• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 51 subscribers
  • Views 497 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.

CC2540 programming fail on 50% of batch

roue
Prodigy 90 points
Other Parts Discussed in Thread: CC2540

Hello,


I designed a module with a CC2540, and launched a pilot run. I found around half of the modules are ok (programming with CC Debugger ok, working ok, BLE peripheral communication ok,... everything ok).

Other 50% have hardware troubles. Programming fails randomly (at verifying). However, when the programming is sucessful (sometimes), the chip seems stuck, the software doesn't start. Sometimes, it starts if I don't disconnect the CC debugger right after the programming. I can see the two Crystals at work (32k and 32M). After reseting the board several time (either by disconnecting the power supply or by playing the reset pin), the chip doesn't start (Crystals offwork). Resetting over and over has no effect. I can restart the board, by programming again (but, it very difficult because I need to try many many times before getting sucessful verify).

There is a TX pin of UART connected to an external FDTI adapter to USB for logs. After the chip start the Crytals (when it can), the logs are partial, or messed, very random.

When the Crystals are started, I have measured the frequencies, they are both ok.

Knowing that 50% of the boards are ok, and others are working randomly, I stuck...

Anyone has a clue on that issue?

  • 0
    TI__Mastermind 32615 points
    Moving it to Bluetooth® Low Energy Forum
  • 0
    TI__Intellectual 1840 points
    Hi,

    Please share your schematic, for us to review and revert.

    Please ensure that both the FTDI and CC254x chip are at 3.3V logic level and that you have connected GND of FTDI to GND of CC254x.

    What make crystals are you using? Please share datasheets.

    We would recommend to use the crystals, that we have tested and validated.

    http://processors.wiki.ti.com/index.php/CC_Crystals

    Regards,

    Vijay
  • 0 in reply to Vijayasarathy Shastri
    Prodigy 90 points

    Hi,

    Please find the requested documents here:

    On this batch, the Xtals are :

    32MHz : TXC reference : 7M-32.000MEEQ-T

    32KHz : Abracon : ABS07-32.768KHZ-T


    The FTDI adapter has the same level as the module (3.3V)

    For the next batch, the Xtals will be from RAKON (TEMEXPRESS) which is better quality than NDK. The datasheets are also in the shared folder. I am waiting for samples. Please check it and let me know any comments.

    Best regards

    Benoit Roué

  • 0 in reply to roue
    TI__Intellectual 1840 points

    Hi,

    My comments after reviewing your schematics.

    1) The DCOUPL pin should be connected to GND through a 1uF cap. But it is also pulled up to Vcc [ DCOUPL pin supplies 1.8V, and should never be tied to Vcc, which would be 3.3V, or used as supply for external circuits]. So cut the trace/track that connects to Vcc.

    2) Reset cap C9, should be 100nF/0.1uF.

    3) The loading capacitance that the 32Mhz crystal, sees should be as per values in the "Crystal" section of the datasheet. Also account for parasitic capacitance[ check Pg8, 32Mhz table of datasheet]

    4) The RF trace should have a solid ground plane below and should not be so close to the crystals. Also the RF trace and the RF antenna should have a "keep put area".

    5) The crystal trace to the OSC Q1,Q2 pins should not be cross or should be as near and straight to the device as possible.

    6) Remove C8 and D2. I think it is not needed.

    Regards,

    Vijay

  • 0 in reply to Vijayasarathy Shastri
    Prodigy 90 points
    Hi Vijay,

    Thank you for your feedback.

    1. You are right, the problem comes from that... After cutting pin 40 from VDD, it works fine. I am so surprised to see that about 50% of the boards can work with 3.3V at this pin...
    2. in the reference design (CC2540EM), this capacitor has a value of 1nF. You said it should be 100nF. which one is correct?
    3. The loading capacitance will be adjusted to make the fine tuning of the frequency as soon as I will get the samples with the final Xtals
    4. The RF trace has a solid ground place below , of course, maybe you didn't see it clear, but it has on "Mid Layer 1". Compared the reference design, the RF trace and the 32MHz Xtal are very similar. In my design, the 32kHz xtal, is much more close to the chip than in the reference design. Yes, the keep out is not too wide, only 15 mils. However, this RF trace is not too long, and I have already checked the performances; it's ok.
    5. Pins should not be cross? I don't get your point. The 32MHz Xtal is as the same location as the reference design. Also the trace are very similar to the reference design.
    6. C8 removed already, D2 is to protect the chip in case of connector inserted with wrong polarity of power supply.

    Best regards

    Benoit