Using clone ELM327 Bluetooth OBDII adapter with Leaf

[sorphin]

I'm charging again to 100% tonight (since i have some running around to do tomorrow).. I'll lower the Ahr again to see if that will put me at 100% cap finally. Then I'll post updates again (in the other thread I guess).

 

[TonyWilliams]

Whatever calculates the SOC% is also wrong, because the car will be at about 2.5% SOC for "dead".

SOC comes straight from the car so you will need to talk to Nissan and tell them they are wrong.

I don't think I'm on Nissan's speed dial for info, in either direction. Its obvious that the (')Gid count is way off for LeafSwe at dead (showing 8.2% when 2% to 2.5% Gid is typical). I have no idea why LeafSwe's car would show 7.6% SOC at that point. That value clearly does not reflect the remaining energy left in the battery, but the other data (cell voltages) suggests that it is indeed dead.

 

[surfingslovak]

Whatever calculates the SOC% is also wrong, because the car will be at about 2.5% SOC for "dead".

Yes, that would be one possible explanation. Phil has previously confirmed that turtle should occur at about 2 to 2.5% true SOC. Cliff's recently posted plot show that what we think is the SOC, could be just anther expression of stored energy.

 

[TonyWilliams]

Whatever calculates the SOC% is also wrong, because the car will be at about 2.5% SOC for "dead".

Yes, that would be one possible explanation. Phil has previously confirmed that turtle should occur at about 2 to 2.5% true SOC.

It's well established data, so there must be some reason that LeafSwe's data is so far off. I don't want folks thinking that 7.6-8.2% is normal at dead... it's not:


New, non degraded battery at 70F degrees temperature

SOC% - kWh usable ------ Gid --- Gid% --- Gid2% --- 2013 dash %

95.x% - 21.0 -- Full --- 281 --- 100% --- 100% ---- 100%
------- 10.5 -- Half --- 144 --- 51.2% --- 50% ---- 50%
-------- 3.1 -- LBW ----- 49 --- 17.4% -- 15.3% --- 17%
-------- 1.3 -- VLB ----- 24 ---- 8.6% --- 6.2% --- 8%
2.x% --- 0.0 -- Turtle --- 7 ---- 2.5% ---- 0% --- (---)

 

[surfingslovak]

It's well established data, so there must be some reason that LeafSwe's data is so far off. I don't want folks thinking that 7.6-8.2% is normal at dead... it's not

Yes, agreed and understood. Although it was discussed before, I believe that nobody has volunteered to run the car to turtle while using Jim's app. The first data point, if confirmed, would indicate that what we believe to be the true SOC could be just another expression of stored energy. Cliff posted couple of days ago, that there appears to be a linear 1:1 correlation between Gids and the "SOC" value. That should not be possible, based on what Phil has posted earlier. I think the good news is that we should be able to accurately compute Gids from the "SOC". The bad news is that the true SOC is likely remains elusive.

 

[Stoaty]

Is this what people want to see?

Nope. It's perfect the way it is. We already know that most batteries are going to be fairly well balanced (e.g., 30 mv or less, and we get that number anyway on Screen 1) and want to see the differences between the cells. Especially, I would want to identify low cells (so far, #35 is consistently a bit low on my Leaf, but not by all that much). It's easy to do this with the present implementation, would be a lot harder if the differences are minimized.

 

[surfingslovak]

The red is about 6600 data points, and green is a
linear fit to the data, using gids = soc * .31 - 15.
While the .31 will change with battery age, I do not know if
the -15 gids will change, or is fixed.
Still as a first guess, using this offset might improve simulated gids calculation.

Cliff, thanks so much for doing this. LeafSwe just corroborated what you saw couple of days ago. The "SOC" value is likely not the true SOC, but an expression of stored energy. Because of this, we should be able to calculate Gids from it accurately and it would also explain why the other formula did not work.

I tried to align the values against 80% and full charge, and this is what I believe would be a good start:

Raw Gids = "SOC" x 3.1 -17

Would it be possible for you to discharge the car over a large range of SOC values? Ideally from full to either LB and VLB. Possibly turtle (the last step could be accomplished by running the heater in the confines of your garage).

 

[surfingslovak]

It's well established data, so there must be some reason that LeafSwe's data is so far off.

Based on cliff's prior work, and by aligning this to screenshots posted by others, I believe that Gids can be calculated using the following formula. Should this be true or at least close enough, then LeafSwe was at 6.56 Gid when his LEAF hit turtle. If nothing else, it sounds like a plausible number to me.

Raw Gids = "SOC" x 3.1 -17

of in LeafSwe's case:

7.6 "SOC" x 3.1 -17 = 6.56 Gid

One possible complication is that folks with older packs still see 80% on an 80% charge and about 96% on a full charge. Could this be confirmed? If true, then we will need to multiply the Gid value calculated above by CAP value. What we could be looking at then is:

Raw Gids = "SOC" x 3.1 -17 x CAP

CAP is likely cut off at 100%, which means that the pack will never go over a certain Gid value (281). What was the highest "SOC" reported to date? Would anyone know?

