Post Go back to editing

LTC3171I regulation tolerance

Need to make sure I understand how to calculate the output voltage regulation tolerance, or range.  Datasheet states Vout = 0.6V * (1 + R1/R2).  So tolerances of R1 and R2 are obvious, but am I correct in assuming 0.6V here is the internal voltage reference?  Datasheet has a parameter called Vregulated, with a min/max range of 0.591/0.609V.  Is this the value range to be used for the 0.6V number in the output voltage equation? 

The application generates 1V nominally.  There is another datasheet parameter called Vloadreg.  For the "I" version of the LTC3731, and at 1V, the datasheet implies 0.1% nom, and 0.% max.  Is this in addition to the Vout equation?  Also confused why the value is positive for voltages programmed below 1.2V, and negative for voltages programmed above 1.2V.  Am I reading this correctly?  (Glad it wasn't programmed for 1.2V, wouldn't know which to use...) 

Bottom line:  Don't know how to calculate a worst case voltage regulation range over temperature for this part based on this datasheet.  Any other datasheet parameters that would impact the regulated output voltage range?

Parents
  • SORRY!  Wrong part number in the title:  Supposed to be LTC3731

  • Hello,

    The LTC3731 VREGULATED specification (reference) of 0.591V to 0.609V is the voltage measured at the (+) input of the differential amplifier with the ITH pin at 1.2V. The ITH pin is the output of the error amplifier and controls the inductor current. As the load rises ITH will rise and as the load falls ITH will fall. An ITH of 1.2V represents 50% load, approximately, while an ITH of 0.5V and 2.0V represents no load and full load, respectively. Therefore, the VLOADREG (load regulation) parameter specifies the slight drop of IN+ going from no load to mid-load and from mid-load to full load.The drop in IN+ is due to the output resistance of the error amplifier which is a gm type error amplifier. The output resistance is approximately 200K, typical, for the LTC3731.

    Summing the worst case variation of the VREGULATED and VLOADREG parameters along with the resistor values will provide a conservative estimate of the output voltage regulation. Another factor listed in the data sheet is VREFLNREG (line regulation). However, it is only 0.03% / V. Its variation is much smaller than the other two specifications. Therefore, it does not need to be factored in.

    Another part to consider for a single output, triple phase converter is the LTC3829. It is a newer part. Here are the key differences between the two parts:

    • Reference accuracy over the -40C to +85C temperature range:
      • LTC3829: +/- 0.75%
      • LTC3731: +/- 1.5%
    • Load regulation over the -40C to +85C temperature range:
      • LTC3829: -0.1% for ITH going from 1.2V to 1.6V
      • LTC3731: -0.7% for ITH going from 1.2V to 2.0V
    • The LTC3829 has optional non-linear control for a fast load step response, the LTC3731 does not have this feature.
    • The LTC3829 has AVP which reduces the total output excursion during a load step, the LTC3731 does not have this feature.
    • The LTC3829 has an integrated INTVCC regulator, the LTC3731 requires an external 5V supply for bias.
    • The LTC3829 can be used for input voltages up to 38V. The LTC3731 can be used for input voltages up to 30V.
    • Both have a similar output voltage range and switching frequency range

    The LTC3829 has better specifications of the two. It is worth considering.

    Best regards,

    Mike

  • Thanks Mike, lots to absorb here.  Unfortunately, the design is done, and I'm just tasked with reviewing and verifying their analysis.  I will forward the upgrade part info to them in case they ever decide to upgrade the design.  Much thanks for this info.  The actual design uses the Industrial grade part, so the worst case regulation is supposed to be over the full range capability of this part. 

    So for a full load range analysis, from no load to max load, the Vloadreg contribution is thus +/-0.7% for the industrial grade part, correct?  I believe I understand the rest, just got confused how to read the Vloadreg spec.

Reply
  • Thanks Mike, lots to absorb here.  Unfortunately, the design is done, and I'm just tasked with reviewing and verifying their analysis.  I will forward the upgrade part info to them in case they ever decide to upgrade the design.  Much thanks for this info.  The actual design uses the Industrial grade part, so the worst case regulation is supposed to be over the full range capability of this part. 

    So for a full load range analysis, from no load to max load, the Vloadreg contribution is thus +/-0.7% for the industrial grade part, correct?  I believe I understand the rest, just got confused how to read the Vloadreg spec.

Children
  • Hello,

    The LTC3731 is a good part to use but for future designs, consider the LTC3829.

    Going from no load to mid-load the voltage at IN+ of the LTC3731 will drop by 0.7%, worst case. Going from mid-load to full load IN+ will drop another 0.7%, worst case. Therefore, the total drop going from no load to full load is 1.4% worst case. Keep in mind that the VREGULATED specification is measured at mid-load. Refer to the graph below.

    Best regards,

    Mike