. Just getting to 20-25c should make a significant difference... but might screw up Stoaty's aging model...

Note this would only work obviously for those with battery heaters i.e. 2011 with "cold weather package", 2012's, and 2013's.

 

[JeremyW]

Here is the drive home on August 8, 2013. Pack Temperature was 37C. I reached 40 regen units (equivalent to almost 7 bubbles)! and 42kW max regen (post sw update but higher temperature and lower SOC than in the morning)

Wonder what the the battery voltage and other parameters (like Hx) were at when regening that hard!

 

[GregH]

Would be interesting to see this max regen figure graphed and scaled against regen current (ignoring drive current)

 

[FalconFour]

Would be interesting to see this max regen figure graphed and scaled against regen current (ignoring drive current)

^ Second this. I think there's definitely a bug in the battery controller available-regen algorithm... it simply doesn't make sense to have it keep cutting the available regen down every time power's taken back out of the pack. Doing that should open up more available regen, I'd think! :?

 

[DaveEV]

Would be interesting to see this max regen figure graphed and scaled against regen current (ignoring drive current)

Pack voltage would be interesting, too.

 

[TickTock]

Gary,
Do you know if anyone ever conclusively determined 1DA5&D6 to be RPM with 1LSB=0.5rpm? I see a few posts where that was speculated and it seems to have held up until now. However, I am starting to wonder if it is really motor volts. Volts and Speed do track for electric motors but I think it may be volts since when I multiply motor Amps (1DA3&D4) with this field I get a power number that very closely tracks the power in and out of the battery. Specifically, if I treat 1LSB of motorAmps = 0.5A (just like the packA) and 1LSB of 1DA5&D6 as .05V instead it tracks very well with a 90% efficiency. Maybe just coincidence, but the units do seem to line up with A & V.

 

[TickTock]

On further thought, RPM does makes sense, too. Using straight math taking the gear ratio and the radius from axle to pavement for my tires, I come up with a scalar of 1/227 to convert the 1DA treated as RPM_x2 to miles per hour. Empirically, I have established that a scalar of 1/220 results in a reading matching the Leaf's own speedometer. That's only a +3% error (suggesting speedometer reads slightly fast). So... maybe it really is RPM and the field we think is motor amps is actually torque? I know torque and current are proportional, I am just trying to establish the actual units of these fields.

 

[TickTock]

[moving from LeafApp thread]

The data I am displaying comes from a single CAN message 0x180. It has been labeled as Motor Amps but whether that is true or not only Nissan engineers knows for sure. This is all reverse engineered by looking at traces and see how the data changes. I am assuming the units are whole amps but they also could be 0.1 KW units as the max seems to approach 800 amps on my scale which if KW would be 80 KW.

Very good theory. I noticed that the 180 message is positive when driving in reverse while the 1DA messages are negative in this case. Both (180 & 1DA) go negative during regen. This supports your proposal that 0x180 is actually power.

It (CAN:180.23) is not power. However, I get very good tracking with the power from the battery when I multiply the CAN 180.23 with CAN 176.01 (after dividing by ~3.6) or CAN:180.23*CAN:176.23 divided by ~30. Below is the latter (although both 176 messages yield the same results with the right scalar).

 

[Turbo3]

If I divide 176.23 by 8 I get almost exactly 176.01. The plots virtually overlay each other.

I will update my test version of Leaf Spy Pro to capture 0x176 in addition to 0x180 and see what it looks like. What are your exact conversion constants as 3.6 and 30 are not 1:8 (close at 8.333).

I am also thinking that the sequence numbers at the end might be used to match up the pair samples (i.e. seq #3 180 goes with seq #3 176).

 

[TonyWilliams]

I thought I'd throw this out for the guys that are light years ahead of me in cracking the CAN bus. Do we already have a catalog of the charging messages from the BMS?

For instance, what message opens and closes the main contactor?

The thinking process is that a CHAdeMO receptacle could be bolted onto a non-CHAdeMO LEAF and with the correct programming, a relatively cheap upgrade could be made.

 

[TickTock]

I thought I'd throw this out for the guys that are light years ahead of me in cracking the CAN bus. Do we already have a catalog of the charging messages from the BMS?

