Quick way to check battery when buying used Spark?

[NORTON]

...adodgerfan has 70K miles with his 2014. He wrote that he saw 17 kWh from almost empty to full using DCFC. At 94%, that would mean 16 kWh for the battery capacity. At 10K miles per year, he drove equivalent of about 7 years, yet he sees about 5 years worth of degradation. He has 2014, so you have to scale my plot for higher initial capacity. It seems his 2014 is following linear degradation curve for his usage pattern...

? "using DCFC" ???

Again, measuring what goes into a battery is NO WAY to measure capacity.
Especially at a DCFC unit. The first thing I hear is the AC compressor and the cooling fans ramping up. That is all power used during a charge.
And batteries in general are never 100% efficient at goes-in / goes-out power.

We should be glad we have the Leaf page with data. Some EV's don't give you this data.
Use it.
Plot it, because if varies a lot.
That is the only way to plot your pack's capacity.

 

[SparkevBlogspot]

DCFC is not the most accurate way, but it does give you a clue. Since he maxed out at 17 kWh, even 100% efficient would mean 17 kWh battery (about 3 years degradation from my plot scaled for 2014). I find DCFC to be about 94% efficient (again, described in my blog how that's derived), but if the weather was hotter or other circumstance for more energy used other than charging the battery, battery capacity would be even less than 16 kWh.

I did use "leaf button" to get better estimate of battery capacity few times after 100% charge at home, and they fit well with blog post data (about 18.0 kWh after 1.2 year, 17.5 kWh recently). I briefly mentioned it in the blog post, but I'm gathering few more "leaf button" data points for another blog post.

Considering MrDRMorgan seem to be getting very similar number as mine, what I find may be typical, at least for 2015. Not sure if he used DCFC as often as I did. I went to extra lengths for simulating "living in apartment without home charging" scenario, which added probably 20 or more sessions.

 

[user 1116]

@SparkevBlogspot
So you have to live in California to even be considered for the Chevrolet 8 year battery warranty?

 

[user 1116]

@SparkevBlogspot
From your prediction, the 2015/2016 may degrade slower, maybe 1 to 2 extra years until it reaches 50%. The 2014, being 1 year older, therefore technically you'll get probably 2-3 less years until 50%, am I reading your post/blog correctly? (I read your disclaimer)
Lastly, do we know for sure that lithium batteries degrade in a linear pattern?

 

[MrDRMorgan]

"4. Avg. Capacity: 17.57 5. Last calculation: 17.75"

Your average capacity is less than your last calculation? That could only happen if you gained capacity in time. I wouldn't say it's impossible as my next blog post will show some "interesting" patterns.

adodgerfan has 70K miles with his 2014. He wrote that he saw 17 kWh from almost empty to full using DCFC. At 94%, that would mean 16 kWh for the battery capacity. At 10K miles per year, he drove equivalent of about 7 years, yet he sees about 5 years worth of degradation. He has 2014, so you have to scale my plot for higher initial capacity. It seems his 2014 is following linear degradation curve for his usage pattern.

I have to stress that his usage could be very different from mine. Still, there's some hope that the degradation is more linear and not as severely affected by charge-discharge cycle at least for one car.

I think this would fall under "interesting" patterns. I plot all kinds of graphs trying to make sense of the data. Here is my average data by 1000 mile increments:

2014: kWh
4k - 5k: 18.89
5k - 6k: 18.77
6k - 7k: 18.63
7k - 8k: 18.13
8k - 9k: 18.06
9k - 10k: 17.67
10k - 11k: 17.38

2015: kWh
8k - 9k: 17.45
9k - 10k: 17.57
10k - 11k: 18.03
11k - 12k: 18.06
12k - 13k: 17.90
13k - 14K: 17.74
14K - 15k: 17.60
15K - 16k: 17.36
16k - 17k: 16.86
17k - 18k: 17.02
18k - 19K: 17.47

You can see that the 2014 seems to be degrading much more than the 2015. in fact, the 2015 looks like it has only degraded about 0.5 kWh over ~10k miles. All of this data is derived from L2 charging only.

