Better (reverse) SOC meter already in the car?

[TonyWilliams]

The temperature gauge had five bars, which means that its large thermal mass kept the pack warm. The Leaf is usually parked in a covered garage structure with about 55 F ambient temperature and the car wasn't exposed to the elements for more than four hours.

I suspect that in the winter, the biggest single "bang for our buck" will be preheating the battery, since I suspect that it will hold a significant portion of that pre-heat for a LONG time.

Otherwise, freeze your butt off with the heater off !!!

 

[91040]

Good work 91040 on mapping the data. I have a question on the charge timer: in your data set there's no more increment after 21hr, but in Slovak's test he was able to run it down to 25hr charge time. Any explanation for the inconsistency? I'm very curious about the reading on the low range.

Also, what's the correlation between gid and the on-board SoC bars?

again, thanks for sharing the data.

My car was delivered in May 2011 with the post-A/C problem firmware (new bars). surfingslovak may have newer firmware and/or a 2012 model.

Check tonywilliams chart for correlation between Gids and bars. http://www.mynissanleaf.com/viewtopic.php?f=31&t=4295&hilit=range+thumb" onclick="window.open(this.href);return false;

 

[lukati]

hat we need is another volunteer to see if the ~4.4 Gids/hour value holds true for other cars.

You meant 4.4% per hour, right? The gid value is around 11 or 12. Nice collaboration otherwise, this is really fun. Good luck with your new Leaf tomorrow! I received slightly diverging data the other night. I wish I had Gary's SOC meter, but perhaps I can borrow one locally for a test drive.

I will be using this SOC meter on my trip home from the dealer tomorrow. My first electric trip in the Chicago cold will be about 65 miles. Wish me luck!

Good luck. I think that you are well prepared for your trip. Preheat while plugged in before leaving the dealer !!!!

4.4% per hour, yes. Thanks. I corrected the post.

Thanks for the good wishes. Everything went really well today. Dressed warm, preheated the cabin at the dealer, and used the seat heater. Using cruise control at 40-50 mph at ~32F I averaged 4.9 miles/kWh over 69 miles and had a SOC of 25% (17 hours), three bars, and 27 miles in the DTE display when I got home. Range anxiety is for the birds.

 

[surfingslovak]

Good work 91040 on mapping the data. I have a question on the charge timer: in your data set there's no more increment after 21hr, but in Slovak's test he was able to run it down to 25hr charge time. Any explanation for the inconsistency? I'm very curious about the reading on the low range.

Also, what's the correlation between gid and the on-board SoC bars?

again, thanks for sharing the data.

My car was delivered in May 2011 with the post-A/C problem firmware (new bars). surfingslovak may have newer firmware and/or a 2012 model.

I have a new datapoint: turtle mode at 28 hours. I'm on firmware 3NA0B, which implies MY 2011 vehicle. This was referenced in one of my earlier posts as well. It looks like we don't have a good model for the behavior of the 120V charging time display yet. I haven't reviewed my trip notes in detail, but I have a hunch that the delta between low battery warning and turtle is four hours on this display. So even though the absolute readout might vary, perhaps it can still be used to gauge onset of power limited mode.

The 240V display actually seems to be more consistent. Although both Tony Williams and I have managed to bring it down to 11 hours on two separate occasions, it usually shows 7:30 hours just before I hit turtle.

 

[yoyofella]

I have a new datapoint: turtle mode at 28 hours. I'm on firmware 3NA0B, which implies MY 2011 vehicle. This was referenced in one of my earlier posts as well. It looks like we don't have a good model for the behavior of the 120V charging time display yet. I haven't reviewed my trip notes in detail, but I have a hunch that the delta between low battery warning and turtle is four hours on this display. So even though the absolute readout might vary, perhaps it can still be used to gauge onset of power limited mode.

Interesting data again Slovak. This is really weird, 28 hr for 120V seems excessive, so I guess there's not a very good correlation between charge timer and SoC at the low end, where it's probably more useful for people.

I wonder if there's another variable used in addition to SoC to calculate the charge time. Temperature could be a possibility, was this done in extreme temperature?

 

[thankyouOB]

have you adventurers developed a consensus on what the charge meter says when the car is at turtle?
are the safest values to use for 100%?
24 hours on 120
and
7.5 hours on 240?

By safest, I mean for those of us who dont Tony-Williams it with abandon?

and do you think the 120 value is more precise?
I note from my driving that at the outset of the use of a charged battery (even 80% full) the 240 measure goes down in 10-minute increments, while the 120 goes down in 30-minute increments.

Right now, after reading all the above posts, the general consensus I get is that the
to-charge 80% values at turtle are:
17 for 120
and WHAT??? for 240.

two questions on 80% charge values:
-is that correct 17 hours on 120
-do the values change if you have charged to 100% using the default rather than using the timer at 100% or 80%?

 

[thankyouOB]

?

 

[surfingslovak]

have you adventurers developed a consensus on what the charge meter says when the car is at turtle?
are the safest values to use for 100%?
24 hours on 120
and
7.5 hours on 240?

