LMR38020: Design equations, in part's data sheet, do not seem to yield the same results as the WebBench design results

hello Mark,

The product expert is out of office now. TI responses may be delayed during the year end holiday season until Jan 2nd, 2024. Thank you for your patience and happy holidays.

Best wishes,

Ryan

Hi Mark

I am curious that why would you choose LMR38020 since your application is only 4.2V - 6V and Io is 300mA.  We have lower Vin_rated buck converters. What's your end quipment and can Vinmax=36V cover your application ? Is Iomax=300mA?

For the calculation

1. Please note that in the equation(10), the Iout_max is the max rated current for the device, which is 2A.  The calculation result varies under every different Vin/Vo/fsw. But in real word we will need to select a normal inductance value, like 3.3uH/6.8uH/10uH....  

The Webench recommended the L from a preset vault. The difference might because the Vin is really small in this application. You can use the inductance according to the equation. But please use the right Iout_max. 

2.Does the calculation in WEBENCH or the equations change if the inductor is shielded?--No

3. AUTO mode/Forced FPWM mode-- Yes the suffix of the part determine which mode is used.   The difference only happens under light load, when the FPWM keeps switching   under the fixed fsw and remain in CCM mode , while the AUTO mode device will enter DCM mode and will decrease fsw as the load decrease. 

4.The equation doesn't consider the AUTO or FPWM mode since they have no difference except light load. 

5.The AUTO mode has higher light load efficiency. But FPWM has smaller Vo_ripple under light load.  If Vo_ripple is cared more , I think it would be better to select the LMR38020FSDDAR.

Thanks and best regards.

Hello Gui,

Thank you for replying so quickly.

I did have a couple of additional questions and a few answers to your questions.

The LMR38020 was selected for the following reasons:

A. We have multiple voltage rails (5V, 3.3V, 1.2V) in our design and the primary rail (5V) is derived from a 24VDC power supply.  We may also change this DC supply to 36V in the near future, so a high input voltage rated part was important.  Our application is industrial in nature and there are electrically coupled surges and interference that requires us to protect the input side of the part with a TVS device to limit those occasional over-voltages. The maximum clamping voltage of a 36V TVS is ~58V, so this also factored into the selection of the part.

B. Secondary factors were: i) the broad select-able switching frequency; ii) the spread-spectrum capability; iii) a 2A output rating (useful for other designs we have).

C. A tertiary factor was limiting the number of items in the Bill of Materials and maximizing single component use across the rest of the design.

Now for some additional questions/comments:

1. I changed my spreadsheet to make Iout_max = 2A.  This changed the resulting inductor value to 3.4uH but it is still 3x lower than what the WEBENCH results recommended (10uH).  I've attached the WEBENCH schematic for reference.  This is worrying since the two value aren't even close.  I'm going to use the WEBENCH's recommendation on the design, but I'd recommend that TI has a further discussion in a separate app note to cover how a user should resolve this.

2. A follow on question to #1.  Does the actual load current have no effect on the selection of the inductor?  If it does have an effect, where in the equations is the user's maximum load represented in selecting the inductor?  I'm sure that duty cycle would be different, but what factors are being considered outside of the equations that drive the tool's inductor selection if not the equations that are provided?

3. My next question has more to do with a problem with WEBENCH's modeling limitations.  The other rail that I was intending to use the part for was a 1.2V output @1.1A.  I've written the FAEs on this forum that WEBENCH doesn't allow modeling below 3.3V for this part, even though it should.  The response I received confirmed that the part is capable of providing a 1.2V output and that I should use the equations in the specification.  My quandary now what do I do in selecting the inductor if the equations are so vastly different than the tool?

4. Based upon my dilemma with #3, can you recommend another similar part with the following:

a. 1.5A-2A rated output, Vin_max = 10V minimum, 12V preferred (there is a TVS on the 5V rail with a clamp voltage of 9.2V)

b. Synchronous, spread spectrum.

c. WEBENCH modeling at 1.2V output, using a nominal 5V Vin.  The 5V rail is from an LMR38020 (24V/36V to 5V).

d. Similar HSOIC-8 footprint, or smaller alternate.

BTW, I was trying to attach the WEBENCH .sgv file and can't seem to see how to do it; can you help guide me?

Thank you.

Best regards,

Mark Meyerhofer

Hi Mark

For your information, we have LMR51610/06, which is 1A/0.6A,  65V Vin,  0.8V Vref, small, easy to use and with better cost.  But it has no spread-spectrum and fsw is fixed as 400kHz/1.1MHz. Will you consider this device?

Choose a 2A device can cover further needs. But in this design, you will need to select the inductor with current rate larger than the peak current limit of LMR38020(which is 3.2A).  Which adds the inductor size and cost.

For the Vin 12V device, it's from another business line and it would be better if you can start a new thread about this.

BR.