Right now I am making note of the GOM estimate at full charge and, as I drive, compare the starting GOM with the total of the distance traveled plus the current GOM reading to see if it makes any sense. After all, RANGE is what we are concerned about and it seems we should see a drop in range if the battery degrades. As an example, I fully charged the 2015 early this morning and, at 11am, I drove to Modesto and back - 40.6 miles. A/C was on 100% of the time and was set to 75 deg. F and automatic mode. Outside temp was 89 - 91 deg. F and road speed ranged between 45 and 55 mph. The full-charge GOM read 92 miles before I left. Upon arriving home I had the following on my displays:
GOM: 57
Trip: 40.6 miles @ 5.8 mi / kwh.
Energy Info screen: 36 - 5 - 0 %, 7.0 kwh used [17.07 kwh battery capacity]

While GOM is very dependent on outside temperatures and driving habits, comparing monthly average GOM values year-to-year may give some indication of loss of range.

 

[SparkevBlogspot]

So you have to live in California to even be considered for the Chevrolet 8 year battery warranty?

Since I'm in CA, I can only comment on CA cars. SparkEV was not made officially available outside of CA+OR (2016 in MD as well), so I don't know how they'll handle the warranty. Best is to ask Chevy, though some jurisdictions could have different policy even if Chevy rejects warranty.

I can only comment on my 2015. I don't know how 2014 would behave, but people say it has higher capacity at start. If you assume same trend as my 2015, you have to scale it for 2014. That means percentages would stay the same, but absolute capacity in kWh would show higher degradation.

If you assume same trend for 2014 as 2015, that means 2014 already lost at least a year. I got mine at mid 2015, but some got their 2014 in 2013, which means some could already have 2 years more use (close to 4 years total) than my 2015. That's the less life of 2014 that I'm talking about.

Ideally, LiIon would degrade exponential decay (green plot) since "poor" ion receptors would degrade in short time while "better" sites keeps on working. But the real world usage is far different from ideal. Even for non-ideal, cell phone in one's pocket would experience ~100F temperature most times while SparkEV battery with the best battery thermal management among any EV would have easier life (or not, consider DCFC is most powerful among any EV). Basically, we (consumers) cannot know until the battery has degraded enough (ie. 8+ years).

 

[MrDRMorgan]

So you have to live in California to even be considered for the Chevrolet 8 year battery warranty?

Since I'm in CA, I can only comment on CA cars. SparkEV was not made officially available outside of CA+OR (2016 in MD as well), so I don't know how they'll handle the warranty. Best is to ask Chevy, though some jurisdictions could have different policy even if Chevy rejects warranty.

I can only comment on my 2015. I don't know how 2014 would behave, but people say it has higher capacity at start. If you assume same trend as my 2015, you have to scale it for 2014. That means percentages would stay the same, but absolute capacity in kWh would show higher degradation.

If you assume same trend for 2014 as 2015, that means 2014 already lost at least a year. I got mine at mid 2015, but some got their 2014 in 2013, which means some could already have 2 years more use (close to 4 years total) than my 2015. That's the less life of 2014 that I'm talking about.

Ideally, LiIon would degrade exponential decay (green plot) since "poor" ion receptors would degrade in short time while "better" sites keeps on working. But the real world usage is far different from ideal. Even for non-ideal, cell phone in one's pocket would experience ~100F temperature most times while SparkEV battery with the best battery thermal management among any EV would have easier life (or not, consider DCFC is most powerful among any EV). Basically, we (consumers) cannot know until the battery has degraded enough (ie. 8+ years).

After going back through the Limited Warranty and Owner Assistance Information booklet that came with my 2015 Spark EV, I see nothing that says anything about the warranty being limited to ANY state. The warranty covers the Spark EV regardless of where it is located in the USA at the time of the claim.

