back to article Better gadget battery-level readouts in pipeline

You know the way your phone shows a full battery for ages and ages, then as soon as you make or take a call it throws a double six and shuts down? Or the way your mp3 player drops towards empty while playing, then suddenly climbs up to almost-full again as soon as you press stop? Annoying, isn't it? But, according to a report …


This topic is closed for new posts.
  1. Chris Wood

    Good for devices with non-servicable batteries

    I can see how this would be very useful in a device where the battery is never/rarely removed (say, an iPod or mobile phone), as the device would always have a good idea of the initial state of the battery.

    It seems this innovation still wouldn't solve the annoyance of being told half-way through your Wii sports session that the Wiimote batteries you thought were full are in fact empty though. At least the readout would be slightly more accurate, but I assume the estimates would depend on the device knowing what type of AA battery (alkaline or otherwise) you'd popped in.

  2. Silas Humphreys

    And about time too!

    Now if they could just do something about the conspiracy to render older devices useless by discontinuing batteries for them, then I'd REALLY be impressed...

  3. daniel


    About frigging time!

    Voltage depletion is a known phenomena, aka "memory effect" (which is not in reality somthing different): The battery can output (for example) 1.2 volts, and the power management system cries foul at 1 volt. This is ok, but as the battery ages, the battery voltage will drop from 1.2 to 0.9 in 5-10 minutes, then sustain 0.95 volts for the next hour... and this is within the power supply criteria for the device being powered.. but too late, beep beep, low batt and shutdown...

    I remember playing with the batteries on my first luggable laptop (386 SX 16 with 2 mb of Ram and 40 mb hard disk...): After a full charge, the portable would die with a low-batt warning after 5-7 minutes, but I could run both filaments in a 50 watt H4 headlight bulb hard-wired into the battery terminals for a further 20 minutes before starting to noticably dim, then 50 minutes dim but usable light (for reading in my bedroom), and 20 more minutes as a red glowing filament before fully empying the battery. With foresight, I probably helped destroy battery this way (deep discharge is not a good idea), but it seemed a good idea at the time to fully discharge the cells - and after playing this game for a few cycles, I managed to get a 30-40 minutes of use out of the battery pack, as opposed to the 5-7 minutes runtime beforehand...

    So, about time that someone made a decent battery sensor, and now all is missing is some decent power management software to go with it.

  4. Anonymous Coward
    Anonymous Coward


    sorry for being slow here but isn't this something Sony's Camcorders have done for years?

    "Built-in Microprocessor communicates with camera, and accurately indicates remaining battery time in minutes. "

  5. Chris Purcell

    Current versus impedance?

    Someone check my physics here. A battery gives direct, not alternating, current; hence, measuring impedance is just measuring resistance. The current flowing through a circuit is voltage divided by resistance.

    So how is measuring the current and voltage any different from measuring the impedance and voltage? (Aside from the method you use, of course - I'm referring to the tangible benefits.)

  6. Anonymous Coward
    Anonymous Coward

    @daniel - deep discharge

    deep discarge is a Good Thing for Ni-Cd cells, which have a significant memory effect... so go ahead if you're still using that '386

  7. G R Goslin

    deep discharge?

    H,mmm. Deep discharge of Ni-Cad BATTERIES, is definitely a bad idea. The capacity of the cells in such a battery tend to vary somewhat in capacity, and on deep discharge at least one of them will be driven into reverse charging by the remaining cells. The back charged cell will be quickly destroyed. This was the usual fate of early portable (?) PC's. the owner was determined to get the last scrap of power from the thing and was soon complaining that the battery life had dropped to next to nothing. The manufacturers replied to this criticism, by arranging to cut off the power before the voltage dropped to the point where any one of the cells output dropped to zero voltage.