 

[JeremyW]

Tested Leaf Battery app for the first time today, drove from 80% charge to Dead.
I'm a novice, but why SOC 7,6% when the car is dead ?!

I think the minimum cell voltage holds a clue: 2.692 volts. I bet low voltage cutoff for the leaf is 2.7 volts. Perhaps there is some left in the rest of the pack, but once the weakest cell hits LVC, it's game over.

Note the deviation between min and max cells in this case is around 300mv. I've observed ~100mv around 20 true gids. My pack spread is normally 25mv when not so low.

 

[Stoaty]

Based on cliff's prior work, and by aligning this to screenshots posted by others, I believe that Gids can be calculated using the following formula. Should this be true or at least close enough, then LeafSwe was at 6.56 Gid when his LEAF hit turtle. If nothing else, it sounds like a plausible number to me.

Raw Gids = "SOC" x 3.1 -17

of in LeafSwe's case:

7.6 "SOC" x 3.1 -17 = 6.56 Gid

One possible complication is that folks with older packs still see 80% on an 80% charge and about 96% on a full charge. Could this be confirmed? If true, then we will need to multiply the Gid value calculated above by CAP value. What we could be looking at then is:

Raw Gids = "SOC" x 3.1 -17 x CAP

CAP is likely cut off at 100%, which means that the pack will never go over a certain Gid value (281). What was the highest "SOC" reported to date? Would anyone know?

Doesn't work well for me, although not terribly far off. Pretty good at 75% SOC, but at 33.5% SOC (what I had tonite), the formula predicts 30.9% Gids, while my Gid meter read 27.7% Gids. The formula does about the same for me as Simulated Gids, although may be a tad better.

 

[Turbo3]

Whatever calculates the SOC% is also wrong, because the car will be at about 2.5% SOC for "dead".

SOC comes straight from the car so you will need to talk to Nissan and tell them they are wrong.

I don't think I'm on Nissan's speed dial for info, in either direction. Its obvious that the (')Gid count is way off for LeafSwe at dead (showing 8.2% when 2% to 2.5% Gid is typical). I have no idea why LeafSwe's car would show 7.6% SOC at that point. That value clearly does not reflect the remaining energy left in the battery, but the other data (cell voltages) suggests that it is indeed dead.

You do realize the Leaf is going to try to protect the lowest voltage cell regardless of the actual remaining energy availabe in all the cells. Turtle will not be at any fixed lower energy level. Once the weakest cell crosses the cutoff voltage that is it.

Just look at the difference between the max and min cells it is 307 mVolts. One or more very low cells forced turtle even at 7.6% SOC. I don't see anything wrong with that SOC causing turtle with that much cell difference.

 

[surfingslovak]

Doesn't work well for me, although not terribly far off. Pretty good at 75% SOC, but at 33.5% SOC (what I had tonite), the formula predicts 30.9% Gids, while my Gid meter read 27.7% Gids. The formula does about the same for me as Simulated Gids, although may be a tad better.

Thanks for doing that, Stoaty. If I calculated it correctly, then it's 78 Gid vs 87 Gid, which is a significant difference. Would you mind sharing your CAP value? Upon further reflection, I believe the "SOC" meter in 2013 LEAFs could be Gid-based also. Aligning the "SOC" value from Jim's app to the 2013 SOC dash gauge would be interesting. I think we might see a linear fit there also.

 

[surfingslovak]

You do realize the Leaf is going to try to protect the lowest voltage cell regardless of the actual remaining energy availabe in all the cells. Turtle will not be at any fixed lower energy level. Once the weakest cell crosses the cutoff voltage that is it.

Yes, of course. We had very consistent data from earlier LEAFs. I don't believe that any owner has seen turtle at 7.6% SOC. Personally, I would expect some variation, especially as the pack ages, but the data has been very consistent with new LEAFs. I have seen turtle anywhere between 5 and 8 Gid, if I'm recall correctly.

 

[garygid]

Suggestion:

The 0.7 volt range for the cell-pair bars is too large to show any useful
information. However, the rapidly changing auto-scale is almost (but not)
as bad.

I suggest a "fixed", but adaptive, range of around 100 mv, and 200 or
300 if necessary to show all the CP data. Just move the vertical axis origin
by perhaps 50 mv, but only when needed to show all the data.

Setting the base x-axis value at some multiple of 50 mv is probably
sufficient to maintain easy to read values on the axis.

This would be very similar (actually identical to) what you have done
with the adaptive scaling on the horizontal axis of the histogram graph.

In fact, I think that you could use exactly the same values?

 

[Turbo3]

You do realize the Leaf is going to try to protect the lowest voltage cell regardless of the actual remaining energy availabe in all the cells. Turtle will not be at any fixed lower energy level. Once the weakest cell crosses the cutoff voltage that is it.

