The (De-) evolution of My Laptop Battery

Update: the GitHub repo for this data/script is now available.



Today my MacBook Air is one year old. That's not exactly an officially recognized holiday, but it does mean one thing very cool:
I have one year of data on my laptop battery, recorded every 1 minute of computer usage!
Epic

A little backstory:
I started occasionally keeping track of my laptop's battery several computers ago. Near the end of my previous computer's life I realized I could automate this collection of data. By keeping a record of the battery charge every minute my computer is being used, I am able to track the health of my notebook, as well as study my own computer usage in remarkable detail.

In a previous blog post I noted that it would only take a negligible amount of hard drive space to keep such a record for the entire life of your computer at 1-min sampling (though I under predicted the amount of space by about 3x). The previous battery study also provided me with subject matter that was included in a quantified-self art exhibition in Ann Arbor last year. While I obsessed about what to buy for this latest computer, battery life never factored in to my equation. Every model seemed to boast more than enough capacity.

Without further ado, here is what 1 year (152,411 samples) of battery capacity data looks like for my 2012 MacBook Air:

Look at that awesome sampling! I use this computer as my work and home machine. It's actually my only regularly used computer (big monitors are for wimps). I try to take good care of the computer, cleaning it, letting the battery cycle, keeping the harddrive with lots of space.

When you fold my computer usage over a 24 hour period, looking at how frequently the notebook is in use (ignoring the battery data), you get this wonderful diagram:
Bam! You can clearly see when I'm (usually) sleeping, and my diminished use of the computer around dinner time. 

This figure can be taken to the next level (I've shown versions of this in my previous posts about computer batteries). Here we have the logs folded over Time of Day as a function of time, one little dot for every minute the notebook is in use.
Now we've got some good trends! Normally I'm very regular about starting my day and taking a break for dinner/relaxing in the evening. A notable change in behavior occurred at Days=300-ish, when I started my summer internship at Microsoft Research (they happily provided me a computer for work). At Days=345-ish I went on a proper vacation, and didn't open my laptop once!


But back to the battery life, and here's where it gets really interesting! I've noticed my Air's battery just ain't quite what it used to be. When I compared the record of battery capacity with that of my old MacBook Pro, my jaw hit the floor:


Comparatively the Pro destroyed my newer Air! Even though the sampling is quite sparse for the first year, the Pro's trend was clearly more flat. My usage behavior is basically identical. Both the Pro and the Air were my primary (only) computer being used at home and work. Both computers were being used for astronomy grad school.

My current battery cycle count is 223. My MacBook Pro got to around 750 before I replaced it, roughly  giving the same number of battery cycles per year as now. According to Apple, the Air's battery is designed to last 1000 cycles at 80% capacity, nominally over 3 years (length of the Apple Care Warranty I purchased). Based on this figure, I'm skeptical that it will reach this target. Further, if I took it to the Apple store for inspection, it's not clear to me that any problem would be diagnosed (to be clear, the computer is working wonderfully thus far, absolutely the best Apple notebook I've purchased in the past decade)

So what's changed? My usage is the same, as far as I can tell. Some others have reported issues, describing dramatic battery decay in 2012 era Mac notebooks. Maybe these new batteries are just not built as well? My MacBook Pro had a capacity of 4600mAh, while the MacBook Air boasts a factory capacity of 6667mAh. Perhaps the engineering required to provide 45% more power for a computer almost half the size means the battery isn't as durable.

What ever the reason, I'd love to know:
A) what kind of performance are other people getting out of similar aged notebooks?
B) what is Apple's response to such issues?
C) are there other similar datasets available for comparison?

The MacBook Air has been my favorite Apple computer yet, I can't see myself going back to a bulky full sized laptop ever again. Maybe these new "all day" batteries won't have this problem. Happy birthday, computer.


The GitHub repo for this data/script is now available.

61 comments:

  1. My 2 yr old MBA has 397 cycles and has been demanding to have the battery serviced for a few months (one of these days I'll get around to actually doing it. Sigh...

    ReplyDelete
  2. I got a base model MacBook Pro last year. I have 303 cycles and am at about 345 days. iStats Pro says I'm at 90% health. On the other hand, my 2008 white polycarbonate MacBook's battery lasted about 1000 cycles before I replaced the computer and is just at 85% health. I have to say that the older batteries seemed to last longer.

    ReplyDelete
  3. I would suspect the heat. In the Air, the battery is thin and everywhere, even near hot components, whereas in the older models it was a monolithic block in a relativly cool corner.

    ReplyDelete
    Replies
    1. It'd be interesting if temperature data were also available, at the granularity you can get from an app like iStat Menus. That might shed more light on the heat degradation hypothesis.

