The top Bluetooth evangelist Nick Hunn states in his blog "For Bluetooth low energy it should be easy. One of the clever aspects of the standard is that the Bluetooth low energy technology is incorporated into the next generation of normal Bluetooth chips, so an increasing number of mobile phones will incorporate it for free. That gives a critical mass of hundreds of millions of handsets that can act as gateways or displays for a new generation of products.... Bluetooth low energy will gain market traction." "Because of its integration into handsets, Bluetooth low energy is assured of success in reaching critical mass."
What he forgets is that the industry context of Bluetooth Low Energy is totally different to the context when Bluetooth Classic was first to the market.
With the introduction Bluetooth Low Energy, Bluetooth chips have split into two groups: dual mode and single mode. The dual mode Bluetooth chip is for cell phones and note books. It supports both Bluetooth classic and Bluetooth Low Energy, whereas the single mode chip is only for sensors. The table below lists semiconductor companies that have announced their Bluetooth Low Energy products.
| Single Mode | Dual Mode | |
| Broadcom | x | |
| CSR | x | x |
| EM Micro | x | |
| Nordic Semiconductor | x | |
| ST-Ericsson | x | |
| TI | x | x |
There are only two semiconductor companies announced doing both, TI and CSR. CSR announced dual mode first, then early this April it announced the plan of single mode. But besides Bluecore 7, the dual mode architecture, I have not found anything more publicly about their single mode plan. TI discloses more of their BLE work through their blog. TI's single and dual mode chips are from two different divisions, Low Power RF and Wireless Connectivity Solutions. The Low Power RF organization is the acquired company Chipcon in Norway.
What this tells us from outside is that dual mode and single mode are two different animals, that requires different semiconductor expertise. Dual mode cares less about the power consumption, as it will be powered by lithium Ion rechargeable batteries. Single mode is all about coin cell battery operation for at least 1 year. Looking at the chip architecture, single mode BLE part is an SoC with full BLE stack implementation; while the dual mode part is cut at the Host Controller Interface, with the upper Bluetooth layers and applications deployed on the cellphone hosts or the notebooks' CPU. So technically, the single mode part is technically a more challenging endeavor. Karl Torvmark of TI gives a very clear description of this architecture difference in between BLE dual mode and single mode chips in his blog What is 'Sing-chip"
By realizing this fact, we can challenge the statement that cell phones provide the critical mass to Bluetooth Low Energy and assure its success.
Let's first review another fact not much in favor of Bluetooth - the ~20% attachment rate world wide, which has been reached after ten years span. Attach rate in simple explanation is the percentage of cell phones whose Bluetooth are in use (linked to an earphone or connected to PC...). Bluetooth organization has given itself a targeting or forecasted attach rate 8% for the first year of BLE. It remains to be seen for how long Bluetooth Low Energy can reach this goal.
Though BLE's targeting markets range from sports, to medical, to homes, even to industrial application, the low hanging fruits are the personal sports sensors (heart rate monitors, speed and distance sensors) already conquered by other radios like ANT, Nike+ and the legacy 5KHz analog systems. In the trend that digital sensing to replace analog, ANT and Nike+ are far ahead, neck to neck with approximate 5M each in shipping volume shipment in recent 2-3 years.
Do not look down this 10M volume. The usage model of ANT and Nike+ is basally one on one and multiple on one, which I mean 1 sensor actively connecting to 1 hub (a watch, a dangle ...) or multiple sensors to 1 hub. People buy them in pairs or in package. The model of Bluetooth is to pre-invest the hub device base, and hoping companies will be attracted to develop accessories. In the classic Bluetooth world, this model works perfect and survived and grew companies like CSR from start-up. The same business model will not work as easy in the market established by others but new to Bluetooth Low Energy.
Companies making single mode chips are not fooled by the Bluetooth cell phone base illusion. They are very clear about the up and downs of BLE and its competitors. Nordic semiconductor, the company providing the low power expertise to Bluetooth SIG, is actively selling three solutions, BLE, ANT and their own wireless desktop protocol. They never say they will become a pure player of BLE. They are now the vendor of both ANT and Nike+, as well as Logitech in wireless mouse and keyboard market. In the past July, a new generation of ANT chip nRF24AP2 family were announced.
TI is becoming more and more a protocol agnostic company. In the low power RF arena, TI has both 802.15.4, the Bluetooth and sub- 1GHz radios to support protocols like Zigbee, Bluetooth, WirelessHART, SimpliciTI, Sensium etc. TI has joined ANT+ alliance. I will not be surprised if one day TI announced an ANT chip family.
Bluetooth SIG may be very regretful to take a lot of inputs from Nordic in defining the BLE physical layer, that is not the same as the competing ANT and Nike+, but still too close. Too close that these single mode chip companies can design one semiconductor architecture to serve both the BLE and the competing technologies, putting eggs in more than one baskets. So what? So Nordic and TI can sell the chips no matter their customers want to develop BLE, or ANT or Nike+ sensors. So the presumed scale of economics advantage of BLE chip will be taken away by the competitors, which will be cheaper, at least not more expensive than BLE.
Nick is correct that cell phone access will provide Bluetooth Low Energy the traction, but cell phone is unable to offer the critical mass. And.... the cell phone traction may not as solid as he or people think.
Keep coming back for more discussions. Welcome comments.
Nearly 12 months on from your original post and Bluetooth Low Energy has been adopted. I should first declare an interest - I work for CSR..
ReplyDeleteYou make some interesting points on Bluetooth Low Energy, and I think that you are correct that the companies involved should not take the market for granted.
Bluetooth Low Energy Single mode devices are now availble from multiple vendors and they do offer some significant comercial benefits over the incumbent proprietary technologies.
The Bluetooth standard is not owned by a single company - it is an open standard. This has two nbenefits. One - There is no per unit licence fee. Two - There is no single company controlling the market. So there is no comercial risk for competeing companies using the standard(Ant is from Dynastream who are owned by Garmin who produce fitness products)
CSR single mode devices can act as master and slave, and this means that you can get the power saving in both the sensor and the watch or cycle computer. This makes a single mode device perfect for existing health and fitness products.
Most Bluetooth silicon will transition from Bluetooth classic only to Dual mode in 2011. So Nick has a strong point for the other market that you mention - PC keyboards and mice. Bluetooth low Energy single mode devices will soon provide keyboard and mice direct to the PC without a dongle (probably shipped with a dongle at first)
Longer term the cell phone access will probvide traction for other markets, but Bluetooth Low Energy will stand alone as a good choice in the end to end products you describe.
Proprietary solution get chrushed by an adopted standard sooner or later, period.
ReplyDeleteIf any sensor will work with any hub then this will win.
Hubs that have energy capacity do not need the BLE to operate but rather to support slave devices that do need low energy transmit and recieve. s even if a speedometer is a hub and the sensors are the slaves I would still want to be able to upload data to my cell phone to look at on a PC (or the phone)at a later time.
People that buy this stuff typically are training or plan to. This requires looking at the data to see ones performance level to make adjustments to optimise performance.
new2coding
16 months on, TI has announced ANT chips and several of the dual and single BT chips are on the market. But no cellphones with BLE yet. You may be right.
ReplyDelete