LM5160: LM5160 flybuck 42V to 57V Vin

I will pass this to the Webench guys and see if they can help.

Jonathan,

Thanks for the quick verification that this is a WEBENCH issue. I have noticed another issue with this LM5160 design. My design inputs are Vin 42V to 50V, Vout 24V and Iout 0.95A. WEBENCH allows me to select the LM5160 for these design inputs because Vin max is 50V. The design seems to make sense except for the flybuck output components. Cout2 is a 4V capacitor but the output is 24V! D1 has a 10V reverse voltage which is also not correct. I checked the recommend Cout2 part number and it is indeed a capacitor with a 4V working voltage. It would be comforting if someone who has worked with LM5160 flybuck designs could check for any other bugs.

Regards,

Ken

Hi Ken,

I'll look into this to make sure that the correct components are being selected for the design, and will do some additional tests to make sure that there are no other bugs with this part.

Regards,
Jonathan 

Hi Ken,

I did fix the bug regarding the diode and capacitor selection.  We're currently running additional tests to make sure that there aren't any other bugs regarding design creation or component selection.  

It may take a couple of days to complete all the testing and get the update released, so in the meantime, if you would like us to create the design for you and to share the design report so that you have a BOM or some basic operating values (power dissipation, currents, duty cycle, etc.), please let me know and I can send that over to you.  

Thank you again for bringing these bugs to our attention and I do apologize if they have cause you any inconveniences.

Regards,
Jonathan

Jonathan,

I have gone ahead and created a WEBENCH design with a Vin max of 50V and using a custom transformer. I am building that design on a breadboard so I am OK for now. I will wait for the update to be released and then run it again, but I will also learn a lot from the breadboard build in the next few days.

One request: Could you please let me know when the update is available for me to use. Thanks.

Ken

Hi Ken,

I've run the tests, and didn't find any other issues.  I am prepping for this update to be released soon.  I'll let you know once it gets released so that you can create your design in  WEBENCH, but I just wanted to give you a quick update on this.

Regards,
Jonathan

Jonathan,

Thanks for keeping me in the loop on this. I have been trying to do an LM5160A design with webench using limiting the input range to 50V. The specs are Vin = 42V to 50V. Vout = 24V at 0.75A, isolated. I attached an LM5161 design (#119) and an LM5160A design (#120). The LM5161 design has a Cac of 1 nF while the LM5160A design has a Cac of 100nF. The other 2 ripple injection components, Rr and Cr, are the same. Why the big discrepancy of a factor of 100? Is this a webench bug, or is the value of Cac not very sensitive? I looked at the LM5161 data sheet and EVM and all examples have Cac of 100nF not 1nF. So maybe that points to a webench bug with the LM5161? Another problem: the LM5161 design allows me to choose an off-the-shelf transformer. I chose a VPH4-0140R from Coiltronics and wired it up for a 1:1.5 turns ratio. I bread-boarded this design and it works. The LM5160 webench design does not allow me to select an off-the-shelf transformer. Is that a bug with the LM5160? I was thinking of replacing the LM5161 with the LM5160A in my bread-board design, but I'm stuck wondering if the Cac and transformer discrepancies will cause problems.

I hope you can shed some light on these webench issues. Thanks.

Kenwebench_design_4441883_119_607135116.pdfwebench_design_4441883_120_474174746.pdf

Hi Ken,

I will need to discuss the Cac discrepancy with the WEBENCH Apps engineer responsible for the LM5161 model in order to know why this issue exists, so I will get back to you as soon as possible on that.

Regarding the transformer selection, the reason why you're able to select an off-the-shelf transformer for LM5161 is because the LM5161 WEBENCH model was created more recently than the LM5160 model, and we've been able to enhance a lot of our component selection algorithms for these newer WEBENCH models.  Since the LM5160 model was created a couple years ago (vs. a few months ago for LM5161), those enhancements weren't included in the transformer selection for LM5160.  I'll double check to make sure that is indeed the case.

I'll try to get back to you as soon as possible regarding Cac so that you can make a decision between using LM5160 and LM5161.

Regards,
Jonathan

Hi Ken,

I've discussed the Cac discrepancy with both the WEBENCH apps engineer and the product line apps engineer, and there shouldn't be any issue with Cac being either 10nF or 100nF or 1nF.  It all depends on what kind of Vout rise time you desire.

It is preferred to have Cac > 3 * Cr for most cases.  Higher Cac values will provide slower rise times for Vout, and lower Cac values will provide faster rise times, so you can make a decision on whether to increase or decrease the value of Cac based on what kind of rise time you are looking for for your design.  For most cases, 100nF Cac cap is sufficient, but again, there's no reason you can't use something different.  

After discussing this with the apps engineer responsible for this, we both agreed that the value of Cac shouldn't affect the selection between using LM5160 and LM5161.

Hopefully this clarifies things for you and helps you make a decision between using LM5160 and LM5161 for your design.

Regards,
Jonathan

Jonathan,

The LM5161 design had Cr = 3.3nF and Cac = 1nF. So clearly Cac was no where near the range of > 3 * Cr. It sounds like Cac = 10nF would be sort of a minimum value, but there is nothing wrong with 100nF. Thanks for getting that clarified.

Ken

Hi Ken,

Just an FYI, the update was released to WEBENCH.  You should be able to create your LM5160 Fly-Buck design in WEBENCH now.  Please let me know if you have any other questions or concerns.

Best Regards,
Jonathan Arzadon