• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 2 answers
  • Subscribers 82 subscribers
  • Views 1194 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TMS320F28235: Flash erase time

Jordan Zhou
Genius 10040 points
Part Number: TMS320F28235
Other Parts Discussed in Thread: C2000WARE

Hi,

Customer is using F28335 for their industrial applications. Now they find that, some pieces F28335's Flash erase time will increase from "less 10s" to "over 20s", after one month filed working. We checked that, the Flash still worked (like reading, writing) in an stable environment (like power supply, temperature 25C and others).

We'd like to know two points:

1) Why does it happen? Please provide more info to share with customer. 

2) Is there any way to decrease the erase time from 20s to 10s, again?

Thanks a lot.

Br, Jordan

  • 0
    TI__Mastermind 45046 points

    Hi Jordan,

    There is a plan in place to add maximum flash timings to the F2833x datasheet.

    The erase times that the customer is seeing are for the entire device flash, correct? Not individual sectors.

    1. There are factors that can affect the time it takes to erase the flash, such as operating temperature/conditions and number of prior program/erase cycles.

    2. Not that I'm aware of. Is the customer operating the device at the maximum frequency when performing flash functions?

    Best,

    Kevin

  • 0
    TI__Genius 15385 points

    Hi Jordan,

    Continuous write/erase cycling on Flash will degrade erase time, as the way the data is stored is through electron-hole injection. Do you know how many flash write/erase cycling did it go through in that 1 month. Unfortunately you will not be able to decrease the time on that device, as the device seems to have degraded.

    Typically we see this kind of degradation after >1K of write/erase cycling. 

    Regards,

    Nirav

  • 0 in reply to Kevin Allen18
    TI__Mastermind 45046 points
    Also please see the E2E FAQ on F05 flash: e2e.ti.com/.../757585

    -Kevin
  • 0 in reply to Kevin Allen18
    TI__Genius 10040 points
    Kevin & Nirav,
    Thanks for your comment.
    Some info about Flash erase:
    1) F28335 is not working at max 150MHz when erasing Flash. What's your suggestion when erasing?
    2) During 1 month time, there are erase/write, but I'm not sure how many times. I'll talk with customer to know more.
    Br, Jordan
  • 0 in reply to Jordan Zhou
    TI__Mastermind 45046 points

    Hi Jordan,

    When you say the F28335 is not working when erasing at the max frequency what exactly is happening?

    I'd have the customer review the f28335 flash_api example within C2000ware for how to properly configure the device for erasing/programming at the max operating frequency. Directory location below:

    C:\ti\c2000\C2000Ware_1_00_06_00\libraries\flash_api\f2833x\28335

    There's also a doc in C2000ware that outlines the same steps to take:

    C:\ti\c2000\C2000Ware_1_00_06_00\libraries\flash_api\f2833x\28335\docs

    The biggest things to note are:

    1. Modify Flash2833x_API_Config.h for your targets operating conditions. i.e. un-commenting the proper CPU_RATE

    2. Initializing PLL to proper CPU frequency

    3. Copy the flash API to RAM

    4. Initalize the Flash_CPUScaleFactor variable to SCALE_FACTOR

    Best,

    Kevin

  • 0 in reply to Kevin Allen18
    TI__Genius 10040 points
    Kevin,
    Thanks for your suggestions.
    Currently, we can see that:
    1) When F28335 working in 150MHz, the Flash erase/write costs more time;
    2) No matter Flash API or JTAG operation, Flash costs the same long time during erase/write;
    I'm asking them to collect more info about how they use Flash during, like how many times they erase/write during these months. You will be updated when I get these info.
    Thanks again for your support.
    Br, Jordan
  • 0 in reply to Jordan Zhou
    TI__Mastermind 45046 points

    Jordan,

    For #1, you're seeing flash erases/programs taking longer at 150MHz than at lower frequencies?

    Another thing to check is the number of flash wait-states they have configured, as the minimum required is 5 for 150MHz operation. Please see the table below and the FBANKWAIT register in the Sys Control Guide:

    http://www.ti.com/lit/sprufb0

    OK, that sounds good. Let us know when you can.

    Best,

    Kevin

  • 0 in reply to Kevin Allen18
    TI__Genius 10040 points
    Kevin,
    After another round test with customer, here's the summary:
    1) Over 10 pcs F28335's Flash erase time (total Flash) is over 40s, after 1 year filed running;
    2) Filed temperature is quite good like: 10~35C;
    3) Their application doesn't do any Flash erase/write operation; the issue was reported, just because they updated the firmware and found it;
    4) Replace the F28335 with another one on the same board, the erase time (total Flash) is about 15s, which is normal;
    5) The normal and "un-normal" F28335 current consumption during erasing, are the same;
    6) If the F28335 total Flash erase time is over 40s, each sector's erase time are the same, there's no specific sector costing longer time;
    7) All of the F28335 are erased by both JTAG and customer's bootloader, the erase time are the same.
    Is this normal?
    Customer has the concern on the max erase time. Can you provide me a draft max time?
    What's the detailed logic for Flash erasing by Flash API? Is there any Flash API to identify the erase time?
    Thanks a lot.
    Br, Jordan