Improving GOM estimate?

My 2022 40kWh leaf has a good GOM - very linear. My short commute always costs me about 3% indicated battery charge each way.

We just bought a 2016 24kWh leaf for my wife and the GOM drains nonlinearly. It falls quickly in the upper percentages and slowly at the bottom. For example, if it's at 80% it will drop to 79% after just 0.5km, at GOM readings of 30-70% it will drop about 1% per km, and at the bottom, if it's at 20% GOM it might stay stuck there for 1.5km or more. It's not from driving conditions because I've observed the same behaviour over many months and a wide variety of roads.

So, question 1 - is there any way to improve the linearity of the GOM?

Question 2 - I've heard that draining the car to turtle mode might help recalibrate the GOM at the bottom. In this case, would I have to actually drive the car or could I drain it in the garage using the A/C? According to leafspy, with a GOM reading of 20% I have 3.5kWh remaining in the battery. With the AC set at max temp drawing 1.5kW, that's under 3 hours to totally drain the battery, and I can do it in the comfort of a garage with a charge point instead of risking getting stranded on the road. Would this idea work?

Seems reasonable to me but I think it will unnecessarily stress the battery to drain all the way dead for calibration. Down to VLBW should be fine. Also assuming you have a heat-pump you'll get higher drain rate by putting into defrost mode, as this will force the resistive heat on, along with the heat-pump. Also, keep in mind that the vehicle "rebalances" the modules at the high SOC%, and this can take some time even after the main charging has finished, so leave the car plugged in for a few hours extra to allow these small current-shunt events to complete.

As far as GOM I rarely use it but when I do I find it helpful to use it as 2nd-order information. Specifically I keep a running tab on the difference between the range estimate and the distance left to my destination. And then based on the trend of THAT difference, I can adjust my driving as needed to keep a reasonable buffer. Accounting for terrain, wind and weather along the way is something that comes with experience.

Nubo said:

Seems reasonable to me but I think it will unnecessarily stress the battery to drain all the way dead for calibration. Down to VLBW should be fine. Also assuming you have a heat-pump you'll get higher drain rate by putting into defrost mode, as this will force the resistive heat on, along with the heat-pump. Also, keep in mind that the vehicle "rebalances" the modules at the high SOC%, and this can take some time even after the main charging has finished, so leave the car plugged in for a few hours extra to allow these small current-shunt events to complete.

As far as GOM I rarely use it but when I do I find it helpful to use it as 2nd-order information. Specifically I keep a running tab on the difference between the range estimate and the distance left to my destination. And then based on the trend of THAT difference, I can adjust my driving as needed to keep a reasonable buffer. Accounting for terrain, wind and weather along the way is something that comes with experience.

Click to expand...

What would be the cell voltage at cutoff? I believe damage only starts at around 3.0V but with a bit of extrapolation I'd say the car cuts out quite a bit before that, I think not less than 3.2.

Would this low-end calibration make the GOM more linear? Do I have to do a full charge? Usually I give it 4 hours of timed charging when it gets to 30%, so my battery spends its life oscillating between 30 and 70%. I live on an island only 33km (17 miles) long so half a charge on the 24kWh battery is enough for 3 days of commuting. We want to keep this battery as healthy as possible for as long as possible.

LMF5000 said:

What would be the cell voltage at cutoff? I believe damage only starts at around 3.0V but with a bit of extrapolation I'd say the car cuts out quite a bit before that, I think not less than 3.2.

Would this low-end calibration make the GOM more linear? Do I have to do a full charge? Usually I give it 4 hours of timed charging when it gets to 30%, so my battery spends its life oscillating between 30 and 70%. I live on an island only 33km (17 miles) long so half a charge on the 24kWh battery is enough for 3 days of commuting. We want to keep this battery as healthy as possible for as long as possible.

Click to expand...

I do a complete drain of my 2020+ pack a few times a year now, but mainly to get the cells as balanced as possible. It does help the range estimates as a bonus.

The lowest voltage before damage happens is 2.500V, the highest being 4.299.

If my memory serves, for the Gen1, turtle mode happens around 3.0V or 0.5 kWh of capacity remaining if all cells are still above 3.0V. Total shutdown happens when any cell reaches 2.500V or lower. The reason I use all those zeros is because these cells are measured down the thousandths of a volt, so if a cell was 2.599, it won't trigger a shutdown. If no cells ever get below 2.500, then total pack shutdown still happens at 0.3 kWh of capacity remaining. Probably as a safe mode just in case.

Will it make the GOM more linear? Not on a Gen1 Leaf. The GOM still way over-estimates range when the battery is full, unless you can really stick to driving 5.0 m/kWh somehow. When using Climate Control, unless you are driving around at less than 30 mph, usually isn't possible in real world driving conditions.

Your plan for long battery life should work if you never need more than 80% of the battery to complete your driving task.