 

[SparkevBlogspot]

But SparkEV was only made officially available in CA+OR (and MD). One could argue that they bought the car in CA originally and moved out of state, or that selling used car anywhere in the nation should honor the warranty. But I don't think there's any car from Chevy that was restricted like SparkEV, so making things crystal clear with official written notice from Chevy would keep peace of mind rather than having to fight them in court should warranty issues arise.

 

[PhilPen]

The degradation of the Tesla lithium cells seems to follow a different curve, in this case there seems to be a relatively fast small loss (5%) in capacity followed by a plateau/slow decline.

https://electrek.co/2016/11/01/tesla-battery-degradation/

 

[SparkevBlogspot]

Better data for Tesla can be found here. (more recent, May 2017)

https://steinbuch.wordpress.com/2015/01/24/tesla-model-s-battery-degradation-data/

Tesla data is comparing many cars, and dubious way of gathering the data (miles and by various people). For example, one could be blasting the heater, and nothing for the other. What I have is pretty spot-on for my car since I'm using energy. I also verified using "leaf button" data collection.

Since I only have 20K miles, that's what I wrote about; just the facts and simple extrapolation. I have about 400 charge cycles and 800 days, and that roughly agrees with Tesla data. Tesla shows about 6% but I have about 8%.

SparkEV would go through more than 3 times the number of charge cycles of Tesla for given number of miles as well as DCFC charging rate that's over 2.5 times higher than Tesla. Considering how I ran the battery down many times and so many DCFC, and it's held up almost as well as Tesla, it's a testament to how great SparkEV is.

My next blog post will have more nuanced info, looking at theory and various SparkEV degradation and Tesla's. You can probably tell from adodgerfan data that the degradation may not be so bad. However, one should be prepared for the worst if looking to buy a SparkEV: 8.001 yrs / 100.001K miles and 65% capacity remaining. That means 2014 is riskier than newer years.

 

[MrDRMorgan]

Better data for Tesla can be found here. (more recent, May 2017)

https://steinbuch.wordpress.com/2015/01/24/tesla-model-s-battery-degradation-data/

Tesla data is comparing many cars, and dubious way of gathering the data (miles and by various people). For example, one could be blasting the heater, and nothing for the other. What I have is pretty spot-on for my car since I'm using energy. I also verified using "leaf button" data collection.

Since I only have 20K miles, that's what I wrote about; just the facts and simple extrapolation. I have about 400 charge cycles and 800 days, and that roughly agrees with Tesla data. Tesla shows about 6% but I have about 8%.

SparkEV would go through more than 3 times the number of charge cycles of Tesla for given number of miles as well as DCFC charging rate that's over 2.5 times higher than Tesla. Considering how I ran the battery down many times and so many DCFC, and it's held up almost as well as Tesla, it's a testament to how great SparkEV is.

My next blog post will have more nuanced info, looking at theory and various SparkEV degradation and Tesla's. You can probably tell from adodgerfan data that the degradation may not be so bad. However, one should be prepared for the worst if looking to buy a SparkEV: 8.001 yrs / 100.001K miles and 65% capacity remaining. That means 2014 is riskier than newer years.

I too collect the Energy Information screen data to try and understand what is going on with my 2014 and 2-015 Spark EVs.

2014:
At 1000 miles on the ODO my calculated average battery capacity was 19.81 kWh.
At 10,000 miles on the ODO my calculated average battery capacity was 17.38 kWh
-Down 12%

2015:
At 1000 miles on the ODO my calculated average battery capacity was 18.61 kWh.
At 18,000 miles on the ODO my calculated average battery capacity was 17.47 kWh
-Down 6%

I question just how good the Info Screen data is in making this calculation. Friday, I drove my fully charged 2015 Spark EV 40.6 miles to Modesto and back. The Info screen showed:
36-5-0 % and 7.0 kWh for a capacity calculation of 17.07 kWh.

