CC2541 versus nRF51822

Gentlemen,

I apologize if this is not the correct place to post this question.

I have a potential project to start a BLE project but their requirement would be start a new design with the CC2541.

However, I would like to tell them to use the nRF51822 but I want to present good technical arguments.

Is there anyone here familiar with the CC2541? Please PM me and let me know.

Thanks, Gil

  • We currently use both chips in different aspects of our system, and personally I like the Nordic chip better. When we started our project, the SDK for the nRF51822 wasn't quite as developed as the CC2541, but I think it has come a long way and a long term goal of mine is to eventually replace the TI with the Nordic part.

    Another company in our building was recently going through a side by side comparison of the nRF51822 and CC2541 in a beacon application (simple sleep & advertise), and has decided to go with the Nordic part due to substantially lower power consumption in his testing.

    Another concern if this is a hobby project or you work for a small company is the toolchain cost. To do anything meaningful on the CC2541 you need to buy an IAR Embedded Workbench license which will run you around $4000. Compared to the free Kiel setup you can use with the Nordic part this is a little absurd.

    Finally, the support with Nordic through the DevZone is fantastic. Usually forums are a hole into which you throw your consciousness when you are completely out of any other options, but responses here are unusually helpful with Nordic employees frequently chiming in with helpful information.

    The only drawback to the nRF51822 is it still seems to be changing a bit (for the better) as it matures as a product. They are open about the remaining bugs people have found and most can be worked around pretty easily, you just need to pay attention to which version of the chips you are getting and reading through the most recent Product Anomaly Notice (PAN) to see if anything effects you.

  • I like the first answer to this question and agree with it.

    I also started with the TI chip and moved away from it quite quickly.

    1. The toolchain. TI is totally wedded to IAR, you want to build for that chip, you need IAR and it's expensive, and it's not very good either(*). True that Nordic's primary development tool has been and is Keil, but you can at least get a limited version for free and Nordic does support and supply templates for straight gcc development, which is always free. People have successfully used Eclipse, I use Crossworks (I have the non-commercial license but if I ever needed the commercial one, it's not that expensive) and the latest devkits support MBED too. ARM is clearly working on making MBED better and more standalone, there's a whole community behind it with bucketloads of available source code and libraries to aid development. That is worth time and money.

    2. The TI chip is based on a pretty old processor. That's partly why the toolchain support is as described above. I think the Cortex M0 is a better processor the ARM series is a better place to be going forwards.

    3. Support from Nordic is excellent. Take a look through the support forums here and look at the quality, in-depth, answers from the guys there, Ole, Stefan and the rest. Not every question gets answered, but most of the understandable, interesting ones do :). I've opened a number of support cases as well asking for help ordering kits to help with the underlying sniffer protocol and they've helped me out professionally and rapidly every single time. I don't know how good the TI support is, I never had to use it, but great support can make the difference between a successful and unsuccessful, timely or late, project.

    4. Nordic continues to develop the line. They recently introduced chips with more RAM, which I think is necessary as BTLE devices start to get more complex and we move from BTLE 4.0 to 4.1 and I guess soon 4.2, they are all going to take more on-chip resources. The latest release silicon fixes most of the product anomalies (i.e. silicon bugs), so you're making a choice at quite a good time.

    It does all of course depend on what you're building and your engineering constraints. Obviously you should check specs for power consumption etc, speed, available GPIOs, I/Os etc. but absent any shopstoppers there, I'd never go back to the TI chip.

    (*) - if you look hard enough there is a huge long thread on the TI support forums about the lack of open source tools, it's rumbled on for two years, it's a big issue and lots of people on that thread bailed out and went Nordic at some point.

  • We also started with the TI CC2540/41 Chips. We had a lot of trouble. Then we moved to the nRF51822/422 Chips and this was the best decision. For us the biggest PLUS is the support. The support is just great. The support from TI is not so bad but you cannot compare it with the support from nordic.