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.

TPS51916: Component Selection

Part Number: TPS51916

Hi Team,

My customer is working on a DDR3L power supply design that will use the TPS51916.

We have come up with some questions:

  1. They have been mostly using Webench as a basis for the design, and had a question on the implementation of the Cvttref. Mainly, it is a 220nF COG/NP0 Class-1 capacitor, and there are no selections within Webench that are class 2 (X5R to X7R or so). This particular combination of cap value and cap class make for a limited number of products available, and the ones that are available are quite large in size (1206 is the smallest). Is there a very specific reason for having a Class-1 cap on this output, or can we have a higher capacitance Class-2 cap that would have a guaranteed minimum capacitance of .22uF based on deratings of DC bias and Temperature? Would the varying capacitance value over temperature cause a problem?
  2. Can we get some clarification on the best operating modes between the D-CAP and D-CAP2? It would appear that the D-CAP will have a better operation with regards to transient response. Is there any reason to consider the D-CAP2 aside from a desire to keep with MLCCs?
  3. We are a bit curious on how the transient response is better at lower switching frequencies… The general rule we have assumed is the bandwidth is about 10% of the switching frequency… It would seem that raising the switching frequency would raise the control bandwidth… Is that not the case for this supply, and that general rule does not apply?


  • Hi Mitchell,

    1. Can you use X5R or X7R capacitors for VTTREF decoupling. I think Wenbench just want to maintain 0.22uF min with COG/NP0, you definitely can use X5R or X7R with higher capacitance to guarantee 0.22uF min capacitance. 

    2. D-CAP needs higher ESR Cout, which usually yields higher Vout ripple. Also since the reference is compared with the valley of Vout ripple, the output accuracy will depend on the magnitude of output ripple. D-CAP2 supports all ceramic output cap application, with slightly slower transient than D-CAP. 

    3. For D-CAP mode control, there is no error amplifier, so the connect of bandwidth doesn't apply. With lower switching frequency, the effective duty cycle can be higher during load step up (since the max duty is limited by min off-time), so the transient performance at load step up is better with lower switching frequency. However, at load step down, lower switching frequency usually will give higher overshoot.