Later that day, my wife drove the same car to a Walmart and back - about 8 miles. The Info screen changed to:
40-7-3 % and 8.9 kWh for a capacity calculation of 17.8 kWh.

Go figure! I now average my data by month and by 1000 mile increments to get a better view of what is going on. I have also started looking at the full-charge GOM readings at the start of a trip and compared that number against the sum of the end-of-trip GOM number plus the ODO reading.

After all is said and done, to me, RANGE degradation is what I want to see. Right now my full-charge range for my 2014 is 105 miles. My 2015 is at 95 miles.

 

[SparkevBlogspot]

Since there aren't many sig figures, you need many data points picked at random and leverage central limit theorem.

In your example,

36%, 5% could've been 35.5% and 4.5% If you use those,
7/(35.5+4.5) = 17.5 kWh

40%, 7%, 3% could've been 40.49%, 7.49%, 3.49%. If you use those,
8.9/(40.49+7.49+3.49)=17.3 kWh

When I estimate, I charge to full and record the data at random interval as much as I can. At one time, I recorded over 100 data points for 100% down to 20%. Each % reading has more than one corresponding kWh reading. I once counted three kWh reading for one % reading.

Best might be OBD, but second best might be go-pro pointed at the screen and driven at night. I tried during the day, but the glare wiped out the video.

 

[MrDRMorgan]

Since there aren't many sig figures, you need many data points picked at random and leverage central limit theorem.

In your example,

36%, 5% could've been 35.5% and 4.5% If you use those,
7/(35.5+4.5) = 17.5 kWh

40%, 7%, 3% could've been 40.49%, 7.49%, 3.49%. If you use those,
8.9/(40.49+7.49+3.49)=17.3 kWh

When I estimate, I charge to full and record the data at random interval as much as I can. At one time, I recorded over 100 data points for 100% down to 20%. Each % reading has more than one corresponding kWh reading. I once counted three kWh reading for one % reading.

Best might be OBD, but second best might be go-pro pointed at the screen and driven at night. I tried during the day, but the glare wiped out the video.

Good points - thanks.

 

[NORTON]

..You can see that the 2014 seems to be degrading much more than the 2015. .... All of this data is derived from L2 charging only....

You're also plotting 'Charge Data' only ?

I can see where this can give you accurate 'Cost of Driving' data, which, in the end, is what matters with Cost of Ownership.

But it is not measuring the Battery's Capacity.

Charging takes a lot more power than what the battery puts out in power.

My charge data would change seasonally by a lot. Heating in the Winter and Cooling in the Summer.
And even in the best of times, Spring and Fall, when the least amount of Thermal Management System would be needed,,
the BMS continues to use power to Balance the cells, and whatever 'housekeeping' is taking place when the car is done charging but still drawing 600-800 Watts for 15-25 mins after the car reports is done charging.

Maybe '14's and '15-16's have different TMS and BMS protocols.

I always get 2 separate SMS txt msgs when the car tops up at a public charger. One from Onstar, later, one from ChargePoint.

Some owners reported problems with their cars continuing to draw ~600 Watts for a long time after 'Charge Complete'. Whatever happened with that problem???

 

[MrDRMorgan]

..You can see that the 2014 seems to be degrading much more than the 2015. .... All of this data is derived from L2 charging only....

You're also plotting 'Charge Data' only ?

I can see where this can give you accurate 'Cost of Driving' data, which, in the end, is what matters with Cost of Ownership.

But it is not measuring the Battery's Capacity.

Charging takes a lot more power than what the battery puts out in power.

My charge data would change seasonally by a lot. Heating in the Winter and Cooling in the Summer.
And even in the best of times, Spring and Fall, when the least amount of Thermal Management System would be needed,,
the BMS continues to use power to Balance the cells, and whatever 'housekeeping' is taking place when the car is done charging but still drawing 600-800 Watts for 15-25 mins after the car reports is done charging.

