SOC data: 281-based, New-Bars

[TonyWilliams]

Does the LEAF Service Manual talk about SOC "mapping tables"?

With a constant load, according to the service manual, the SOC bars are not linear. Close, but no cigar:

Bar..Min...Delta
12…230…n/a
11…200…30
10…180…20
..9…160…20
..8…130…30
..7…110…20
..6…..90…20
..5…..70…20
..4…..40…30
..3…..20…20

 

[evnow]

Does the LEAF Service Manual talk about SOC "mapping tables"?

No - my guess is Nissan would have a mapping table to translate the voltage under various temperatures to SOC. They would also monitor current in & out. Some combination of these 2 would give them SOC.

Also, the areas of the SOC vs Voltage curve where voltage goes down faster (i.e. a curve that is less flat), it is easier to get more accurate SOC. It is in the falt area that it is difficult to estimate the SOC using the voltage.

 

[SparkyEV]

With a constant load, according to the service manual, the SOC bars are not linear. Close, but no cigar:

Looks pretty linear to me, within quantization error:

 

[garygid]

These Nissan Li battery cells do not have a "flat spot" in their voltage/discharge curve.

I believe that they have a fairly constant slope in the mid-range.

 

[garygid]

The 230 in this Nissan "minutes to 2-bars" discharge test gives numbers strangely in the range of the "SOC" raw data, especally when you add another 20 for each of the two remaining (old) bars, and perhaps 11 more for zero-bar turtle mode: 230 + 40 + 11 = 281.

Remember, these would be Old-Bars.

 

[garygid]

The discharge data "approx" times (rounded to 10 minutes?) are, most likely, really about 23 minutes per Old-Bar.

Does not tell us much when we realize that the loss of bars is controlled by some function of some "SOC-type" value.

 

[evnow]

I think we should do more tests in the first bar to figure out what is happening. I remember someone writing about how the miles in the top bar increased after a couple of full charges from 3 to 6 miles.

May be boomer's cells are not balanced and thus discharge rather quickly at the top ?

 

[TomT]

I consistently get about one half the miles from the top bar that I get from the 11th and remaining bars... It's been that way since new.

I think we should do more tests in the first bar to figure out what is happening. I remember someone writing about how the miles in the top bar increased after a couple of full charges from 3 to 6 miles.

May be boomer's cells are not balanced and thus discharge rather quickly at the top ?

 

[evnow]

I consistently get about one half the miles from the top bar that I get from the 11th and remaining bars... It's been that way since new.

May be we should do some remapping. Linearity is important for any SOC "gauge".

Let us consider 270 to be the max - and the top 20 units should be translated to 10 units ...

 

[davewill]

I consistently get about one half the miles from the top bar that I get from the 11th and remaining bars... It's been that way since new.

May be we should do some remapping. Linearity is important for any SOC "gauge".

Let us consider 270 to be the max - and the top 20 units should be translated to 10 units ...

Well, the obvious thing to do is to run the exact discharge regimen from the service manual, and log our SOC number vs. time and see just how linear, or not, it turns out to be. We would then have the data necessary to correct it, if needed. Who wants to leave their car running for several hours in the name of science?

 

[Boomer23]

I consistently get about one half the miles from the top bar that I get from the 11th and remaining bars... It's been that way since new.

May be we should do some remapping. Linearity is important for any SOC "gauge".

Let us consider 270 to be the max - and the top 20 units should be translated to 10 units ...

Well, the obvious thing to do is to run the exact discharge regimen from the service manual, and log our SOC number vs. time and see just how linear, or not, it turns out to be. We would then have the data necessary to correct it, if needed. Who wants to leave their car running for several hours in the name of science?

I can do that. I have solar power to burn.

How do you guys want me to collect the data? Just SOC% and SOC# at each bar drop, time from start of study? Do you want Guessometer values as well?

 

[davewill]

...How do you guys want me to collect the data? Just SOC% and SOC# at each bar drop, time from start of study? Do you want Guessometer values as well?

The guys with the hardware may have more ideas, but I would connect a laptop and log ALL the CAN data during the test. That way it can be analyzed and reanalyzed as we learn more about what's in there.

 

[TonyWilliams]

A dyno test at 60mph from 100% to Turtle would be a worthwhile test, but of course, would cost a couple dollars [edit: just called the local shop in Miramar, $150/hr]

If we can plot how many miles per "281 SOC" unit, we might find a trend that might point us somewhere. At the very least, you could write a program to correct for non-linear parts of the discharge at what should be an absolutely sterilized steady "real world" state.

Assuming we get 80 miles, that average would be 3.5125 units per mile. Then, a correction for each mile that is above or below the 3.5125 value to hopefully achieve a mostly realistic SOC gauge with the data we know today.

 

[evnow]

Well, the obvious thing to do is to run the exact discharge regimen from the service manual, and log our SOC number vs. time and see just how linear, or not, it turns out to be. We would then have the data necessary to correct it, if needed. Who wants to leave their car running for several hours in the name of science?

Why do you think the usage will be linear ? I'm sure in the beginning the heater will consume more heat than after a steady state is reached, for eg.

BTW, boomer, were you running AC in that test run ?

 

[TonyWilliams]

Well, the obvious thing to do is to run the exact discharge regimen from the service manual

Why do you think the usage will be linear ? I'm sure in the beginning the heater will consume more heat than after a steady state is reached, for eg.

I agree. That time thing even says "approximate". Hardly what the goal would be for exacting data, but cheap and easy to do, and perhaps instrumental.

 

[DaveEV]

Well, the obvious thing to do is to run the exact discharge regimen from the service manual, and log our SOC number vs. time and see just how linear, or not, it turns out to be. We would then have the data necessary to correct it, if needed. Who wants to leave their car running for several hours in the name of science?

Why do you think the usage will be linear ? I'm sure in the beginning the heater will consume more heat than after a steady state is reached, for eg.

That's why the instructions say to set the temperature to max, fan speed to max and to leave the windows open - shouldn't take long for a steady state to be reached and wouldn't be surprised if the heater kept pumping at max power. Pretty easy to test if this is an issue or not.

 

[Boomer23]

[

BTW, boomer, were you running AC in that test run ?

Yes.

 

[Boomer23]

A surprising and somewhat troubling reading this morning. I had charged back up to 100% last night, after yesterday's run. This morning, the LEAF dash showed 12 bars, but I plugged in the CAN device and the reading was only 95% SOC. I didn't see the SOC # reading, sorry. I had the LEAF timer set to start at midnight and NO end timer set, so It didn't shut off prematurely due to a timer. I did go back and check my TED, and it showed that the charging stopped at 05:41 with a proper "jagged" ramping down of power use, which looks typical of any of my 100% charging sessions.

I drove the car a while with the SOC gauge connected, and this 95% reading was not an artifact. The reading was solid and diminished from there as I drove.

Any ideas why I didn't get a 100% SOC reading this morning?

 

[garygid]

Still with Old-Bar firmware, I just "charged to 80%" with always-active timers.

The SOC "should" be the same between old and new bars, but my
281-based SOC-Meter reads 82.2% => raw 231 ... (with 281 = 100%).

So, if the LEAF thinks that 231 is 80%, that would mean that
the LEAF is using about 289 as 100% ... that would be strange.

Maybe getting a "full" charge to 281 or 278 is as full as our
already-diminished battery can go?

Or, some of the equalization eats away at the top percent or so?

So, maybe the car charges to "About 80%", to match everyting else
that is Nissan "approximate"?

 

[aqn]

Has anybody gathered data across many many runs and look for a trend or average?