  8. Mike Banahan

    @Chris Purcell

    Measuring the impedance of the battery usually means measuring not the voltage and current (which gives the impedance of the load, not the battery) but the internal resistance of the battery which tends to rise as it discharges. So you would watch the drop between open-circuit and on-load voltage and combining that with knowledge of the characteristics of the battery, divine more about the state of charge than you could know otherwise. To be really smart, that could be watched with time and the battery characteristics learned by observation rather than programmed-in.

    And yes, with DC, impedance is the same as resistance but since with AC it's not, it's more generic (and saves thinking) to talk about impedance unless you want to be very specific.

  9. Andy

    What's the big deal?

    My laptop batteries - from Macs, but I assume the same applies Windows-side - have had perfect charge readings for at least half a decade. I've always assumed that the fact phones don't show such reliable readings was due to the manufacturers cheaping out on components, not because the tech wasn't there.

  10. Brian

    Philips Shaver

    My Philips shaver tell me exactly in minutes how long the battery has left. I've always wanted a phone that did the same.

  11. Arclight

    @GR Goslin

    We had this debate years ago at work,two of us bought identical new Bosch cordless drills. The place I bought mine from advised deep discharge before a recharge, my mate was told not to. 5 years later mines still going strong (batteries are recharged on average twice a day), and outlasted the drill itself. His gave up the ghost after two years. Most likely its because you don't actually get a total discharge, but the memory effect on ni-cd's will kill them quicker

  12. George Grenley

    Impedance answer

    Impedance, in a DC world, is resistance, yes. The exact internal resistance, often called source impedance, is a pretty good measure of the state of charge of a particular type and size of battery.

    This new method won't be perfect, but it has good potential to be a lot better.

  13. Charles Manning

    Nothing really new here

    Proper coulomb counting technology has been around for at least ten years and is commonly used in most laptops, PDAs, high end phones etc. These do require good testing and calibration firmware but unfortunately this is often not done well.

  14. Smokey Joe


    A phone won't ever get as close as a phillips. Not even an iPhone. Don't believe the hype!

  15. Anonymous Coward
    Anonymous Coward

    Yes, but....

    ....will it work with the NoPoPo pee-powered batteries?

  16. Anonymous Coward
    Anonymous Coward


    One could model a cell or a battery of cells as a perfect constant voltage source with a variable series resister and large parallel resister. The series resistor value get larger as more power drawn, or as the state of charge drops. The varying resistance, then, would be called the 'impedance'. I suspect TI's trick is to asses the instantaneous slope of the function and relate that to "remaining time".

  17. Daniel

    deep discharge and the memory effect

    First off, there is a difference between "deep discharge" and "drained dead." If you drain a ni-cd cell completey dead, you are decreasing it's life, Not a lot, but the effect is cumulative. Worse yet, if you have many cells in series forming a higher voltage string, one cell is likely to drain off before the others, and may develop a negative voltage ("cell reversal"). This can kill a cell very quickly (you actually boil the magic gasses out of the emergency vent), and that's that - you now have a dead cell, and a useless battery that will soon be dead if you don't do something about the dead cell (usually uneconomical for small handheld batteries).

    However, if you have control circuitry which will cut power before this happens, a deep discharge is recommended by many ni-cd (formerly most/all) manufacturers in order to prevent memory effects. However, it should be noted that true "ni-cd memory" is actually a very rare phenomenon, only seen when a cell is consistently used within a very small percentage of it's capacity. Far more common, and often mistakenly identified as "ni-cd memory" is simple voltage depression due to overcharging and age. This is the same cell aging issue mentioned in this article, and not one that is limited to ni-cd chemistry. Contributing to the confusion is that deep discharge and recharge may help recover some of that lost capacity, so the "folk remedy" for ni-cd memory is the same as a real remedy for the far more common problem of voltage depression.

    Finally, ni-cd cells that sit a long time can form conductive crystals called dendrits internally that eventually short out the electrodes. Regular drain and charge cycles will dramatically slow dendrite growth. Once dendrites have reached a certain state, they can be sometimes be "cleared" by hard charging the cell, but it's a temporary fix at best, as the crystal forms again fairly quickly. A cell showing signs of dendrite growth (failure to hold a charge for) should be considered EOL or near EOL even if temporarily recovered.

