- Charge frequently and don't allow full discharges (Li-ion batteries don't have memory);

- Keep batteries at temperatures lower than 30ºC;

- Charge with voltages lower than 4.1V/cell.

That's basically what it says, yeah. It's cool that you can just skim the three graphs and take away the main points of the article.

But that's hardly news.. The problem is that achieving these guidelines on a burning MacBook without an option to remove the cell is impossible :-/

Good thing the engineers know this all along and try to design so that the conditions are met to some extent.

I would just add to this, don't keep battery at 100% (unplug it when it's done charging)

> I would just add to this, don't keep battery at 100% (unplug it when it's done charging)

Absolutely incorrect. What makes you think that keeping a 100% charged battery plugged in bad? The device has its own charging circuitry that stops charging when full.

Anedoctal evidence, my SOs Dell's battery is basically dead(about 2-3minutes of charge) even though it has spent 95% of time plugged in fully charged.

Laptop forums are full of stories like this.

I would have thought the charging circuitry takes care of this?

Thinkpads offer a way to limit the percent State of Charge (SoC) while charging. AFAIK this is the only laptop which does so.

I do this manually by removing the battery from my laptop once it reaches 70-90% SoC (and I'm plugged in, obviously). This is superior for battery protection because the temperature is limited to room temperature, but inferior because the laptop shuts off on power loss.

Are the battery life savings really worth the hassle and potential for lost work?

Since I've picked up my Thinkpad, I really wonder why Lenovo doesn't stick in a small internal battery ontop of the removable (say good for like 10 minutes or something...) so that: a) On power loss (battery removed), it can still dump into hibernate (the real one) b) So you can hotswap the external batteries. I have two batteries, and when I'm out for long periods, it annoys me that I have to dump it into hibernate so I can swap batteries. It's just an extra few minutes that really annoy me.

Some laptops have room for two batteries. My old FSC Celsius H did.

Wouldn't total mAh delivered be a better metric than charge cycles, since less power is delivered per cycle with shallower discharges? Some quick calculations, assuming a linear loss of capacity over time:

Assume an example rated battery capacity of 1700mAh

The average capacity over the life of the battery (i.e. reaching 70% capacity) is 80% (assuming batteries start at 90% and drop linearly to 70%).

At 100% DoD, 500 cycles * 80% * 1700mAh = 680Ah

At 50% DoD, 1500 cycles * 80% * 50% * 1700mAh = 1020Ah

At 25% DoD, 2500 cycles * 80% * 25% * 1700mAh = 850Ah

At 10% DoD, 4700 cycles * 80% * 10% * 1700mAh = 639Ah

So clearly a 50% DoD is better than 25% or 10%, but where does the optimum lie?

   Depth of discharge |    Discharge cycles

    100% DoD          |         500

     50% DoD          |        1500

     25% DoD          |        2500

     10% DoD          |        4700

So....4700 discharges with 10% DoD. Isn't that just like 470 full (100% DoD) discharges in terms of total energy supplied? In other words, the battery lifetime can be described by a single amount of energy (e.g. 1000 kWh). It doesn't matter if you use that much energy interspersed with frequent recharges, or with full discharges each use.

The products are 500 discharges, 750 discharges, 625 discharges, and 470 discharges, so it turns out that it does matter a bit; you can store 50% more total energy in the battery over its lifetime if you discharge it 50% each time than if you discharge it either 100% or 10% each time.

Shallow and deep discharges are not as good because they cause the battery to spend more time in the "danger zone" of full or zero.

Notice the 50% figure results in 750 full-charge equivalent cycles. Also, these figures are general and the magnitude varies based on the exact battery type.

Hmm. I have 15 minute interval logs of battery stats for my mbp since the day I have it logged (with gaps when it was asleep/off) .... about 20,000 entries.

I could extrapolate the data based on my usage patterns (the charging/discharge/etc are all in there)

Anyone know offhand how to tell the charing circuits in the mac to stop charging at 50% capacity? I'm thinking I could easily automate that for a few weeks or months and check out the log data...