Post Go back to editing

Stability for LT1764 and LT176X Linear Regulators

Category: Hardware
Product Number: LT1764
Software Version: NA

I get this question relatively frequently these days so I will post the question and then answer it myself in another post:

What are the requirements for COUT so the LT1764 linear regulator and similar LT176X LDOs are stable?

Parents
  • My perception is that the LT1764 and similar LDOs require ESR for stability in general BUT those LDOs were designed when the typical COUT was tantalum and not ceramic, PLUS tantalum capacitors at that time had higher ESR than typical tantalum and polymer capacitor have now, so now LT1764 circuits can have stability problems because customers want to either use ceramics for COUT or modern low ESR tantalums or polymers that don’t have enough ESR either.  

     

    Further, the LT1764 datasheet does not do a good job explaining how much COUT ESR is necessary for stability.

     

    I recommend that  customers consider the guidance in the LT1764A datasheet about COUT and ESR.  The LT1764 and LT1764A probably do have a subtle feedback loop difference, but it is also important that the LT1764A datasheet does a much better job characterizing the LT1764A (and probably the LT1764) stability based on VOUT, COUT and COUT ESR.  The LT1764A datasheet specifically mentions that it may be necessary to add ESR outside of the chosen COUTs.  I show that text in the image below:

     LT1764A DS COUT ESRE_1.png

    You can see that the LT1764 is sensitive to COUT ESR just using LTspice.  I uploaded my evaluation circuit.  But it is best if customers perform a load transient test on their hardware and make sure they have some margin to prevent instability due to part-to-part or temperature variation.

     

    Lastly, the LT1764 and similar LDOs are also sensitive to the amount of CIN depending on the input power network, but the datasheets do not mention that either.  Make sure CIN is sufficient for your circuit and include some margin for part-to-part and temperature variation. 

      1764_1.asc 

Reply
  • My perception is that the LT1764 and similar LDOs require ESR for stability in general BUT those LDOs were designed when the typical COUT was tantalum and not ceramic, PLUS tantalum capacitors at that time had higher ESR than typical tantalum and polymer capacitor have now, so now LT1764 circuits can have stability problems because customers want to either use ceramics for COUT or modern low ESR tantalums or polymers that don’t have enough ESR either.  

     

    Further, the LT1764 datasheet does not do a good job explaining how much COUT ESR is necessary for stability.

     

    I recommend that  customers consider the guidance in the LT1764A datasheet about COUT and ESR.  The LT1764 and LT1764A probably do have a subtle feedback loop difference, but it is also important that the LT1764A datasheet does a much better job characterizing the LT1764A (and probably the LT1764) stability based on VOUT, COUT and COUT ESR.  The LT1764A datasheet specifically mentions that it may be necessary to add ESR outside of the chosen COUTs.  I show that text in the image below:

     LT1764A DS COUT ESRE_1.png

    You can see that the LT1764 is sensitive to COUT ESR just using LTspice.  I uploaded my evaluation circuit.  But it is best if customers perform a load transient test on their hardware and make sure they have some margin to prevent instability due to part-to-part or temperature variation.

     

    Lastly, the LT1764 and similar LDOs are also sensitive to the amount of CIN depending on the input power network, but the datasheets do not mention that either.  Make sure CIN is sufficient for your circuit and include some margin for part-to-part and temperature variation. 

      1764_1.asc 

Children
No Data