Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

That crappy degradation chart made less crappy (you know the one)

This site may earn commission on affiliate links.

darth_vad3r

Well-Known Sith
May 6, 2019
1,574
1,178
Canada
By now everyone's seen this chart:

DST-cycles-web2.jpg


It's crap. Why? Because the Dynamic Stress Test (DST) Cycle they are using on the x-axis is the actual test cycle that only goes X% based on the legend (i.e. 10% for 75-65%, 60% for 100-40%) and *not* a battery "cycle" that would make comparing these lines more equivalent.

In reality each line should be scaled horizontally to match.

Being annoyed by this chart for the umpteenth time, I decided to (try to) do something about it.

First I recreated my own version of the crap chart by populating some of the underlying data points so that I could use those to extract more meaningful info.

My version of the same old crap:
crap-1.png


Looks pretty good, right? I mean, it looks like crap, but it looks the same(ish) as the source chart ... or close enough.

So, based on that I labelled the datapoints by the number of rated miles consumed based on 100% = 310 miles...
crap-2.png

The whopping gigantic datapoint circles are roughly equivalent 'mileage' based on using the largest number available from the most-DST-cycled info (260,000 miles). Some of the others are at 260k as well, others at 279k. Roughly.

Here's the same slightly less crappy chart, with less clutter (removed the labels):
crap-3.png

My takeaway is ... there's still something to be learned from the chart. There's not massive amounts of difference, but black, blue, red are all below 90% at the 260-279k miles mark. What do they have in common? They all charged to 100%. Red did the best cycling 100-40. Black worse 100-25, Blue worst 100-50. Why? I dunno. My takeaway ... 100% = bad.
Green is a bit better above 90%, maybe 91%. That charged to 85% and down to 25%. Takeaway, 85% is better than 100%.
Light blue is again a bit better maybe 92%. That one charged to 75% and down to 25%. Takeaway, 75% is better than 85%.
Orange, way the F off to the right, looking all awesome in the original chart ... is not so awesome. That one is in the middle of green and light blue and it charged to 75% and only down to 65%.
Purple (mauve?) is actually the winner. It also charged to 75% but only ran down to 45%. Takeaway? I dunno 75% is better, and super shallow isn't best, but averaging closer to 50% is better? Or that datapoint is crap (it is higher than the purple trend).

Overall takeaway ... 75% better than 85% better than 100%. Difference after a ton of miles might be 93% of original capacity vs 89% of original capacity.

DISCLAIMER: (i.e. don't come at me bro) ... this is *NOT* based on *Tesla's* cells or their super-de-duper awesome improved magical chemistry that might be waaaay amazeballs better than this. We don't know. But anyways, given this data that came from somewhere that might be half-decent but was visualized in a crap way and keeps getting thrown around incorrectly all over forumss and youtubes and the twitters... it might learn us sumthin' gud? I dunno. Take it as is.

If you found this useful, click like, don't forget to subscribe, and blah blah blah.

All content except the original crap chart, Copyright (C) 2019 Darth fricking Vad3r.
(If I see this in someone's youtube video, so help me ... I will not hestiate to force choke you through the interwebs!)

That is all.

Original crap chart from: https://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries but who knows where they got it from? Do you know? If you do, let us know.
 
Pffft... 10 minute time limit to edit a new thread post?

Here's the TL;DR takeaways reformatted:

My takeaway is ... there's still something to be learned from the chart:
  1. There's not massive amounts of difference, but black, blue, red are all below 90% at the 260-279k miles mark. What do they have in common? They all charged to 100%. Red did the best cycling 100-40. Black worse 100-25, Blue worst 100-50. Why? I dunno. My takeaway ... 100% = bad.
  2. Green is a bit better above 90%, maybe 91%. That charged to 85% and down to 25%. Takeaway, 85% is better than 100%.
  3. Light blue is again a bit better maybe 92%. That one charged to 75% and down to 25%. Takeaway, 75% is better than 85%.
  4. Orange, way the F off to the right, looking all awesome in the original chart ... is not so awesome. That one is in the middle of green and light blue and it charged to 75% and only down to 65%.
  5. Purple (mauve?) is actually the winner. It also charged to 75% but only ran down to 45%. Takeaway? I dunno 75% is better, and super shallow isn't best, but averaging closer to 50% is better? Or that datapoint is crap (it is higher than the purple trend).
Overall takeaway ... 75% better than 85% better than 100%. Difference after a ton of miles might be 93% of original capacity vs 89% of original capacity.
 
Great analysis.

I would argue that in the "big picture" this is all "the same" - That is, after 260,000 mi - your battery capacity will be somewhere between 88-93% of original - a 5% spread, on a LR RWD Model 3 car that' amounts to 16 mi.

