• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 78 subscribers
  • Views 712 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.

When you use MSP430i2041, frequency accuracy at room temperature is better with internal resistor rather than external 0.1% 20k ohm resistor.

Koyo Kegasa
Expert 1260 points
Other Parts Discussed in Thread: MSP430I2041

I received the samples of the MSP430i2041 this year.  Thank you very much!

I measured the oscillation frequency of DCO clock using spectrum analyzer. 

When I chose internal resistor, DCO frequency was 16.3890 MHz. 

Frequency error = (16.3890-16.3840)/16.3840*100 = 0.03 % = 300ppm.   Extremely small error!  Lower than the CO2 level in the atmosphere. 

When I chose external resistor ( measured resistance is 20.000 k ohm ), frequency was 16.365 MHz. 

Frequency error = (16.365 -16.384)/16.384*100 = -0.116 %.  Very small.  But, a little bigger than DCO with internal resistor.

So, I think  "If you want to get the smallest frequency error at the room temperature, I should use the internal resistor mode for DCO."

I think, the parasitic capacitance of the chip resistor (including the printed circuit board pad and wire capacitance) made the DCO frequency lower.

Is it possible to add or subtract some tuning_value to the CSERFCAL to get more accurate frequency ( to compensate the parasitic capacitance ) ?

CSERFCAL = *(TLV_address_for_parse + TLV_CAL_CSERFCAL) + tuning_value_1;

  • 0
    Guru 74080 points
    In your specific case: Maybe.
    Look at the datasheet at page 16. Using the internal resistor, the DCO has a maximum absolute tolerance of 0.9% at 3V! Depending on your device if it was manufactured on monday or on wednesday. Using the external resistor the maximum error is 0.25% (20k, 0.1%). And the drift is a tenth of the internal resistor (if used a type that has 50ppm). The supply voltage drift is the same.

    So if you can ensure that your temperature stays the same all the time and you are building up one single circuit where the frequency is proved to have such a less error, then do that. But if you want to start a mass production then the external one should be preferred.
  • 0 in reply to Dennis Eichmann
    Expert 1260 points
    MSP430i2041 has calibration data for DCO with internal resistor.
    Maybe, calbration value is made at specific temperature ( such as 25 degrees C ).
    So, I think that, If you use the MSP430i2041 at the same temperature and at the same Voltage, you will get extremely accurate frequency.
    As CSIRFCAL is 8 bit data. If the frequency error correction step is 300ppm, you can cover +-3.84% error this way.
    And you can get the +-300ppm accuracy.
    Am I right?
    At what temperature the calibration is done?
    At what voltage the calibration is done?
    I want to know it!

    One more thing.
    MSP430i2041 has calibration data CSIRTCAL.
    I think this one is to compensate for temperature coefficient.
    MSP430i2041 has internal temperature sensor. It knows the temperature it is working in.
    So If you wish, you can measure the temperature, and correct the DCO frequency using the CSIRTCAL data accordingly.
    Am I right?

**Attention** This is a public forum