I have more data points now, and a theory. I'd say that we are getting closer to a usable solution. We may need more people to try this at some point or another, to confirm accuracy of this approach.

Bear with me for a moment, it's a bit complicated, but here is a brief explanation of how the reverse SOC could work. We are nearly certain that the hourly increments (or decrements) on both the 120V and the 240V charging gauges are linear and they represent a constant amount of energy. It's roughly around 0.8hWh for the 120V display and about 2.8kWh for the 240V display.

What's variable is the point of reference. What do I mean by that? Well, the charging time display seems to adjust the distance to T0 whenever you plugin the car. This means that if I plugin the car into a trickle charger for 10 minutes, the board computer will try to recalculate distance to 0 hours charge time, and mess up the scale. This is how I likely ended up with 28 hours on the 120V display at one point and 11 hours on the 240V display on another.

The obvious workaround is to charge to full on a trickle charger to reset the gauge, and then only use 240V for opportunity charging. I'd ignore the 240V display at that point, because it will get out of whack with frequent partial charges. The 120V on the other hand, should remain accurate, so long you don't use the trickle charger for partial fill-ups. I hope this makes sense so far.

I believe that the car will go into turtle mode once you reached 26 hours on the 120V display. I'm in the process of confirming that. Filled up to 100% via trickle charging couple of days ago and the Leaf will be near empty today or tomorrow. I would like to repeat this with an 80% charge, and I believe that turtle mode will engage at 21 or 22 hours in that mode.

Anyway, that's all I have. I've been diligently gathering data, but I didn't have time to share it all due to other commitments. Thank you for keeping the thread alive, and if you wanted to help with the validation process, please chime in and post any results and observations you might have.

Best,
George

 

[thankyouOB]

have you adventurers developed a consensus on what the charge meter says when the car is at turtle?
are the safest values to use for 100%?
24 hours on 120
and
7.5 hours on 240?

I have more data points now, and a theory. I'd say that we are getting closer to a usable solution. We may need more people to try this at some point or another, to confirm accuracy of this approach.

Bear with me for a moment, it's a bit complicated, but here is a brief explanation of how the reverse SOC could work. We are nearly certain that the hourly increments (or decrements) on both the 120V and the 240V charging gauges are linear and they represent a constant amount of energy. It's roughly around 0.8hWh for the 120V display and about 2.8kWh for the 240V display.

What's variable is the point of reference. What do I mean by that? Well, the charging time display seems to adjust the distance to T0 whenever you plugin the car. This means that if I plugin the car into a trickle charger for 10 minutes, the board computer will try to recalculate distance to 0 hours charge time, and mess up the scale. This is how I likely ended up with 28 hours on the 120V display at one point and 11 hours on the 240V display on another.

The obvious workaround is to charge to full on a trickle charger to reset the gauge, and then only use 240V for opportunity charging. I'd ignore the 240V display at that point, because it will get out of whack with frequent partial charges. The 120V on the other hand, should remain accurate, so long you don't use the trickle charger for partial fill-ups. I hope this makes sense so far.

I believe that the car will go into turtle mode once you reached 26 hours on the 120V display. I'm in the process of confirming that. Filled up to 100% via trickle charging couple of days ago and the Leaf will be near empty today or tomorrow. I would like to repeat this with an 80% charge, and I believe that turtle mode will engage at 21 or 22 hours in that mode.

Anyway, that's all I have. I've been diligently gathering data, but I didn't have time to share it all due to other commitments. Thank you for keeping the thread alive, and if you wanted to help with the validation process, please chime in and post any results and observations you might have.

Best,
George

THANK YOU!!!
this is great and I will add what I can.
I do trickle charge to full at work some days, though the earlier part of the charge is 240.
I am doing that today, so I will be able to make some observations.

 

[thankyouOB]

I talk a walk and a swim, and was thinking about what you were saying: the 120 to-charge meter is NOT useful or reliable unless you have fully charged on 120 to full?

that seems to negate it as a tool.

 

[evnow]

I think the more we dig, the more clear it gets why Nissan didn't want to give a numerical SOC meter.

But, their smoothed post OCV data (our Gids) are pretty darn good for an SOC meter. Pretty darn consistent with the Battery Warnings, and Turtle.

Actually the "gids" are not linear - atleast near the top (and most probably near the bottom). I had a thread about it sometime back. That probably reflects the difficulty in getting a good measure of SOC near the top & bottom.

 

[surfingslovak]

the 120 to-charge meter is NOT useful or reliable unless you have fully charged on 120 to full? that seems to negate it as a tool.

Yes, I share this concern, and I DO think that there is a valid workaround. Charge the car to either full or to 80% using the 120V EVSE to recalibrate the display. Then don't use the trickle charge for partial fill-ups, use 240V for that. The 120V display should remain accurate, it won't recalibrate its reference point until and unless you use the trickle charger again.

I would like to validate that this works and that we understand the somewhat quirky charging display behavior. If you had time and inclination to help with this, that would be great. Otherwise, I expect to have something to share with you within a week or two.

 

[lukati]

the 120 to-charge meter is NOT useful or reliable unless you have fully charged on 120 to full? that seems to negate it as a tool.