      Delete
    2. Alas I didn't save any temperature data. I too think heat may be a culprit.

      I'm going to put my cron script up soon, will try to add temperature to the output!

      Delete
    3. Can't wait to use your script on my MBP!

      Delete
  4. can you put your cron script on github !

    thanks

    ReplyDelete
  5. Did you ever think about the fact that macbook air doesnt give you the option to replace the battery (yourself) and this might be a nice moneymaker for apple?
    Over the last 5 years hardware became so powerfull that you dont need to replace it every 2 years to be able to use the latest operating system/software.
    My guess is that maybe apple opted for the less quality battery so you would be forced to replace your notebook earlier. Or at least would pay apple for replacing your battery.

    ReplyDelete
  6. This comment has been removed by the author.

    ReplyDelete
  7. Great post! I have to check but i think the MBA battery is also a molded Li polymer battery. So there some chemistry differences and mAh is only part of the power equation for batteries. The current draw and voltage play a role as well. I know the monolithic battery in my 2008 Aluminum MacBook died more suddenly. perhaps the MBA's battery has more graceful degradation.

    ReplyDelete
    Replies
    1. Hmmm...it seems the 2009 MBP does have similarly improved battery chemistry. http://www.anandtech.com/show/2783/apple-s-2009-macbook-pro-battery-life-to-die-for

      Delete
  8. Your posit about the new style of battery being to blame is most likely the correct one. The new batteries use a nanostructured cathode that greatly increases surface area, the downside being these smaller features degrade much quicker resulting in the loss of capacity you see.

    ReplyDelete
  9. TomE: My 15" Retina MBP has 201 cycles and 270 days of use on it (similar to this author's usage pattern) and also has a non-removable battery, but it has only degraded to 95% of it original capacity. I don't think Apple deciding to put in a less durable battery in order to make $130 three years later on a fraction of users is a sufficient explanation.

    ReplyDelete
  10. make an app that records and plots that data, and release it for free. encourage people to submit their data (anonymized) to you for correlation.

    ReplyDelete
    Replies
    1. Btw. There's already CoconutBattery for Mac which can log your Macbook's battery status, although it doesn't do that automatically. It also can upload the data to the author's server and show how's your battery faring compared to other users. They have already 120k samples.

      http://www.coconut-flavour.com/coconutbattery/

      You can also checkout the different models and their average battery life

      http://online.coconut-flavour.com/

      Delete
  11. There's an aspect of the graph that really caught my eye. From about days 222 to 260 the battery seems to be making a distinct recovery. From 260 to 282 it sharply crashes, then seems to be recovering again from 282 to 306. I realize the data is noisy, but those trends are clearly outside the usual range of deviation for the data. Based on the dates perhaps you can speculate what was happening differently there? Maybe there's some sort of charging or usage behavior that helps the battery recover capacity? Things like draining the battery more or less between chargings, the charging frequency or duration, or maybe storing it in a warmer or cooler place while charging?

    ReplyDelete
  12. If your battery drops below 80% before doing the rated number of cycles it is covered under apple care. Apple Technicians have a test for this issue.

    ReplyDelete
  13. "According to Apple, the Air's battery is designed to last 1000 cycles at 80% capacity, nominally over 3 years (length of the Apple Care Warranty I purchased). Based on this figure, I'm skeptical that it will reach this target. Further, if I took it to the Apple store for inspection, it's not clear to me that any problem would be diagnosed"

    I wouldn't be so sure. If the battery has less than 1,000 cycles on it, is less than 3 years old and has a full charge capacity that is less than 80% its original capacity, I am pretty sure they're going to replace it for you, since you have an AppleCare Protection Plan. Actually, if it does this inside the one year warranty, I'd expect them to replace it as well.

    ReplyDelete
  14. So I have no data logged or anything, but I bought my MacBook Air roughly one year ago and iStat reports 78% health after 336 cycles.

    ReplyDelete
  15. Letting the battery drop right down before charging back up again (deep cycling) is bad for modern batteries. You will get a better battery life letting it only go down to around 30%

    ReplyDelete
  16. You don't specify your charge-discharge methodology. Modern batteries dislike:

    - Being fully charged.
    - Being fully discharged.
    - Being charged quickly.
    - Being discharged quickly.
    - Being charged while hot.

    Every one of those will significantly reduce the decay.

    Most high-end notebooks today ship with quick chargers.

    To maximize battery life even with your existing charger, you should avoid charging above 90% and discharging below 10%.

