This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

A business opportunity killer

Other Parts Discussed in Thread: Z-STACK, CC2538, CC3200, CC2630

More than one year ago I have migrated all my projects from AVR32 to TI ARM platforms. The reasons for this selection were both, tools and devices, namely CCS, TI-RTOS Tiva and SimpleLink devices. I have created building blocks for USB, Ethernet and WiFi applications porting my framework for those devices. I have invested many man-months to integrate my framework with TI-RTOS, bought expensive debugger with trace and many development kits. The WiFi module is certified - life is good.. I have now promising business opportunity for high quantity of small sensor devices requiring self-healing mesh network. I have seen the TI videos about the wireless solution so I was confident it will be an easy ride, similarly to the WiFi module. The reality hit me as a wall.. I cannot use the CCS/TI-RTOS, because the Z-Stack is available only for IAR  Workbench, or 6LoWPAN uses Contiki and there is no certified device module. This is an opportunity killer.  Obviously I can buy certified ZigBee modules from third parties, use them as the wireless modem for my MCU, but it will put me out of competitive price range - I need to run my FW on the wireless module. I would have to buy the $5k IAR Workbench just for one device. I have heard the explanation about more efficient IAR compiler, and yes Atmel was in similar situation, but with introduction of new devices with more memory this went away and because the availability of the full source code, I could modify and customize their stack to fit my needs without the IAR environment. The excuse that the Z-Stack is using some IAR libraries does not seem to be valid, as all the projects and libraries for the Tiva and SimpleLink can be developed using bothh the CCS and IAR studios. I can see the frustration of many loyal TI developers, and would expect TI to make a decision to port the Z-Stack to CCS, to be on the same level as the SimpleLink devices. Also a move to certify the CC2538 as a module should generate enough of business to justify it. Otherwise I do not understand the TI's strategy for the ZigBee devices in the IoT market. Please give me some positive news, so I can go ahead with my project.

Thank you,
Dalibor

  • As I know, there is no specific plan from TI to support CCS to build Z-Stack. TI employees won't reveal their roadmap on E2E forum. I would suggest you to contact TI local representative to get related information.
  • I know and this is from my point of view a really big TI's strategic mistake and the E2E community should push for rectification. Considering the smart building blocks as Tiva Ethernet solution and CC3200 Wi-Fi certified module, where the communication stack sitting in the ROM is TI-RTOS and CCS compatible, suddenly for ZigBee there is a void and you can see the frustration of community members, forcing them to go a different way than TI. And the  problem is simple, just replace some IAR library functions, I cannot imagine which ones, as there was no IAR dependency in Atmel's ZigBee stack, just the devices memory limitations initially. The CC2538 has plenty of resources to tolerate slightly less efficient compiler. I doubt it is a bigger task for TI's guru than 1-2 days. I would be even willing to try myself for free if they promise that the CC2538 certified module is coming, so I can build my self-healing mesh network with it with a competitive cost...

  • I agree. IAR is a POS. The fact that specific versions (yes, point releases) must be used for specific Z-Stack builds is insane. Z-Stack and TI SoC is good quality,  but I'll avoid anything that depends on IAR in the future.

  • I agree with you as well.

    It is my impression that TI doesn't do any major SW development and IAR does it for TI. Then, TI forces its customer to buy IAR license which increases the cost of development significantly.

    TI is good at creating technology, but too bad to document and provide support. We use MSP430 energy library v2.0 for an e-meter, TI's CCS release can't be even compiled because of too many syntax mistakes :) and TI still doesn't care about it. CC3200 can't be used at crowded networks, it malfunctions, TI support team just ignores when you point a problem at a product. CC26xx is very bad supported in terms of documentation and available sources. TI's products pages are also very bad, many application notes and related nodes are not linked to the products.

    90% of the support would be decreased if Yikai Chen leaves the forum :) It wouldn't hurt TI to employee and dedicate 40-50 engineers in the support, I believe that TI would earn more if it provides a better support mechanism for its products. Hopefully, TI would employee better managers that would improve the support mechanism.
  • Hi,

    I am going to address the CCS vs IAR question for 802.15.4 products in this forum first.

    We appreciate the feedback from the community, and we are receptive of it. One of the best instruments the e2e portal gives the community is to provide it to us so we can improve our offering and the best experience for our customers.

    Historically our 802.15.4 products have been running on 8051 core based products, and IAR has been the IDE and toolchain used. As such because of legacy reasons and the existing customer base familiarity with existing tools, the full addition of CCS as well is therefore taking some time. 

    I will forward the generic feedback about documentation gaps to the respective teams in MSP430 and CC3200; what I would recommend and what would help us if you can go to the next level and help providing some specific examples about it in the respective forums, so actions can be taken accordingly.

    Thanks,

    TheDarkSide  

  • For CC2xxx, I think people (especially new to TI's products ) would appreciate to have product related App notes ( e.g., e2e.ti.com/.../application-notes  and https://e2e.ti.com/support/wireless_connectivity/w/design_notes/design-notes-section ) on a product page. CC2538 have couple of them, but more could be added. In addition, " design check list" for the custom HWs would be nice to have, as others provide.

    It would shorten the HW development time, if it would be possible to access the reference design files as a Eagle CAD project and in other commonly used EDA programs' format.

    processors.wiki.ti.com/.../Contiki-6LOWPAN is not a good for a starter. maybe, TI could copy and improve github.com/.../wiki which is more educational.

    http://processors.wiki.ti.com/index.php/CC26xx_Optimal_Load_Impedance should be linked to the CC2630/50's product pages.



    I am looking forward to download IAR release of Contiki, is it going to be published this week?

  • Just get information that TI would release wiki for Contiki project on IAR by end of this week.
  • Thank you for sharing the information

  • You are welcome.
  • Again, for me and several other E2E fellows, if this news does not mean that it will run on TI-RTOS and is portable to CCS, it was not effort in the right direction. What is the reality?

    Thank you,

    Dalibor

  • I use both TI and Silicon Labs Zigbee solution and they all need IAR to build the project. I think if TI can port Z-Stack to CCS, that's would be more competitive.
  • Freescale uses IAR also, but NXP seems eclipse (GCC) based.
  • Yikai, do you have more information if it is going to be released today or next week?
  • Not get any update until now. I think it might be delayed.... : (