Webench solution for DC-DC converter selects unnecessarily large resistors?

Thanks Gerold,

I presume its ok then to downsize all the resistors in the circuit to smaller 0402 packages.

I am now wondering whether I can down size the capacitors Webench selected as well?

For example Cout, 10microFarad, was sourced with a voltage rating of 16V, although the output voltage will be 8.7V<10V, so it seems again that the 0805 package size is unnecessary and I could go with a 0402 sized capacitor, with 10V rating.

The question applies even more to the input capacitors: Webench wants to put two huge 1206 package capacitors with 22 microfarad in parallel. Given that I don't have much space, would there be a problem using two 22 microfarad capacitors in a smaller 0805 package? or could I even replace these two capacitors with a single 47 microfarad 1206 package capacitor? Why does Webench put capacitors in parallel anyway?

I guess my basic question is: Does webench have a good reason for selecting these bigger packages and for putting capacitors with the same capacitance in parallel?

I read that ceramic capacitors can drop well below their capacitance rating for higher voltages in their range and the more so the smaller they are, so the smaller capacitors I want to select might drop down to 30% or less of their nominal capacitance and not do their job properly in the circuit anymore? Does Webench purposefully select capacitors in order to counteract this capacity drop?

How sensitive is the TPS61088 to capacitance drop? What is the worst that can happen, if I downsize the capacitors as indicated above - (this is for driving an LED) ?

thanks a lot for your help,

Stefan

Hi Stefan,

    In this design, the resistors are not in the power path and it should be okay to use a smaller size component. For the output capacitor, I would recommend to use the WEBENCH suggested component with the voltage rating of 16V for your 9V output design. For a 9V design, having a 10V capacitor would be very close to the rating for the capacitor and also ceramics that operate close to the rated voltage would have a large derating of more than 50%. Regarding the sizes of ceramics, yes, the smaller the size, the higher the effect of derating at a specified voltage. WEBENCH considers the effect of voltage on the derating of capacitors but does not consider size. You can use the WEBENCH recommended value of capacitance and use any of the capacitor vendors to get information on the actual derating of the component as the size varies.

Let me know if you have more questions.

Regards,

Gerold

Hi Stefan,

     The Rlim and Rt resistors pick up 0402 size now for the TPS61088. Please give it a try now. Also the Rfbt picks up a 0603 resistor based on its availability in the tool. To see if there are other components that can be selected, you can click on the "Select Alternate Part" button in the BOM page within the tool.

Regards,

Gerold

Hi Gerold,

thanks very much for your reply. In order to save space, I want to replace the two1206 input capacitors with 22uF each, with a single 1210 47uF capacitor I found, which has a similar derating of 20% of its capacitance for the 3-4V input. Do you think its ok to do this replacement? Or is there any reason why two conductors in parallel are better than a single, bigger capacitor? Why does webench put conductors of the same capacitance in parallel? I know that sometimes a higher capacitance is put in parallel with a lower one to smooth out different frequencies in the voltage, but I don't see the point of using two capacitors of the same capacitance?

Thanks for your help,

Stefan

Hi Gerold,

regarding the alternative parts selection in webench for the comp and comp2 capacitors for the compensation network of the TPS61088, I noticed that the suggested capacitors vary hugely: For the comp capacitor, the voltage rating varies between 630 and 1 volt and the package size varies between 15 and 2 mm square. For the suggested alternatives for the comp2 capacitor even the capacitance itself varies between 10 times more and 10 times less than the original value! It seems unlikely that I can just select any combination of these very disparate alternatives and the DC-DC converter will still work? Are you sure Webench is configured correctly with respect to the capacitor alternatives?

I would be grateful for a quick answer, this is starting to hold things up over here,

thanks a lot,

Stefan

Hi Stefan,

For the TPS61088 Ccomp2 , Ccomp capacitors, the select alternate parts need to be tightened to prevent selecting components which are far from the target. We will have this addressed immediately.Also, the package size is not a constraint in the select alternate part, so it will list all components that are within the limits specified irrespective of the package.  You can constrain the design to select a minimum component size by going to the Visualizer-> Advanced Options -> Minimum Package and select either 0402, 0603 etc. Currently we don't allow the user to set the maximum package but you can set the Maximum Component Height if that is a constraint in your design again using Advanced Options.  

Also to answer your question on having two capacitors in parallel of the same size, is because WEBENCH found the solution to either have a lower cost or met the Irms current rating or had a lower ESR than a single larger capacitor with the same capacitance. The reason for having a small value capacitor in parallel with a high value capacitor (example a 1uF cap in parallel a 100uF electrolytic) is to reduce the high frequency noise where the 1uF behaves more capacitively at higher frequencies than a 100uF capacitor at that frequency - the ESL of the capacitor will start to play a role at such high frequencies. If you ignore the parasitics, 100uF in parallel with 1uF is as good as a 100uF cap.

Regards,

Gerold