    ReplyDelete
  17. Does this change any of the parameters or weights in your "Awesomeness Parameter?"

    ReplyDelete
    Replies
    1. Good question! I have been thinking about how to factor battery life in to my next computer purchase, though its not clear yet to me how to get enough data about battery lifespan (maybe thru Coconut Battery?)

      Delete
  18. Any chance that running the script once per minute had an appreciable affect on the battery life? I'm wondering if the script prevented it from staying in a lower power state as often as it otherwise would have.

    ReplyDelete
  19. Thanks for the info!

    On my 2011 MBA (MacBookAir4,1), I have the following stats ... not sure why yours are so different on the newer model:

    MaxCapacity = 4113
    DesignCapacity = 4680
    CycleCount = 542
    DesignCycleCount9C = 1000

    ReplyDelete
    Replies
    1. The Coconut Battery app seems to confirm something fishy is going on if you check their database. Which MBA is this?

      Delete
  20. Buying apple hardware is not very smart itself but then wondering they are becoming worse with every product cycle is pretty much typical of apple users. They are produced in the cheapest china factories, Paragon mostly, and get usually beaten by every 200 Euro Netbook in every aspect of robustness. But on the other hand, you got some nice looking furniture so be happy.

    I have sampled the battery data of my EEE (light, smaller and booting faster than any apple) for over five years with three OS: XP, Ubuntu 10.04 and 12.04. I only recently replaced my battery after 4.5 years with a 10 hours battery which did cost me €30 - my first battery started at 6 hours and went down to 3 in the end. Though I might add that the new one seems to age a bit faster than the last one.

    ReplyDelete
    Replies
    1. EEEpc's are absolute garbage, at least you are running Ubuntu, hopefully you're mainly coding with it so you don't have to use the terrible trackpad much.

      EEEpc's are not fast.Obviously. Atom far slower than dual core i5 or i7, and your drive is not capable of 400+Mb/s sustained read/writes like 2012 MacBook Airs are or 750+Mb/s sustained read/writes of the 2013 MacBook Airs.

      Delete
  21. Degradation is bad, sure. But I'd be more curious about the results. What was the original runtime of the macbook versus the original runtime of the air, based on one full charge. What was the runtime at the end of the data cycle. If the air still has more runtime than the macbook at the same age, its still better despite a steeper degradation. Just means they put in a lighter, more powerful battery that also, unfortunately, degrades faster than the heavier, less powerful battery.

    ReplyDelete
  22. Jim, I posted about this on GigaOm where there was an article about your blog post.

    You are clearly very intelligent and perceptive, and your effort is to be applauded, but your data presentation is "iffy". Specifically in the "summary" graph: Capacity vs. Time. Your choice of a non-zero baseline on the vertical axis really presents a skewed portrayal of the difference between the two machines.

    I refer you to Edward Tufte's work on data presentation: it can be "work" to read and a bit esoteric, but the principles of data presentation are sound and the writing is illuminating, particularly for those concerned with data presentation.

    Anyway, interesting post, good luck as you pursue your doctorate, and, finally, after three years, my MBA 11" is still quite serviceable in term of battery life.

    ReplyDelete
    Replies
    1. Thanks for taking the time to comment here as well!

      Fortunately I have read my Tufte, all 4 books worth. I'll own the choice of the vertical axis scale (since I obviously made it) as being one of the MOST frequent ways people lie with data. The worst offenders are of course TV stations who make bar charts.

      Still, I'm unrepentant living in my sin of vertical skew. The graph effectively visualized the story I wanted to tell: the difference in the two computers, rather than the absolute performance of either one.

      I'd argue my presentation is both accurate and precise, but I welcome your disagreement.

      bottom line: I don't make any $ off this site (ads suck) I'm just thrilled you're reading and that we're having this conversation. Cheers!

      Delete
  23. I wonder if the new MBP's have the same issues.

    Also, for the sake of your eyes, I hope you're using flux:
    http://justgetflux.com/

    ReplyDelete
    Replies
    1. I'm a huge fan of flux, use it on all my machines!

      Delete
  24. My 2.5 years old (32 months) MBA is holding strong at 82% capacity and 602 cycles.
    In the last year it has only decreased from 85% to 82%.

    ReplyDelete
  25. The Mac Book Pros battery of about the same capacity as the Air. You are comparing the electrical charge of the batteries without taking into consideration the different electric potentials (voltage) that that the batteries provide.
    In order to compare them fairly you have to calculate the total amount of electrical enegery stored (Watt Hours):

    MBP: 4600 mAh * 10.8 V = 49.68 Wh
    Air: 6667 mAh * 7.2 V = 48.02 Wh

    ReplyDelete
  26. Would it be possible to post versions of your graphs on a scale of 0 to 100% rather than 80 to 100%?