The GOM does not change linearly because its not linear. The LEAF builds a hidden reserve from a full charge to 10%.

For more info than you will probably ingest.

https://daveinolywa.blogspot.com/2020/06/e-plus-summer-range-test.html

I haven't been following this topic. Has it been noted that the Gen 2 GOM is much improved over Gen I and 1.5? It still can't predict the future of course, but it now seems to learn your average driving efficiency, and to use that more than the energy consumption of the last few minutes. I'm sure the improvement varies by the driver, but it works quite well for both my housemate and her SV+ and for me, with my SV 40.

LeftieBiker said:

I haven't been following this topic. Has it been noted that the Gen 2 GOM is much improved over Gen I and 1.5? It still can't predict the future of course, but it now seems to learn your average driving efficiency, and to use that more than the energy consumption of the last few minutes. I'm sure the improvement varies by the driver, but it works quite well for both my housemate and her SV+ and for me, with my SV 40.

Click to expand...

Agreed. It no longer seems swayed by unusual circumstances like mountains, etc.

To be honest I don't really look at the GOM mileage estimate. It's usually far too optimistic. Malta is like a never ending city with nothing but constant stop-and-go and very large speed swings every few hundred meters. The e-pedal mode on the ZE1 is worth its weight in gold.

I do look at the GOM percentage, but even 30% of charge is enough to cross the whole country so I only really use it to plan which days to charge on.

I found my ZE1 to be quite linear except the first bit when it's just come off the charger (it tends to stick at the charged value, say 75 or 76% for a little longer than usual). By linear I mean my commute always consumes about 3% whether it's at the top (>70%), middle (~50%) or bottom (<30%) of the GOM reading.

My ZE0 not so much. The same commute will drain 10% at the top (eg. 70 to 60%) but only say 5% at the bottom (say 30 to 25%). I will try and find the time to drain it to VLBW and see if it improves the accuracy of the percentage readout.

By far, the best improvement available for the GOM is to ignore it

LeftieBiker said:

I haven't been following this topic. Has it been noted that the Gen 2 GOM is much improved over Gen I and 1.5? It still can't predict the future of course, but it now seems to learn your average driving efficiency, and to use that more than the energy consumption of the last few minutes. I'm sure the improvement varies by the driver, but it works quite well for both my housemate and her SV+ and for me, with my SV 40.

Click to expand...

I disagree, the miles remaining display in my 2019 SL Plus is no better (and no more linear) than the displays in my 2011 and 2015 SL's were. The SOC% is reasonable in the 2019 and was reasonable in the 2015. The 2011 did not have the percentage display.

A few deep discharge and full charge cycles may help the accuracy of the percentage and miles remaining displays, but the miles remaining really depends upon just the past few miles driven.

Edited to add: I agree with Sagebrush--just ignore the miles remaining display.

There have always been drivers whose behaviors matched up well with the GOM, and those (most) for whom it is useless. The gen 1.5 version was useless to me, but the gen 2 version is not. I still live in the same house, and still drive the same routes, so I conclude that the GOM is better. I'd suggest figuring out if the G2 version works for you, and if not, then ignore it.

Since making my post I've deeply discharged the 24kWh leaf and fully charged it, and currently I'm getting just over 1.3km per percent, whereas before I was getting about 1km per percent.

I drove the car until it was reading 8%, then parked in the garage and left the heater and all electrical loads running at full power (drawing about 3.8kW total). When the battery reached 5%, the percentage indication turned to "---". I kept discharging until leafspy recorded 0.8kWh left in the battery. Then gave it a full overnight charge to 100% until my inline plug wattmeter read 0 power consumption from the socket (and the 3 blue dashboard lights turned off). The GOM indicated 132km remaining, which is higher than any previous reading with this car.

I'm still discharging the battery after that. So far almost 60km driven and the battery is at 55%. The linearity is improved somewhat. We'll see once it gets to the lower percentages.

LMF5000 said:

I do look at the GOM percentage

Click to expand...

The percentage (aka SoC) is not the GOM. The GOM refers to the car's estimate of range. The SoC is (well, for our purposes) not an estimate.

Think of these differences in an ICE car:
1. The car says 5 gallons of gas remain in the tank
2. The cars says that you can travel 100 miles

(2) is the value to ignore

If you want to get into the weeds of the SoC value linearity, it comes down to whether the car is reporting Ah or kWh remaining.

Oops. That explains a lot. Throughout this thread I meant to improve the linearity of the SoC readout. I already ignore the GOM then

OK --- with that dispensed with ...

The reported SoC when the battery is mostly discharged is off because of weak cells. I think the most practical way to deal with this is to buy LeafSpy/OBD and go by the percentage of GIDs remaining. That calc appears to use the weakest cell, which IS what determines remaining usable energy.

No charging routine or exotic dance is going to improve the weak cells. They have gone through an irreversible chemical degradation.