• State TI Thinks Resolved
  • Replies 25 replies
  • Answers 7 answers
  • Subscribers 62 subscribers
  • Views 340 views
  • Users 0 members are here
Support feedback
Options
Options
Related

LM25141: Need assistance in designing a 5V Buck Power supply using the LM25141

Mark Vandewaetere
Prodigy 20 points
Part Number: LM25141
Other Parts Discussed in Thread: LM51772,

Hi,

This is an urgent request for assistance in designing a 5V power supply using the LM25141. 

  • Vin range is from 6V - 36V
  • Vout 5V @ 12A

This is our second attempt to design this power supply. Our first attempt used an LM51772 buck boost with a Vin of 4V -30V range. This failed miserably! Please reference the TI forum link here.

https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1568086/lm51772-webench-design-for-a-5v-dc-does-not-work

Our customer has requested that we redesign this power supply using the new Vin stated above.  We decided to use the LM25141 Buck controller since we have already implemented a working 3.3V 16A version on this design and want to simplify the design process. We have an extream short respin time because of time lost with the first attempt.

Using your Webench online tool, we are getting slightly different values and components compared to the 3.3V version. So would need assistance in comfirming this new 5V design and verify that the working 3.3V version will also work with this new Vin requirement. We'd would prefer not to have to redesign the 3.3V at this time.

Would you please assist us?

Thank you,
Mark Vandewaetere

  • 0
    Prodigy 20 points

    Here is the Webench Report for both the new 5V supply.
    LM25141RGER ‑ 6V-36V to 5.00V @ 12A Report.pdf

    And the exisying 3.3V supply (using the original Vin requirements).
    Webench LM25141 3.3V 16A Report.pdf

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    The slightly different recommendations in components between the two applications is expected. Please see this design calc tool that can help you lock down your design better and estimate your efficiency/stability with greater control. https://www.ti.com/tool/LM25141DESIGN-CALC

    Once you do have the design calculator filled out, it should be able to tell you if your design is valid. Please feel free to send it over filled out and I can look it over to make sure nothing is overlooked.

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Thanks Joshua. I will do. I did notice that this calculator is for the "-Q1" version. Are they the same?

  • 0 in reply to Mark Vandewaetere
    Prodigy 20 points

    Here is the filled in calculator. I made some assumptions for iCin and Cout. For M1 and M2 I selected the same MOSFETs we're currently using on the 3.3V design. I not 100% certain I got all of these MOSFET parameters correct.
    LM25141-Q1_quickstart_calculator-iPCBLayout.xlsm.zip

    Mark

  • 0 in reply to Mark Vandewaetere
    Prodigy 20 points

    A question from my colleague, could we simply take the 3.3V circuit (see earlier attachment) and switch the FB pin from VDDA to GND, to work for our 5V?
    Even if our Vin range has changed and we need a lower current out (12A vs 16A)?

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    Yes, you should be able to. However, please keep in mind that your inductor current ripple may change and thus your application may not be as optimized for your 5V circuit (the optimal range for your inductor current ripple with respect to your max load is about 20-40%).

    I will take a look at the calculator and get back to you with feedback by the end of the week.

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Joshua,

    I have a quick question. We noticed that the inventory of the LM25141 is very low. Could the LM25141-Q1 be direct substitute? There seems to be more inventory of the -Q1. 
    If so, should we calculate for the -Q1 instead and used the non Q1 as the sub?

    Mark

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    More or less, yes the LM25141-Q1 can be used as substitute. However you will want to utilize the calculator again and select different components (some of the internal parameters may vary slightly). Please send over that calculator filled out and I can review that instead.

    Thank you,
    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Joshua,

    Now I'm confused. As far as I can tell, the calculator for both the LM25141 and the LM25141-Q1 are the exact same calculator. Literally the same link, same file.
    This was my question I posted on Nov 3rd. Please review.

    Both datasheet links show the following..


    and both hyperlinks are "">www.ti.com/.../LM25141DESIGN-CALC"
    If they are the same calculator, what other components would I use?

    Another question regarding their interchangeability... While the Webench tool does not allow me to enter output current higher than 10A for the "-Q!", it does allow me to enter 12A for the  LM25141. Why is this? why is the "-Q1" limited?

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    Sorry for the confusion. You are correct, they have the same calculator and should have similar internal parameters. Taking a look at the inputs from the calculator, I see no issues with your application. It looks to be stable and your crossover frequency is reasonable for your application. 

    In terms of Webench, it may be that the reference data that Webench is pulling from for the Q1 was only taken to 10A. Regardless, as this device is a controller, the only current limit for it will be based on your utilized inductor or sensing circuit and your FETs. If you have want to continue using the LM25141 to simulate the LM25141-Q1, this should be ok. 

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Great! Thank you!

    Can you tell me if I should recalculate the 3.3V as well, or will it work as is given the same input parameters as the 5V?
    Refer to Webench LM25141 3.3V 16A Report.pdf earlier in the thread.

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    I have a few more questions regarding the schematic implementation of the "quickstart calculator".

    I noticed that there are differences between the "calculated" values (shown in the schematic) and some of the values listed in the BOM.
    In particular;

    • Cin (sch) = 33uF, Cin (BOM) = 10uF
    • Cout (sch) = 68uF, Cout (BOM) = 47uF
    • Cvdda (sch) = 0.1uF, Cvdda (BOM) = 0.47uF
    • Rpgood (sch) = undefined, Rpgood (BOM) = 100K
    • Rcs (sch) = undefined, Rcs (BOM) = 100R
    • Rdemb (sch) = undefined, Rdemb (BOM) = 100k
    • more...

    Which values are the correct ones?