    In summary:

    - Ni-cd memory might as well be a myth for the typical consumer.

    - Voltage depression, however, is real.

    - Deep discharge can be good, but isn't needed everyday

    - Drained completely is BAD

    - Regular use is good (@Arclight, I'm looking at you)

    - Once damaged, there is no real way to fix a cell

  18. Robert Jenkins

    Memory effect

    Nicad 'Memory Effect' though misunderstood is very real.

    It's another side effect of the cold crystallisation that occurs in NiCd cells.

    If a cell is rarely discharged below a certain point, the deeper parts of the plates gradualy change from amorphous to crystalline metal.

    If you then try to discharge the cell at anything like high current, it appears to go flat at it's 'usual' discharge level, as the crystalline metal is much less reactive than amorphous and the internal resistance of the cell gets much higher.

    In this case, a deep discharge (2.2 ohm resistor across the cell & leave for a day or two) will completely recover the capacity of the cell.

    Note that the 'dendrite' crystals are both prevented and destroyed by long-term trickle charging, if you can get the cell charged in the first place (high current 'zaps' help).

  19. Alex

    Carbon Footprint ?

    So if I am reading all the techie bits right it would seem that keeping devices on charge and running them low once in a while is the best option if you want to maximise battery life.

    My question is therefore:

    'How does this affect the current push to reduce Carbon Footprints'?

    I mean, is it better to keep things charging and therefore use more electricity or to let batteries die more often and therefore have to spend more energy on producing new ones (and recycling old ones)?

  20. Stuart Halliday

    Old hat

    Getting the power chip to measure Impedance of the battery is not new. My company has been developing a battery powered device with one of these chips for at least a year now.

    It can also detect the type of battery attached and adjust the battery charging cycle accordingly. So it can take alkaline, ni-cad ni-mh, etc.

    The technology is there and available for hardware developers, if only they'd use it.


  21. Steve

    laptops and mobiles

    Battery indication reliability isn’t so much of a problem for laptops/camcorders as it is for mobiles or MP3 players. The former don’t have significantly fluctuating loads, but the latter do (paused, display lamps off, not in a call). Hence a typical charge remaining indicator can easily be fooled when employed in those devices, but they are working in favourable conditions when used for laptops/camcorders so here they are of course more reliable.

    The better indicators store the previous loaded voltage level so you are not fooled by the EMF when you first activate the device.

    I wonder if this new system will give an accurate assessment of remaining charge without itself significantly draining the battery – well you can’t measure the impedance without taking current.

  22. Daniel Ballado-Torres

    Battery memory

    Hm... at least my HP Pavilion laptop battery suffered from the same evil, my solution was to use Linux, as Windows Power management would insist on shutting down at 5% no matter what I put on the settings. Which sucked when the battery started to die on me: I had about half an hour of actual charge, but XP insisted on hibernating.

    Of course, now that charge is in the order of 10 minutes ... or less. APIC shows me only two values: 100% or 0%.

  23. Parax


    "A phone won't ever get as close as a phillips." Not even a Motorola Razor?


  24. Anonymous Coward
    Anonymous Coward

    who uses NiCd batteries?

    Just how many devices use Ni-Cads these days? Manufacturers have gone over to Li-Ion mainly fpr safety reasons but also for density reasons (and possibly for replacement battery sales!)

    Most devices use Li-Ion or NiMH. These have very different characteristics.

    The biggest is that of internal resistance. This gets larger with age (and temperature) and is a big problem in high current applications such as laptops. They have no memory effect.

    Even though an old Li-Ion battery is fully charged, it cannot deliver enough current due to high internal resistance. This gives very misleading results for remaining time and is a big problem for laptop smart batteries that are over a year old.

This topic is closed for new posts.