LM76003: Challenges with Webench simulation

Obinna Ezeilo

Part Number: LM76003
Other Parts Discussed in Thread: LM76002, , LM76002-Q1

Hi team,

My customer has a few questions around simulating LM76003 on Webench. Please see below:

Value of Cff: According Webench, a value between 7.4pF and 8.2pF has to be chosen. When we calculate Cff according Equation 25 and 26 from the datasheet, we got (with Cout = 94uF, Vout = 24V, RFBT = 68k, RFBB = 3k) a value from 1.64 nF so we choose 2.2nF So which value we should take? 7.5pF or 2.2nF?

Webench is not happy with 2.2nF:

When we try to design the same with LM76002, it is not possible, to change the input to 55V (all values over 53V ending with this Message):

What is the reason for it? What can we change?

Also, when we compare to LM76003, the Phase Marg is different with the same Vin and Iout:

Can you please advise.

Thanks and looking forward to your response,

Obinna.

over 4 years ago

0 Obinna Ezeilo over 4 years ago

TI__Intellectual 2385 points

Webench Design Report

WBDesign4_LM76003 55V 1_5A 68uH.pdf

0 Samuel Jaffe over 4 years ago

TI__Mastermind 25590 points

Obinna,

The LM76002 and LM76003 have slightly different internal compensation parameters to optimize for typical conditions for each part. This is why we're seeing different results for stability.

I was able to create a WEBENCH design for LM76002 by reducing VIN max from 55V to 53V.

CFF is typically on the order of tens of pF or less. And may not be necessary for this high VOUT. Can you please share the equation and values you used to get 2.2nF?

-Sam

0 Obinna Ezeilo over 4 years ago in reply to Samuel Jaffe

TI__Intellectual 2385 points

Sam,

What will happen, when 55V is on the designed LM76002 instead 53V? Would it be a problem?

For the CFF they used the followed equation from datasheet:

With the following values:

Therefore, they chose 2.2nF.

Is it as marked in red correct, that RFBT and RFBB has to be calculated as there were parallel at Equation (26)?

Thanks,

Obinna.

0 Samuel Jaffe over 4 years ago in reply to Obinna Ezeilo

TI__Mastermind 25590 points

Obinna,

55VIN may violate TON_MIN. That would cause the device to reduce the switching frequency to maintain the required effective duty cycle for regulation. This is fine if you are okay with this reduced frequency in this condition.

Yes that symbol means "in parallel" and you are using the equations correctly but the value is very high. I will need to check with the team to confirm why we have this disconnect.

-Sam

+1 Samuel Jaffe over 4 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

Obinna,

From the team, the equations are used as a starting point and CFF should be adjusted depending on each individual board/BOM. I think the abnormal condition of 24VOUT is what gave us such a wild number but WEBENCH does iterative calculations which provides a more reasonable value.

-Sam

0 Obinna Ezeilo over 4 years ago in reply to Samuel Jaffe

TI__Intellectual 2385 points

Sam,

What is the difference between LM76002 vs LM76002-Q1 and LM76003 vs LM76003-Q1? Is it only the AEC-Q100 qualifying for Automotive? Are the technical data and behavior the same?

0 Samuel Jaffe over 4 years ago in reply to Obinna Ezeilo

TI__Mastermind 25590 points

02 vs 03 differences are the current limit and compensation. -Q1 vs non-Q1 is automotive grade (more testing during production). They are otherwise functionally equivalent.

-Sam

Power management

Power management forum

LM76003: Challenges with Webench simulation