Maybe '14's and '15-16's have different TMS and BMS protocols.

I always get 2 separate SMS txt msgs when the car tops up at a public charger. One from Onstar, later, one from ChargePoint.

Some owners reported problems with their cars continuing to draw ~600 Watts for a long time after 'Charge Complete'. Whatever happened with that problem???

I am aware of the charging losses. All of my battery calculation data is taken from the Energy Information screen before I start to recharge.

For DCFC charging, I charge to 85,90 or 95%. I collect the start and end percentage data plus the number of kWh delivered. Currently, there are four separate EVgo ABB chargers I frequent. For my capacity calculation, I use 97% charging efficiency although others think it may be 94%. Between March 2017 and current date, I have charged 19 times and my average capacity calculation result is 17.54 kWh. Maximum capacity so far is 18.12 kWh and minimum capacity is 16.89 kWh.

As long as I am consistent in my data gathering, the actual capacity number will not be as important as seeing if, over time, the calculated battery capacity degrades. Ambient temperature affects battery capacity so I have to keep that in mind too as I collect data.

 

[NORTON]

I am aware of the charging losses. All of my battery calculation data is taken from the Energy Information screen before I start to recharge.

For DCFC charging, I charge to 85,90 or 95%. .... I use 97% charging efficiency although others think it may be 94%. .. Maximum capacity so far is 18.12 kWh and minimum capacity is 16.89 kWh.
... the actual capacity number will not be as important as ..... Ambient temperature affects battery capacity so I have to keep that in mind too as I collect data.

Battery temp when you arrive at a DCFC is also important.
With my car, after flogging it down the highway this time of year when I pull into a DCFC to get a top up, the first thing I hear is the AC compressor and fans working to cool the pack in advance of the 48kW charge that is coming.
And I too never take it to 100%. I rarely take it to 80%. I have places to go!

I don't understand why you guys would gather all this roundabout questionable charging power used data, and then subtract 'xx% charging efficiency'.

The leaf page numbers are simple car generated data. Why bother with questionable DCFC data? So much of that power is going other places, depending on the season and what you have been doing before you arrive at the DCFC.

I doubt I get a range of 18.12 to 16.89 kWh with using leaf page data. Someday I should learn how to make pretty graphs!

 

[MrDRMorgan]

I am aware of the charging losses. All of my battery calculation data is taken from the Energy Information screen before I start to recharge.

For DCFC charging, I charge to 85,90 or 95%. .... I use 97% charging efficiency although others think it may be 94%. .. Maximum capacity so far is 18.12 kWh and minimum capacity is 16.89 kWh.
... the actual capacity number will not be as important as ..... Ambient temperature affects battery capacity so I have to keep that in mind too as I collect data.

Battery temp when you arrive at a DCFC is also important.
With my car, after flogging it down the highway this time of year when I pull into a DCFC to get a top up, the first thing I hear is the AC compressor and fans working to cool the pack in advance of the 48kW charge that is coming.
And I too never take it to 100%. I rarely take it to 80%. I have places to go!

I don't understand why you guys would gather all this roundabout questionable charging power used data, and then subtract 'xx% charging efficiency'.

The leaf page numbers are simple car generated data. Why bother with questionable DCFC data? So much of that power is going other places, depending on the season and what you have been doing before you arrive at the DCFC.

I doubt I get a range of 18.12 to 16.89 kWh with using leaf page data. Someday I should learn how to make pretty graphs!

I also collect full-charge GOM range values. Plotting these values against the time of the year does give a good picture of what to expect from the car. 1 kWh is good for about 5 miles of range. So, 1 kWh of any battery degradation should plot lower when compared to the previous year. That is what I really care about - range degradation. Obviously, this assumes my driving habits and those of my wife are fairly uniform year to year. So far, they appear to be. For example, my 2014 Spark EV average of the full-charge GOM values for Apr 2017 thru June 2017 is 102 miles. The average for almost the same time period in 2016 is 105 miles. That suggests to me that very little battery degradation has occurred over the 1 year time span.

