• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 84 subscribers
  • Views 3734 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.

TMS320F28335 ADC Internal Reference Question

stevenh
Guru 20045 points
Other Parts Discussed in Thread: TMS320F28335


Hello,

Section 4.7 of the TMS320F28335 datasheet states the digital ADC value is based on 3 Volts, i.e. digital value = 4096*(Input Analog Voltage- ADCLO)/3. 

Howerver, according to Note 6 in the ADC Electrical Characteristics (Page 169), the TMS320F28335 uses the difference between ADCREFP and ADCREFM as its reference, which is 0.75 Volts.

What voltage is the TMS320F28335 ADC peripheral using as an internal referece?

Thanks,

Steven

  • 0
    TI__Genius 16281 points

    Steven- we don't publish the details of the internal bandgap but the full scale voltage is indeed 3.0V.  This is the case whether internal reference is used or whether one of the three external reference options are used, namely 1.024V, 1.5V, or 2.048V.  I hope this clears it up for you.

    Regards, Joe

  • 0 in reply to Joe Prushing
    Guru 20045 points

    How can the full scale voltage be 3V when the reference is not 3V? 

    Also, can you please explain Note 6.

    Thanks,
    Steven

  • 0 in reply to stevenh
    TI__Genius 16281 points

    Steven-

    stevenh said:

    How can the full scale voltage be 3V when the reference is not 3V? 

    The sampled voltage is attenuated before being passed on to the converter from the S/H, so the scaled down input uses a smaller reference voltage.  The digital output code still represents an analog voltage with respect to a 3V full scale voltage.

    stevenh said:

    Also, can you please explain Note 6.

    The basic point of this note is to stipulate that the DS spec for internal reference gain error is the max system gain error since the reference is generated internally and the error due to temperature drift is included in the spec.  On the other hand, the DS spec for external reference gain error only comprehends the gain error of the F28335, so the user needs to evaluate and include error from the external reference when calculating total system gain error.  If I still have not answered your question, please be specific about what confuses you in that note.

    Regards, Joe

  • 0 in reply to Joe Prushing
    Guru 20045 points

    Thanks for the information.

    The sampled voltage is attenuated before being passed on to the converter from the S/H, so the scaled down input uses a smaller reference voltage.  The digital output code still represents an analog voltage with respect to a 3V full scale voltage.

    So the sampled voltage is scaled down to the smaller reference voltage (0.75V?) and then scaled up to 3V?  The ADC seems to have relatively large errors.  Are they caused by the scaling? 

    Doesn't note 6 have two parts?:

    first part:
    "A single internal/external band gap reference sources both ADCREFP and ADCREFM signals, and hence, these voltages track
    together. The ADC converter uses the difference between these two as its reference."

    I thought this part was essentially saying the internal reference is 0.75V (as stated in the spec:Voltage difference, ADCREFP - ADCREFM = 0.75V typical)

    Second part:
    "The total gain error listed for the internal reference is inclusive of the movement of the internal bandgap over temperature. Gain error over temperature for the external reference option will depend on the temperature profile of the source used.

    The second part is essentially what you said.  I was more interested in the first part when I first posted my question.

    Steven

  • 0 in reply to stevenh
    TI__Genius 16281 points

    Steven- The input is scaled to match the reference range but the voltage does not need to be scaled back up as there is no analog output.  This pipelined ADC might have larger errors compared to 12-bit SAR designs but it is a tradeoff for the faster throughput (12.5MSPS).  Also keep in mind those are worst case numbers across the full operating range of the device (voltage and temperature).  There are multiple sources of error in this type of architecture, one of which is the S/H amplifier but not necessarily due to scaling down the voltage.  Incidentally, one advantage of scaling the voltage is that the converting part of the ADC can operate off the core voltage saving you power.

    You don't have to be concerned about the actual reference voltage as it is not needed to calculate the theoretical code for a given input voltage, but certainly the delta measured between REFP and REFM is a reflection of the internal reference.  Also adhering to the design guidelines with respect to the caps on REFP and REFM and the resistor on RESEXT is important for proper operation. (http://www.ti.com/lit/pdf/spraas1)

    Regards, Joe

    PS: For some reason I am having trouble getting the above link to work directly from the post, so you may need to copy and paste the URL, sorry!

  • 0 in reply to Joe Prushing
    Intellectual 440 points

    I have a question on the ADC internal reference accuracy.

    What's the accuracy of the internal reference voltage? The datasheet says the gain error +/-30LSB, can I understand the ADC internal reference voltage accuracy =  +/-30LSB plus the temperature drift 5ppm/C.

    This question is about TMS320F28335 processor.