• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 84 subscribers
  • Views 536 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.

TMS320F2800137: Internal oscillators accuracy need to be less than 1%.(-30℃~90℃).

Jim Chen (SEM)
Intellectual 2630 points

Part Number: TMS320F2800137
Other Parts Discussed in Thread: C2000WARE

Hi experts,

My customer want to use F2800137, but they need internal oscillators accuracy to be less than 1%.(-30℃~90℃). Can we achieve it by hardware calibration or software calibration？

BR

Jim

  • 0
    TI__Guru* 85290 points

    Jim,

    The specs in the datasheet define what is achievable here, as you noted past -10C to -30C we fall outside the customer requested tolerance.  There is option further down in the DS, by using an external precision resistor on the ExtR pin to reduce the error term; depending on the tolerance of the external resistor this could meet the 1% the customer needs.  I've given a html link to this section of the DS below.

    https://www.ti.com/document-viewer/TMS320F2800137/datasheet/GUID-7F073018-BBAE-47A1-A426-C26AFF50935B#GUID-578F3FA0-2C0C-47CB-935B-C13B05D127F6

    Best,

    Matthew

  • 0 in reply to MatthewPate
    TI__Intellectual 2630 points

    Hi Matthew,

    Thanks for your reply.

    We can't add external resistor to meet 1% tolerance, so can we use correction code of F28035 to correct F2800137 internal oscillators?

    BR,

    Jim

  • 0 in reply to Jim Chen (SEM)
    TI__Guru* 85290 points

    Jim,

    We cannot use same technique from F28035 on the F2800137.  These devices are in different technology nodes, therefore the internal oscillator implementation is different between the 2 devices.  The intent on the F280013x devices is to have one calibration done at boot that covers the operational range of the device.  I'll see if I can get some more background here, but to my knowledge what is in the DS is the best we can achieve over a given operational range.

    Best,

    Matthew

  • 0 in reply to MatthewPate
    TI__Intellectual 2630 points

    Hi Matthew，

    Thanks a lot for your reply. 

    If the error of the crystal oscillator exceeds 1%, it will cause abnormal communication. Is there any way can fix it other than add external resistor?

    It's a big risk for our project if software calibration doesn't work as you said.

    BR,

    Jim

  • 0 in reply to Jim Chen (SEM)
    TI__Expert 6010 points

    Jim,

    To clarify, if using an external crystal oscillator between the X1 and X2 pins, then the accuracy will be in the very tight PPM range of the crystal.

    The table you referenced above is for the zero-pin Internal Oscillator (INTOSC).  The specs are as Matt indicated and there is not any additional calibration to achieve tighter specification.  (exception is if there is another precision clock reference in the system, but that would be unlikely since it could be used as an accurate reference clock input anyway).

    Which communications interface is this?  If it is UART, then we have seen implementations which have robust communication.  If it is CAN then in any case I would recommend using an external crystal.

    Best regards,

    Jason

  • 0 in reply to Jason Whiles
    TI__Intellectual 2630 points

    Hi Jason,

    Thanks for reply! My customer just use UART but considering aging of communication circuits, they need high precision of the internal oscillator.

    It seems they don't want to add external resistor because of cost and they want to test baud rate tune example in C2000ware SDK. Can baud rate tune example meet the communication requirements under large oscillator error maybe after 10 years?

    BR

    Jim

  • 0 in reply to Jim Chen (SEM)
    TI__Expert 6010 points

    Jim,

    Yes, the baud tuning example will work well, compensating for clock difference error from either transmitter or receiver over 10 years or other factors.

    The only consideration is the baud tuning approach does require a training phase on startup where the C2000 device needs to 'listen' to the UART RX pin to adjust baud to match the other device.  This should be doable but does require the system level 'startup communication handshake' of having the other device send data before the bauds get synchronized.  Hope that makes sense.

    Thanks,

    Jason