    Also, I noticed that in "step 9: IC Polwer Loss", if I select "VCCX Conne...", the IC Power dissipation drops from 0.44W to 0.18W. Would this not be a better choice? I'm assuming that "VCCX Connect to Vout" is what the choice is supposed to be?

    Can you please recommend any other improvements? What about Dither and FPWM? What should these be?

    Mark

  • 0 in reply to Mark Vandewaetere
    Prodigy 20 points

    Another quick question...
    The "quickstart calculator" doesn't specify a value for RT. Neither in the schematic or the BOM. Should this pin be left floating or should I install a 49.9K resister (440kHz)?

    • EN pin will be connected to Vin.
    • PG pin will be left floating
    • DEMB will be tied to VDDA

    Please confirm.

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    The BOM picks from standard values and gives the lowest multiple of the chosen standard value capacitors to meet your application. If you look to the count column, this should show you how many minimum of each component you need.

    In terms of VCCX, this is an optional external power supply input to VCC. When an external signal, or VOUT, that is greater than 4.5V is connected to VCCX, the internal VCC regulator is disabled and this external supply is used instead, reducing the device's power dissipation. Please note that an external signal on VCCX, including if the output voltage is used, cannot exceed 6.5V.

    Depending on your OSC tie and your switching frequency, the option to set your RT resistor will pop up in the calculator. For more details, please see section 7.3.3 of the datasheet that explains how the device looks at the OSC and RT pins to set your switching frequency.

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Joshua,

    I created two schematics using the LM25141-Q1. A 3.3V and the other 5V schematic as follows;

    LM25141-Q1-R02_3.3V.pdf LM25141-Q1-R02_5V.pdf

    From the original "quickstart calculators" I replaced the following;

    Dboot.
    It was originally a standard diode, BAS1602VH6327XTSA1 but the calculator specifies a Schottky diode,60V, 200mA.
    So, I found a SMD110PL-TP Schottky diode, 100V 1A (SOD-123FL package). Please confirm if this is OK. If not can you suggest an alternative?

    Cin (5V and 3.3V)
    The calculator called for 5x 10uF 1210. But because of limited space I selected 2x 22uF 2220.

    Cout (5V)
    The calculator called for 2x 47uF 1210 (94uF). But 94uF seems to mess up the Bode plot. So instead I selected 2x 33uF 1210. Please confirm if this is OK.

    Cout (3.3V)
    The calculator called for 3x 47 1210 (141uF). But 141uF seems to mess up the Bode plot a bit. So instead I selected 2x 47uF 1210 + 33uF 1210 (127uF). Please confirm if this is OK.

    Please note that we are really restricted in board real-estate. So fewer and smaller components is necessary.

    RT (5V and 3.3V)
    Is connected to ground through a 0R resistor. This will give us an option of modulating the 440KHz oscillator if needed.

    VCCX (5V)
    Is connected to Vout through a 0R resistor(internal VCC disabled), with an option to enable VCC through an optional 0R resistor.

    Rsense (5V) resistor. I could not find a 4.5mR resistor, so I used two 9mR resistors.

    Here are the update calculator (5V) and new (3.3V) calculator.
    LM25141-Q1_quickstart_calculators.zip

    Please LMK if these are good to go.

    Mark

  • 0 in reply to Mark Vandewaetere
    Prodigy 20 points

    Hi Joshua. Could you please provide an update?
    This is rather urgent. 
    Thank you!

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    Dboot:

    This should be ok

    Cin: 

    The input capacitance should be ok, but you need to ensure that the capacitors do not de-rate too significantly under the max dc bias that they are set to experience on VIN. This information should be available on your chosen capacitors' manufacturer's site.

    Cout: 

    Can you please elaborate what you mean about messing up the bode plot? If your chosen capacitor has better stability and acceptable crossover for your application, I do not see any issues with using your selected value.

    RT: 

    This is also ok. Please note datasheet section 7.3.3 for more details on this.

    VCCX:

    For the 3.3 application, this can be grounded by default as it will provide no benefits. Leaving an option to externally bias this pin is a good idea however.

    For the 5V application, it makes sense to leave the VCCX connected to VOUT by default. Likewise it is a good option to allow for external bias as well.

    I see no issues for the 5V design calculator inputs. For the 3.3V design, I would recommend recompensating as your crossover frequency is quite low. Doing so to the recommended values pulls the crossover frequency to a more acceptable range for your transient response without making your design unstable (>45deg PM).

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Thank you Joshua,

    Can you also confirm the Dboot?

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    I believe I noted that should be ok. 

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Josh,

    So I’ve taken these 2 circuits into our board labout and discovered that the Cin’s are too large, so I replaced them with the 5x 10uF 1210 50V caps recommended in the calculator. Is there some way I can send you the updated PCB file (not on this public forum)? Is so, in what format (schematic and PCB)?

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    Sounds good. I have sent a friend request. If you accept that, you can send a direct message to me through E2E.

    I would prefer the files be sent as screenshots of the schematic and the board layers.

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Joshua,

    I sent you a couple of private messages since sending you the files. Maybe you have not seen them yet?
    Can you please give me an update of your review?  Sorry. but it's rather pressing at our end.
    Thank you.

    Mark

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    Sure I will take a look at it today and provide some feedback.

    Thank you,

    Joshua Austria

  • 0 in reply to Joshua Austria
    Prodigy 20 points

    Hi Joshua, I sent you files yesterday via private messaging. Just checking.
    What is the best time to reach you? Maybe I’m not timing this correctly.

  • 0 in reply to Mark Vandewaetere
    TI__Mastermind 26200 points

    Hi Mark,

    We can continue there. Sorry, things have been a bit busy but I will review it by EOD today. For reference, I am in the PST time zone.

    Thank you,

    Joshua Austria