• State TI Thinks Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 2 answers
  • Subscribers 87 subscribers
  • Views 2899 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.

Piccolo 28035 Internal Reference Voltage for ADC

Global ET
Prodigy 40 points


Hi,

Is there an electrical specificaton or characterization for the accuracy of the internal 3.3V reference on the Piccolo 28035 ADC?  Document reference is spruge5b section 1.9.1, page 18 (preliminary).

Can the ADC be dynamically switched back and forth between the external and internal voltage references while running our control algorithm?  Some of the ADC inputs are absolute voltages and some are ratiometric to VREFHI, so the best accuracy would be accomplished by converting unknown input voltages relative to either the fixed internal bandgap or the external system VREFHI.  In both cases we would have VREFLO tied to system ground.

How does the internal voltage reference work when VDDA is at its lower limit of 2.97V?  Is the bandgap still 3.3v?  (I'm guessing that the bandgap is actually 1.65V, half-scale, and the ADC converts to 2x this reference.)

We are otherwise considering an external bandgap voltage reference (e.g. 2.500V) with 0.2% accuracy and "correcting" the ADC readings to this reference in software.

The documentation is very minimal and we trying to design with accurate ADC inputs.

Thank you.

Global ET (Farmington, MI)

  • 0
    Genius 5170 points

    Do I , as an user, need to know the accuracy of the internal voltage regulater as long as the vendor provides the ADC accuracy over the temperature range? Will I get better accuracy over the temperature range if external reference is provided? If yes, is there a reference documentation describing that relationship? We have utilized the ADC accuracy information for system design and never really thought too much on the accuracy of the internal reference. Are we making unnecessary trade-offs by not taking that information into account during system design? Would appreciate comments from TI folks on this.

  • 0
    TI__Guru* 81560 points

    Please reference the F28035 datasheet for electrical parameters:

    http://focus.ti.com/lit/ds/symlink/tms320f28035.pdf

    You'll want section 6.11.11 for ADC specs.

    The gain error is what will be impacted by the stability of the internal reference, gain error in this case will be bounded by +/-60LSBs at full scale over full temp(-40 to 125), with typ temp coeff of -50ppm/degC.  Gain error for ext ref(assuming ideal ext ref) is +/-40LSBs, with typ temp co of -20ppm/degC.  The -20ppm/degC would be additive to whatever you external ref temp coef is.

    You can dynamically switch between internal/ext ref to accomodate different inputs, but you will need to give some settling time in this case after the switch, I think 100us would be OK, but I need to follow up on this aspect.

    The conversion range of the ADC is fixed with the internal refernece to 0-3.3V no matter VDDA.  The only restriction is that ADC input not exceed VDDA(nor can VREFHI exceed VDDA at any time).

    If VREFLO is kept at VSSA, VREFHI can go as low as 1.98V if that is of concern. 

    You can certainly force 2.5V on an unused ADC input if your applications requires tighter tolerances, and correct for gain error in SW as you mention.

    I will advise on the ADCOFFTRIM register which will allow you to remove any offset error dynamically in the system.  We have provided an internal connection to VREFLO on ADCINB5 that will allow this correction without using an external channel.  This correction is performed inside the converter, so there is no loss in dynamic range.

    Best,

    Matthew

  • 0 in reply to MatthewPate
    Intellectual 300 points

    hi,

    can you please provide the temp. calibration functions for ADC that should be burned to ROM (and on the kits that you released they are missing) ?

    these functions are in the ADC manual and also the application on how to use them.

    we will be delivering systems to costumers soon and i would like to implement these functions.

     

    thanks

    Ari

     

  • 0 in reply to Aristafari
    TI__Guru* 81560 points

    Ari,

    The functions mentioned in the ADC UG are present in all versions of silicon that are released(that I know of), these have to be ran in order for the device to behave within the datasheet parameters.  If you have silicon without DeviceCal() function populated, both the interal oscillator(s) and ADC will not be in spec(ADC is not usuable if this is true).  Again, this should not be possible.

    How old are these kits you have, can you send the markings on the package as well?

     

    Best,

    Matthew

  • 0 in reply to MatthewPate
    Intellectual 300 points

    hi,

    I was talking about the functions that return the offset and slope of the temperature calibration.

    from the ADC manual, P.40 :

    For the F2803x, call the below factory stored slope and offset get functions:

    //Slope of temperature sensor (deg. C / ADC code, fixed pt Q15 format)

    #define getTempSlope()  (*(int (*)(void))0x3D7E82)()

    //ADC code corresponding to temperature sensor output at 0-degreesC

    #define getTempOffset() (*(int (*)(void))0x3D7E85)()

     

    these function do not exist in the specified address.

    can you please populate them so people that bought the ControlCard etc. can use them ?

     

    thanks

    Ari.

  • 0 in reply to Aristafari
    Prodigy 50 points

    Hello,

    I have also problems with the Controlcard.

    It contains a tmx version of the f28035.

    When i try the execute the getTempOffset() function, i get illegal operation isr trap.

    Is it possible to replace this tmx version with a tms chip so the temperature calibration could work?

    Regards,

  • 0 in reply to Gtracoin
    TI__Guru* 81560 points

    These do not exist on TMX devices, other than using TMS devices which have these populated you could redirect the function calls to something written by you that either has:

    1)The typical values for offset and slope published in the datasheet

    2)Measure 2 temperature points and calculate the slope and offset for a particular device and place those in your defined tempslope/offset functions.  I think there should be enough information in the header file examples and app note to replicate/format this data correctly.  If not please let me know and we can provide this.

    Best,

    Matthew

     

  • 0 in reply to MatthewPate
    TI__Genius 9845 points
    This is an old post but I do have the same question: How can the internal reference be at 3.3V when VDDA is below 3.3V( say 3V)? Could you please elaborate? If I understand your reply correctly, when use internal reference, the full sacle range is still 3.3V even if VDDA is 3V, but the ADC input cannot exceed VDDA. Is it correct? Thanks, Zhiyuan
  • 0 in reply to Zhiyuan Hu
    TI__Genius 16281 points
    Hi Zhiyuan,
    In the future it would be better to create a new post and reference the old post but to answer your questions:
    The input voltage and reference voltage are scaled to a common range less than device minimum recommended operating condition.
    Yes when the internal reference of the ADC is used the full scale range is 3.3V.
    When the ADC internal reference is used and VDDA < 3.3V, the ADC will still convert voltages above VDDA but may not meet the INL specification in the range from VDDA to 3.3V.
    Regards,
    Joe
  • 0 in reply to Joe Prushing
    TI__Genius 9845 points

    Hi Joe,

    Thank you very much for the clarification.

    Regards,

    Zhiyuan