Yes, of course. We had very consistent data from earlier LEAFs. I don't believe that any owner has seen turtle at 7.6% SOC. Personally, I would expect some variation, especially as the pack ages, but the data has been very consistent with new LEAFs. I have seen turtle anywhere between 5 and 8 Gid, if I'm recall correctly.

You really think Nissan is going to kill a cell (and therefore the pack) just to hit a lower gid number. Using gids for the LBW and VLBW is fine but Turtle must be based on the weakest cell just like charging is limited by the strongest cell. (By strongest I meant with the highest voltage while chargeing)

Cell life is the game.

 

[surfingslovak]

You do realize the Leaf is going to try to protect the lowest voltage cell regardless of the actual remaining energy availabe in all the cells. Turtle will not be at any fixed lower energy level. Once the weakest cell crosses the cutoff voltage that is it.

Yes, of course. We had very consistent data from earlier LEAFs. I don't believe that any owner has seen turtle at 7.6% SOC. Personally, I would expect some variation, especially as the pack ages, but the data has been very consistent with new LEAFs. I have seen turtle anywhere between 5 and 8 Gid, if I'm recall correctly.

You really think Nissan is going to kill a cell (and therefore the pack) just to hit a lower gid number. Using gids for the LBW and VLBW is fine but Turtle must be based on the weakest cell just like charging is limited by the strongest cell.

Cell life is the game.

Of course, I didn't say that at all. I'm just re-stating what we have observed out there in the field. This could be interpreted in different ways. One of them is that the pack is well-balanced or perhaps better managed than previously believed. Note that all of our cars were fairly new at the time, and the pack has not aged much. I wouldn't make any assumptions about LeafSwe's car, but it sounds like he could be from a Nordic country and the car is likely less than a year old. Personally, I find it highly doubtful that turtle mode would set in at 7.6% true SOC. That's contrary to pretty much everything we have heard from owners, and from Phil. Granted, he experimented on his own car, which was fairly new at the time.

 

[garygid]

Charging would be limited by the highest voltage cell, which
is probably not the strongest. It could be the cell with
the lowest capacity, which could also become the lowest
voltage cell while driving toward empty.

 

[Turbo3]

Suggestion:

The 0.7 volt range for the cell-pair bars is too large to show any useful
information. However, the rapidly changing auto-scale is almost (but not)
as bad.

I suggest a "fixed", but adaptive, range of around 100 mv, and 200 or
300 if necessary to show all the CP data. Just move the vertical axis origin
by perhaps 50 mv, but only when needed to show all the data.

Setting the base x-axis value at some multiple of 50 mv is probably
sufficient to maintain easy to read values on the axis.

This would be very similar (actually identical to) what you have done
with the adaptive scaling on the horizontal axis of the histogram graph.

In fact, I think that you could use exactly the same values?

Yes, this is the direction I am headed in just like Histograms and Temperature. I already have a variable defined to hold this temporary 'fixed' value.

 

[surfingslovak]

Charging would be limited by the highest voltage cell, which
is probably not the strongest. It could be the cell with
the lowest capacity, which could also become the lowest
voltage cell while driving toward empty.

Of course, Gary. Please have a look at LeafSwe's screen capture again. I believe that the pack voltage is about 307V and about 3.2V on cell level. I'm assuming that these are averages, and they would be in agreement with what I have observed with my car also. The pack appears to be in good health. Could you help me understand the numbers at the bottom of the screen? These appear to be cell voyages per Stoaty's description in the Wiki. However, I don't see how they would correlate to 307V pack voltage. Perhaps I'm missing something. I also wanted to take an opportunity to post what Phil said earlier on the topic:

The LEAF uses a sophisticated top-balance system with an array of MOSFETs that can bypass current around cells that are at a higher SOC than their neighbors, thus allowing those with lower SOC to "catch up" so that all cells have parity. This is all carefully controlled and monitored by the battery ECU located in the battery box.

 

[Turbo3]

Charging would be limited by the highest voltage cell, which
is probably not the strongest. It could be the cell with
the lowest capacity, which could also become the lowest
voltage cell while driving toward empty.

Of course, Gary. Please have a look at LeafSwe's screen capture again. I believe that the pack voltage is about 307V and about 3.2V on cell level. I'm assuming that these are averages, and they would be in agreement with what I have observed with my car also. The pack appears to be in good health. Could you help me understand the numbers at the bottom of the screen?

Wrong. The number (307) is the difference between the highest voltage cell and lowest voltage cell.

2.999 volts - 2.692 volts = 0.307 volts

This is a really big difference.

I am starting to think that people do not understand that you never want to discharge a battery below a certain level. People keep talking about Gids and SOC as if low cell voltage does not matter.

As long as the Pack has energy just keep on draining it out. I am done. Back to coding.

Please move this discusson somewhere else.