Dealing with untethered mobile/wireless batteries

The convenience of an untethered world of mobile phones, personal wearable devices, and the many wireless sensors and controllers in our homes comes at a price: constant attention to and management of the rechargeable batteries that power them.

 

There have been recent proposals for providing other means of maintaining their operation such as super-capacitors, energy harvesting of radio waves and vibrational energy from the ambient environment. But such alternatives alone will not be sufficient. To ensure a constant source of power, rechargeable batteries (currently based on lithium ion cells) will be necessary.

Dependent as we have been on rechargeable batteries for more than a decade, it has always puzzled me that a common device- and protocol- independent standard has not emerged.

 

Until relatively rechargeable battery management has have been barebones. A typical low cost relatively dumb battery scheme often consists of no more than identification resistors connected to the battery pack terminals. Alternately the pack might include a resistive temperature sensor next to the power supply connectors. The measured value of the pull-down resistor indicates the capacity and chemistry information.

 

There are many so-called smart battery management schemes available. But they are either company, industry, or application specific, despite the fact that they all use the same battery types, sizes, and chemistry. The problem will only get worse as we move from a world in which there are almost eight billion mobile phone users to one in which there will be tens and perhaps hundreds of billions of battery dependent wearable wireless Internet of Things devices. This does not count the already existing consumer products that require dependable battery management.

 

My candidate for a standard cross-platform battery management is the Battery Inferface (BIF) specification developed by the MIPI Alliance. MIPI’s 250-company membership has initially targeted BIF at mobile phone and computing devices. Although it has taken longer than I expected to be widely used in mobile devices, I have no doubt BIF it will come to dominate not only there but more broadly as well. There are two reasons for my optimism.

 

One is that the MIPI Alliance’s BIF working group did not try to encompass every aspect of rechargeable battery management. Instead, it limited the focus to how the battery subsystem communicates with the rest of the device within which it resides, which at present is mobile phones.

 

But because it deals with only the hardware and software aspects of the communications interface and nothing particularly device specific, it looks like a good candidate for more power constrained wireless platforms such as wearables and other consumer IoT devices. And because it does not try to do more than is necessary (a problem with many industry standards) BIF offers the promise of also being adaptable to other battery chemistries beyond the lithium-ion currently used on mobile phones and other wireless devices.

 

On the hardware side, BIF is about as minimalistic as you can get, with hardware transceivers that can be implemented in as few as 1k gates using a typical non-bleeding edge CMOS process. This is small enough to be easily incorporated into either mobile phone power management ICs (PMIC) or the digital baseband ICs (BB).

 

Although its small gate count will make it an excellent candidate for IoTs and wearables, also counting in its favor is that the spec adds only a single wire, the battery communications line (BCL), to the two already existing power connectors, VBAT and GND, in a typical mobile phone. Over a single BCL line, the BIF communications protocol has been designed to deliver all the signaling required for a range of management functions: battery presence detection, analog battery identification, as well as a delivery of data, address and command words, inband interrupts, and power-save wake up commands.

 

The second reason I think BIF has what it takes to establish itself more broadly is the software management scheme the working group came up with. To manage all that functionality being sent over one pin connection, they’ve defined an algorithm that allows a developer to define a set of battery charging rules that can be stored in no more than 64k bytes of addressable RAM, ROM, or reprogrammable nonvolatile memory for each slave device in a system. In addition to some generic software drivers, this scheme allows a developer to include functions in the rules-based algorithm for host charging control is stored in a prioritized order specific to the requirements of the application.

 

But even with such a scheme available, there are a lot of low cost “dumb” analog batteries out there that have to be taken into account and identified. The BIF working group has taken that into account by incorporating a pull down resistor connected between the BCL and GND, allowing a BIF based battery subsystem to identify whether the battery is a smart or low cost type, as well as identify the electrical characteristics for a low cost battery.