Post Go back to editing

EVAL-CN0536-ARDZ comparator confusion


I cannot make sense of the feedback section of the schematic. How are VHYS=1.145V and V_POWER GOOD =1.025V obtained if there's no path to ground through DGND via R12 and R13? The input leakage current to the inverting input of U5 seems to be so small that it should have a negligible effect on the voltages.

Therefore either

a) VREF=VHYS=V_POWER GOOD=1.182V and R12, R13 and C13 are pretty much useless, or

b) There should be another 3.7-3.8M resistor in parallel with C13

What adds to the confusion is that in Figure 1 of CN-0536 the voltage divider is drawn correctly, but in Figure 15 has the same problem as in the full schematic, but the V_POWER GOOD is shown as 1.182V.

  • Hi JesusC,

    Thanks for reaching out to help clarify the feedback section of the schematic for CN0536. 

    The LTC1540 has a single comparator with build-in reference, set at 1.182Vref.  That same Vref was used on the LTC1441 as a high voltage power-good indicator.  Please refer on figure 15  of the Circuit Note as it gives the correct circuit diagram. 

    We will fix the issue in the documentation for this Reference Design.



  • Thank you for the answer.

    If Figure 15 is correct and the input of the LTC1441 is supposed to be the same as V_REF of LTC1540, then I do not quite understand why the HYST input of the LTC1540 is taken between R12 and R13. When C13 is charged, V_HYS=V_REF i.e. R12 plays no role, but initially during startup when C13 is discharged, V_HYS will be 0.923 V, which is well below the minimum acceptable voltage specified in the datasheet (V_REF-50mV).

    This happens very briefly, so maybe it will not pose a problem, but I think it would then make more sense to connect HYST directly to V_REF and if R12+R13 and C13 are meant provide some filtering for the input of the LTC1441, then just use a single 580k resistor and a 47p capacitor. Although I would expect the V_REF to be stable enough not to need any extra filtering, given that it already has a 1uF bypass.

  • Hi JesusC, we did some more digging and indeed there is an issue. The initial prototypes of this circuit had the following connections:

    REF pin (1.182V)


    HYST pin (held at 1.145V)


    Power Good threshold (1.025V)


    1.025V (to another comparator input, not used in the final circuit)



    The 3.01M and 866k got omitted on the production circuit. We'll update the circuit on the next build, in the meantime, the expected effect is that the BOOST ENABLE period will be shorter than expected, with correspondingly lower ripple on the high voltage output, and slightly higher quiescent current consumption.

    If this is an issue in your application, you can connect a ~4M resistor across C13.


  • Makes sense now, thank you for looking into that.