So person A, who obsesses over battery settings, constantly changes it, posts all over forms how whoever isn't obsessive about it is unbelievably irresponsible, etc. will have 16 more miles of range after 15 years or so over person B, who chose the 90% set-n-forget attitude.

I know which camp I fall into.
 
  • Like
Reactions: WarpedOne
Need to modify the chart with an "actual miles" x-axis instead of just relabeling data points.

I wasn't 100% sure the DST wasn't adjusted for some type of equivalency between all the cases (like a virtual 0%-100% cycle for all cases). But maybe a few more people will look at that critically and figure it out.
 
Need to modify the chart with an "actual miles" x-axis instead of just relabeling data points.

I agreee but couldn’t easily do this and it took long enough for me to get where I got so I stopped there :)
(you can tell that’s where I was aiming for because I changed the x-axis label first, then gave up and never changed it back :)).

I wasn't 100% sure the DST wasn't adjusted for some type of equivalency between all the cases (like a virtual 0%-100% cycle for all cases). But maybe a few more people will look at that critically and figure it out.

I wasn’t sure originally either, but there are notes comparing the original(?) chart below it at the provided BU link that compare some of the cases and the total “energy units (EU)” provided by each case which you can tell from the numbers is based on DoD % x number of “DST cycles”. This plus the usage of “DST cycle” vs the traditional “cycle” lead me to be fairly confident that is what the chart represents. Of course the chart and captions could be wrong. I don’t know the source data :)

There’s a good (old) thread here that was debating it and came to the same conclusions I did above. It was in the spring, if I find it I’ll post a link.
 
That kind of explains why all those other manufacturers locked out more of the battery than Tesla, to ensure that a more optimal battery life (and to give the perception of faster charging among other things). We'll see how that plays out in the long term (for Tesla so far it seems to be ok? At least until some range reductions in the s and x because of the battery fire stuff)
 
The chart is still pretty difficult to read (aka, can be misleading) without aligning the x-axis among all the lines.

Yes. Asked and answered already ;)

Anyone know how to tell the spreadsheet to auto-scale each series by its numeric labels? I could figure out how to have two y axes, but not seven x axes :) ... other than that I need to merge and sort the mileage columns I created for all 7 data series.

I made the data points of interest gigantic because that’s the new focus for my version of the chart, the next step was to just erase the lines completely :)

Anyways. I agree it’s still a bit crappy and not what I was aiming for when I set out, which is why I labeled it a “less crap” and not “good” chart :)
 
Interesting.

Seems like the best strategy is to figure out your normal usage, and then apply it around 50% SOC. So, if you use 40% daily, then charge to 70% and down to 30%. If you use 20% daily, then 60% to 40%, etc.

I agree with this.
My takeaway would be:
1. Lower max charge is better
2. Keeping state-of-charge near 50% is better
3. Shallow discharge cycles are better, but secondary to staying near 50 state-of-charge

Edit: Having said that, I don't know that it directly translates to a Tesla, given BMS calibration issues.
 
I agree with this.
My takeaway would be:
1. Lower max charge is better
2. Keeping state-of-charge near 50% is better
3. Shallow discharge cycles are better, but secondary to staying near 50 state-of-charge

Edit: Having said that, I don't know that it directly translates to a Tesla, given BMS calibration issues.

BMS calibration issues are temporary. Degradation is permanent.

Given this, if one understands it, I think they should choose health over calibration. If you can occasionally charge up to 90% for a trip and drive down to 20% before supercharging for the next leg, I would think the calibration ‘issue’ would be minimal in the long run even if you otherwise charge 70-50-70-50... or whatever.
 
Again... you're talking a difference of, at worst, 16 miles AFTER 10-15 years of ownership. This is entirely academic.

Is a retirement savings of $1,005,000 "better" than $1,000,000. Yes. Does it make a tangible difference in retirement, probably not.

Yes. Many people enjoy being academic.

Also these are projections ... Prof. Dahn's talks mention models that look good, great even ... until they fall off a cliff suddenly. We don't know for sure what Tesla's cells will look like in 10 or 15 years, so me personally, I will make every little small change I can to potentially avoid my car battery falling off a cliff at 12 years instead of 20 years.

It costs me nothing to charge to 70% most days instead of 90%.
 
  • Like
Reactions: gecko10x
Again... you're talking a difference of, at worst, 16 miles AFTER 10-15 years of ownership. This is entirely academic.

Is a retirement savings of $1,005,000 "better" than $1,000,000
. Yes. Does it make a tangible difference in retirement, probably not.

Since you “disagreed” with my chart above with no commentary (not sure why?) I thought I’d point out your prior error I ignored. 16 miles is 5% of 310.

That’s comparing a $1,050,000 portfolio to a $1,000,000 one (not $1,005,000).

An extra $50k buys you another model 3.