Custom Search

Thursday, October 21, 2010

Bluetooth Low Energy and ANT, who cares?

This post is my read of the interesting press release from TI about their Bluetooth Low Energy and ANT chips (CC2540 and CC257x). I think for anybody closely monitor the progress of BLE and ANT, it is worth investigating this press release.

First of all, did I said this before? TI really does not care about what technologies you choose. What they care about is your choosing their chips. So this is why when we users see Bluetooth Low Energy and ANT are fierce competing rivals, TI has a different view and this view allow TI to announce the two seemingly fierce competing product lines in one release. The key message: our solution is the most completed.
This is TI's unique marketing proposition. There are companies having both dual mode and single mode Bluetooth 4.0 chips, but only TI has also ANT. There are a company that has both BLE single mode and ANT, but it does not have BT dual mode. There are also companies have wifi, BT, GPS multi-comm chips, but again TI is able to add another ANT. Considering other protocols, like 802.15.4 based series, are provided, TI has the most comprehensive solution portfolio for the WSN industry. If I were a customer, I would prefer to consult TI for a neutral inputs of what technology best fit my use case.

This release has reconfirmed that the two product lines, together with 802.15.4 lines are managed by the same low power RF team. This is by all means good to customers. After my double checking of the datasheets, the CC2540 and CC2530 (Zigbee SoC) are indeed as my impression share the same pin-out. When under one organization structure, some negative line competition, for example, resource allocation etc can be avoided or mitigated, and marketing and sales efforts can be more aligned. These are all the advantage that TI customers can take.

Following this thought, I want to prove that if the BLE and ANT are actually based on the same die or the same architecture with tiny differences on things like memory size, I/O port etc. Technically, this should be totally possible as the PHY of BLE and ANT are so similar that the main technical difference could just be the firmware. This is how TI's WiLink part is able to support ANT through a cell phone host software upgrade, isn't? However except for the same packages QFN40, I am not very successful though. There is no CC257x datasheet available yet. Moreover from the information on TI website, there are seemingly more differences than similarities.

Two major technical differences:
  1. The BLE part CC2540 is an SoC, while the ANT CC257x is a network processor.
Network processor is a TI term I believe. It is to describe a chip that has preloaded protocol but no room for application code. So you have to use a host MCU for applications. There is advantages to use network processor, because a product is simple to develop. Application code and protocol are isolated and linked through serial communications, or host controller interface. The drawbacks are the system cost, product size and so on caused by two chips in use.

In fact, all ANT chips from Nordic Semiconductor are actually network processors. Here is my understanding: unless the semiconductor is dual core (one for protocol and one for application), network processor is the only choice for ANT as a proprietary protocol, while CC2540 can use the same core for both protocol and apps. So this difference can not exclude that the die and architecture could be the same.

However, such situation presents a challenge to ANT still. To compete with Bluetooth Low Energy, ANT do need to have SoC in their next generation. With this being said, as simplicity (or KISS or KIDS) is more and more important to developers, BLE SoC may well be with ANT to drive for a dual core architecture. I do not see why TI will not do so. Eventually, the two rivals will again share with the same architecture with less differentiators semiconductor-wise.

It is understandable that some TI guys love to have network processor so to sell more MSP430 MCU. Check their Zigbee SoC and network processor offerings will prove this. Some engineers and companies do prefer to use network processor, so they have the freedom to pickup their preferred and familiar MCU to ensure a simple design and shorter cycle time. Sooner or later, TI will offer BLE network processor too.

Today TI announces BLE SoC and ANT network processor; tomorrow, they will add BLE network processor and ANT Soc into their portfolio.

  1. CC2540 has +97dB link budget while CC257x is at +95dB.

I do not believe the BLE PHY is 2dB better than ANT PHY. (please correct me if I am wrong) From the CC2540 datasheet, its transmission power is upto +4dBm and its receiving sensitivity is -93dBm in high gain mode. This explains the CC2570's link budget.

I have one guess hat may partially explain this difference. ANT is actively using proximity pairing in its fitness and group applications now. And all of the current systems use Nordic radio which has maximum 0dBm transmission power. Allowing 4dBm should cause bad user experience. So the ANT engineers may calculate that 2dBm is still fine. To fully understand the number for CC257x, we probably have to wait until the datasheet published.

(BTW, as aside note, CSR's uEnergy can transmit at +8dBm with receiver sensitivity -92dBm; Nordic's uBlue has 0dBm transmission and EM Mirco's EM9310 4dBm, but neither of the latter publishes sensitivity data.)

PS: I am not sure if the ANT marketing guys have ever involved in creating or reviewing the press release. I guess they were and they were probably happy originally that again ANT bonded with Bluetooth Low Energy. What is frustrating is that some media only pay attention to Bluetooth Low Energy. In a way, some of the articles cited the press release did read like that ANT is part of BLE, or ANT is a sub-brand name of TI's BLE part. ANT guys call for help from ANT fans through its twitter @antplus. I do not think I am a fan of ANT. I am a fan of WSN at most. But I'd like to help to clarify the wrong perception. I'd like to see how smart small ant compete with mammoth with blue tooth. And I wish I could be invited to the next ANT+ symposium just as was @dcrainmakerblog. Ha ha...
Enhanced by Zemanta


  1. Very interesting analysis. One comment: the FSK deviation for ANT is smaller than for BLE, which could explain the 2 dB difference.

  2. I do not know if this is the right forum for my question but I am trying to find out if there is a standard that specifies the behavior of ANT devices. I am currently using Garmin fitness devices that includes a heard rate monitor that interferes with a BlueTooth headset and if possible, how to overcome this issue.
    Peter G

  3. Thanks for reading my blog.

    Regarding to your issue, to my understanding, technically, Bluetooth should not be able to interfere ANT to the degree that catches your notice. Bluetooth is hopping among frequencies while ANT is sitting on one dedicated channel and the signal pulse is in micro seconds. If at one moment, there is a collision between the two, they should be running well apart from each other for the next long period of time.

    I suggest you post your question to and TI's website for an official answer. There are forums there.