 

[user 1116]

There's a 2016 with 5500 miles for $13652 for sale for some time, guessing probably 30 days or so. Wonder how much I should offer, esp considering I almost got a 2015 with 14000 miles for 8500?

 

[PhilPen]

Better data for Tesla can be found here. (more recent, May 2017)

https://steinbuch.wordpress.com/2015/01/24/tesla-model-s-battery-degradation-data/

Tesla data is comparing many cars, and dubious way of gathering the data (miles and by various people). For example, one could be blasting the heater, and nothing for the other. What I have is pretty spot-on for my car since I'm using energy. I also verified using "leaf button" data collection.

Since I only have 20K miles, that's what I wrote about; just the facts and simple extrapolation. I have about 400 charge cycles and 800 days, and that roughly agrees with Tesla data. Tesla shows about 6% but I have about 8%.

SparkEV would go through more than 3 times the number of charge cycles of Tesla for given number of miles as well as DCFC charging rate that's over 2.5 times higher than Tesla. Considering how I ran the battery down many times and so many DCFC, and it's held up almost as well as Tesla, it's a testament to how great SparkEV is.

My next blog post will have more nuanced info, looking at theory and various SparkEV degradation and Tesla's. You can probably tell from adodgerfan data that the degradation may not be so bad. However, one should be prepared for the worst if looking to buy a SparkEV: 8.001 yrs / 100.001K miles and 65% capacity remaining. That means 2014 is riskier than newer years.

I agree the Tesla cells have a much easier life in all regards, they also likely have a much lower average discharge rate in typical use, as well as the mentioned fewer cycles and slower relative charging rates. I also feel that due to range concerns SparkEV owners are likely to try and keep their Sparks fully charged more than Tesla owners further stressing the cells. The battery chemistry is also likely to have some propriety differences, and cylindrical cells are supposed to be more manageable in terms of temperature regulation. We'd therefore expect the SparkEV cells to die faster and possibly with different kinetics than the Tesla cells. However, I think the LG batteries must be at least comparable in performance and possibly better for the price because in the Tesla Roadster upgrade Tesla switch to LG Chem cells to produce an EV with ~400 mile range.
https://www.treehugger.com/cars/tesla-partners-lg-chem-roadster-ev-battery-upgrade-40-more-energy.html

 

[SparkevBlogspot]

Actually, cylinders have smaller surface area, so it would be worse than cube if all things are equal. But cylinders have aluminum walls while prismatics have "plastic" walls, so that makes the cylinders "better". Not sure if SparkEV cells have "plastic walls", but they probably do.

But when applied in packs, packing makes a huge difference, and SparkEV is far better than Tesla. Tesla uses "ribbons" of coolant that snake through the pack and only make contact with about 1/3 the cell area, not to mention high variations from cells making early contact with coolant vs later one. 2015+ SparkEV uses cooling plates between all cells, thus contact area is 80% or more and more uniform cooling for cells.

2014 SparkEV uses cooling plates on the bottom, which means not as much contact with coolant. Not sure if this is part of the reason why some see more degradation for 2014.

Bolt also uses cooling plates on the bottom like 2014 SparkEV. Bolt's DCFC rate is much less than SparkEV. In fact, Bolt's DCFC is less than any DCFC capable EV; Bolt is estimated to top out at 60 kW, which is 1C, which is less than Tesla's 2C peak and SparkEVs 2.3C (2014) / 2.6C (2015+) average to 80%. Cooling might be part of the reason.

Kinda getting off topic, but from charging point of view, SparkEV is the quickest charging among any EV (2.3C/2.6C), and it's unlikely any car will beat SparkEV for the foreseeable future. Assuming 50 kWh is the smallest EV battery in the future, it would have to maintain 130 kW (2.6C) to 80%. Nothing that's announced can do that, and battery sizes are only getting larger.