    ReplyDelete
  27. My 2011 Macbook air is much different.

    Current Capacity 4228 mAh
    Design Capacity 4680 mAh

    90%
    378 Cycles

    Age of the computer, 18 months.

    ReplyDelete
  28. This comment has been removed by the author.

    ReplyDelete
  29. My 15" MBP Retina: 92%, 431 cycles, 387 days old.
    (Went from 93% to 92% while I posted :-)

    ReplyDelete
  30. This is nothing new and has been going on since devices started using lithium ion batteries. It may vary slightly depending on mAh ratings, but it comes down to the chemical makeup of the batteries. The first graph demonstrates it perfectly, as the number of cycles increases, the max capacity of the battery decreases. Your device will still show a 100% charge. This is because as far as the battery is concerned, it is in fact charged to its full capacity. The max capacity just deteriorates over time.

    I would be interested to see data on how long it takes to charge the battery, from 5% to 100% charge, within the same amount of cycles. Theoretically, it should take about 15% less time to charge the battery to "full capacity" after about 350 cycles (based on info from the graph)

    ReplyDelete
  31. Keep in mind it *is* possible you just have a bad battery. Once it drops below that 80% threshold before 1000 cycles, Apple's battery diagnostic that they run at the Genius Bar will flag it for replacement. If you're within 1 year, it's covered regardless. Otherwise, AppleCare gives you three years of coverage.

    ReplyDelete
  32. It would be pretty cool if you used it with rrdtol.

    ReplyDelete
  33. So lithium ion batteries (like in the air) act differently than nickel metal hydride and lead acid batteries. Fully cycling them significantly decreases their lifetime, even though it is recommended for those other battery chemistries.

    ReplyDelete
  34. MBA 2012, 396 cycles, 13 months old, 4738 mA/h max capacity (out of 6700), "Service Battery".

    ReplyDelete
  35. This comment has been removed by the author.

    ReplyDelete
  36. Do you use Carbonite? I have noticed that my battery has gone from 100%->83% in 9 months. However, I use both Caffiene and Carbonite. Carbonite seems to use 100% of one of my threads constantly. This keeps the procs warm. Caffeine ensures that it doesn't sleep.

    Just some thoughts.

    ReplyDelete
  37. This comment has been removed by a blog administrator.

    ReplyDelete
  38. Letting my mbp retina mid 2012 replace today because it crashes everytime at 3% left (approx 10 min) without warning whatsoever (please plug in your power adapter yada yada...)
    226 cycles, exactly one year old

    ReplyDelete
  39. http://topdocumentaryfilms.com/light-bulb-conspiracy/

    ReplyDelete
  40. Very interesting... 14 months old, at 99% battery capacity. So clearly they know how to make a good battery still.

    ReplyDelete
  41. I have a Macbook Air Mid 2012 Model. I took it in for a battery swap after coconut recording 303 charge cycles and 78% capacity. Genuis bar diagnostics test indicated faulty battery. My new battery seems to last less than 4 hours now on light streaming and browsing.

    ReplyDelete
    Replies
    1. This was 20 days ago. My warranty just expired. Didn't bother to buy apple care since battery replacement costs ~$130

      Delete
    2. 4hrs seems very short for a new battery on a 2012MBA. I know all work from Apple is usually warrantied for 90days, so you should be covered. Might as well have it looked at, eh?

      Delete
  42. What radioactivated areas you visited?

    ReplyDelete
  43. Thank you very much for your post. It inspired me to create a similar data collection script. I'm also grabbing a couple of additional items such as battery temp. I've placed my stuff on GitHub at https://github.com/mikegrb/Battery-Logger maybe someone will find it useful.

    ReplyDelete
  44. This comment has been removed by the author.

    ReplyDelete
  45. What about this counter-measure?

    http://translate.google.fr/translate?sl=fr&tl=en&js=n&prev=_t&hl=fr&ie=UTF-8&u=http%3A%2F%2Fwww.greenit.fr%2Farticle%2Flogiciels%2Fmac-allonger-la-duree-de-vie-de-votre-batterie-4731&act=url (translation)

    http://www.greenit.fr/article/logiciels/mac-allonger-la-duree-de-vie-de-votre-batterie-4731 (original in French)

    ReplyDelete
  46. This comment has been removed by a blog administrator.

    ReplyDelete

Inappropriate comments, advertisements, or spam will be removed.
Posts older than 2 weeks have moderated comments.
(Anonymous commenting disabled due to increasing spam)