Yes, I share this concern, and I DO think that there is a valid workaround. Charge the car to either full or to 80% using the 120V EVSE to recalibrate the display. Then don't use the trickle charge for partial fill-ups, use 240V for that. The 120V display should remain accurate, it won't recalibrate its reference point until and unless you use the trickle charger again.

I would like to validate that this works and that we understand the somewhat quirky charging display behavior. If you had time and inclination to help with this, that would be great. Otherwise, I expect to have something to share with you within a week or two.

In an ideal world, the car would know its current SOC and the rate by which it can increase the SOC using either 120V or 240V. Based on this information it could calculate at any time precisely how long the charging would take in either case. You seem to say that this is not what is actually happening, that in effect the car has trouble extrapolating from one voltage to the other and thus tries to keep them separate. When you charge at 120V, the car knows precisely how much time it took to reach 100% (or 80%) and thus can accurately predict how long it will take to reach the same charge state again using 120V. It can only guess the time it would take using 240V. The questions is of course whether it keeps the two voltages completely separate or whether, probably more likely, it re-calibrates both timers (one accurate and one guessed) whenever you charge to a reference point (either 80 or 100%).

 

[thankyouOB]

the 120 to-charge meter is NOT useful or reliable unless you have fully charged on 120 to full? that seems to negate it as a tool.

Yes, I share this concern, and I DO think that there is a valid workaround. Charge the car to either full or to 80% using the 120V EVSE to recalibrate the display. Then don't use the trickle charge for partial fill-ups, use 240V for that. The 120V display should remain accurate, it won't recalibrate its reference point until and unless you use the trickle charger again.

I would like to validate that this works and that we understand the somewhat quirky charging display behavior. If you had time and inclination to help with this, that would be great. Otherwise, I expect to have something to share with you within a week or two.

give me a clearer idea of the charging protocol.
-can i charge from 50% to full on 120?
-does it matter if I chargedon 240 before driving the car and parking it at 50% before charging to full on 120 or whether the 120- charge is in two or three separate charges separated by driving?

 

[abasile]

Running the climate control while charging will also throw off your numbers. While at ~70% SOC and charging/preheating, the car indicated "6 hours" to 100% on 240 V. I seem to recall that the time estimates were skewed until I plugged in the car again, without climate control.

 

[surfingslovak]

In an ideal world, the car would know its current SOC and the rate by which it can increase the SOC using either 120V or 240V. Based on this information it could calculate at any time precisely how long the charging would take in either case. You seem to say that this is not what is actually happening, that in effect the car has trouble extrapolating from one voltage to the other and thus tries to keep them separate.

Yes, that's my understanding as well. I made several observations, which lead me to believe that the algorithm Nissan uses to recalculate the remaining charge time does not work very well. The fact that turtle mode did not correlate with an absolute number on the charging time display, and varied between 24 to 28 hours during my test drives, is likely the direct result of this.

give me a clearer idea of the charging protocol.
-can i charge from 50% to full on 120?
-does it matter if I chargedon 240 before driving the car and parking it at 50% before charging to full on 120 or whether the 120- charge is in two or three separate charges separated by driving

I believe that it's enough if you complete a 80% or 100% charge using the trickle charger. It should not be necessary to run the Leaf down to turtle before charging it.

Once the 120V charging time display has the correct T0 reference for either 80% or 100%, it should work accurately. In my experience, the Leaf will not attempt to recalculate the reference point for the 120V display if you keep charging it with 240V, and that gauge should then provide readouts accurate enough to calculate the SOC.

Obviously, I didn't have time to validate this fully, but the current experiment I'm running (charge to 100% using 120V and run down to turtle) is looking promising. I'm getting readouts that make sense to me. I'll post here tomorrow or on Friday to let you know how it all ended.

 

[thankyouOB]

thank you, surfingslovak.
what does T0 mean in this phrase?

Once the 120V charging time display has the correct T0 reference for either 80% or 100%

 

[surfingslovak]

what does T0 mean in this phrase?

Sure thing, I hope we can figure this out. T0 means zero hours to charge, which is initially represented by dashes on the display after a full charge (-:--). I think it works the same way when the 80% timer is set.

 

[TonyWilliams]

Actually the "gids" are not linear - at least near the top (and most probably near the bottom). I had a thread about it sometime back. That probably reflects the difficulty in getting a good measure of SOC near the top & bottom.

Agreed, it's not linear. My statement is the Gid data is perfect (for me) for an SOC meter. Just a teeny, tiny amount of math is needed on the top and the bottom to make it linear.

 

[surfingslovak]

Actually the "gids" are not linear - at least near the top (and most probably near the bottom). I had a thread about it sometime back. That probably reflects the difficulty in getting a good measure of SOC near the top & bottom.

Agreed, it's not linear. My statement is the Gid data is perfect (for me) for an SOC meter. Just a teeny, tiny amount of math is needed on the top and the bottom to make it linear.

+1

I continue to gather data, and I'm quite impressed with the accuracy this method provides when trying to predict the remaining available energy, especially at a low state of charge. It's very good, but you are correct, the reverse SOC can in the best case be only be as accurate as a gid-o-meter would be.