For instance, what message opens and closes the main contactor?

The thinking process is that a CHAdeMO receptacle could be bolted onto a non-CHAdeMO LEAF and with the correct programming, a relatively cheap upgrade could be made.

Not much is known. User derdlimderdlim added some data on the QCbus section but we don't know much. I wonder, though, if you are over-thinking it. Maybe the QCcan interface is there even if the QC port was declined. If you make the power connection and the QCcan connection, maybe it would connect.

 

[TickTock]

If I divide 176.23 by 8 I get almost exactly 176.01. The plots virtually overlay each other.

I will update my test version of Leaf Spy Pro to capture 0x176 in addition to 0x180 and see what it looks like. What are your exact conversion constants as 3.6 and 30 are not 1:8 (close at 8.333).

I am also thinking that the sequence numbers at the end might be used to match up the pair samples (i.e. seq #3 180 goes with seq #3 176).

I don't use this presently so no exact conversions to share. On CANary, I currently only track the battery power. The ~3.6 and ~30 were hand tuned to get the graphs to visually line up (I was trying to see if there was a nice round number the scalar snapped to if I assume 90-95% efficiency). I didn't notice the 8:1 ratio for the two 176 fields so not surprising my values weren't exactly 8:1. Maybe 176.23 is volts*20 and 180 is amps*2? That put me below 90% efficiency from battery to motor, though, which I thought was too low.

 

[TonyWilliams]

I thought I'd throw this out for the guys that are light years ahead of me in cracking the CAN bus. Do we already have a catalog of the charging messages from the BMS?

For instance, what message opens and closes the main contactor?

The thinking process is that a CHAdeMO receptacle could be bolted onto a non-CHAdeMO LEAF and with the correct programming, a relatively cheap upgrade could be made.

Not much is known. User derdlimderdlim added some data on the QCbus section but we don't know much. I wonder, though, if you are over-thinking it. Maybe the QCcan interface is there even if the QC port was declined. If you make the power connection and the QCcan connection, maybe it would connect.

Good point. I guess I'll try the simple answer!

 

[TickTock]

If I divide 176.23 by 8 I get almost exactly 176.01. The plots virtually overlay each other.

I will update my test version of Leaf Spy Pro to capture 0x176 in addition to 0x180 and see what it looks like. What are your exact conversion constants as 3.6 and 30 are not 1:8 (close at 8.333).

I am also thinking that the sequence numbers at the end might be used to match up the pair samples (i.e. seq #3 180 goes with seq #3 176).

I would go with 29 (or 3.625). That is the value that gives me the same efficiency during regen as during drive (probably not a valid assumption but I'm hoping it is close). This corresponds to 95% efficiency at high power (>30kW) but only 85% efficiency at 10kW (horizontal constant power part of the curve below is ~10kW).

 

[TickTock]

I spent some time last night trying to find a message indicating the number of regen bubbles to help shed some light on the regen reduction riddle. I failed to find anything that directly correlated 1:1 with regen bubbles but I think I found something better. CarCan msgId:260 d:1 appears to contain available regen. I put a watchpoint on it in my CANary for my drive this morning and saw that regen bubbles appear when this message reaches 6, 12, & 18 (I presume I would have gotten the last bubble if I was able to reach 24). There is hysteresis so the bubbles only disappear when it falls to the threshold below that (the 2nd bubble comes on with 12 and off with 6; similarly, the 3rd comes on with 18 and off with 12). Should be illuminating to plot this along with speed, SOC, and temperature throughout the last year. I hope to do that tonight.

I can confirm that the units of message 260.1 is 1kW/lsb. This is power into the battery (as opposed to power out of the motor). So if you have 260.1=23, then you will have 3 or 4 bubbles and, at maximum regen, you will be putting 23 kW into the battery.

 

[JeremyW]

I wonder if this message is there while charging. If so it would be interesting to see it change during L2 or quick charging!

 

[TickTock]

I wonder if this message is there while charging. If so it would be interesting to see it change during L2 or quick charging!

CARCAN is not active during charging (unless you open a door or do something to wake it up). THis message is on the carcan.