LM51231-Q1: LM5123 Application Hints

Part Number: LM51231-Q1
Other Parts Discussed in Thread: LM5122, LM5125-Q1, TPA3223, LM5126-Q1

Tool/software:

Hi all,

i would have to questions about the Usage of the LM5123.

1. BIAS Supply: Is it possible to use a separate Buck Converter (V_IN to 6V) to supply this Pin instead of V_IN or V_OUT? The losses from the internal LDO generating the 5V VDD could be minimized like that. According to the datasheet, this is OK, but in all application diagrams, Schematics, Eval etc. it is never done.

2. Usage of a Output LC Filter: According to the Datasheet, the V_SENSE Pin must be connected to the Drain of the Highside FET. Usually, i would include an LC Output Filter in the Sensing Loop to a) avoid stabiltiy issues, b) improve voltage accuracy.

The actual application has a number of ceramic capacitors on the Output, followed by a Inductor, more ceramics and the bulk elkos. Like this, the Current Load on the Electrolytics is lowered a lot, but the ripple on the drain of the mosfet is quite high. With "classical" boost Converters as the LM5122, i usually conenct the sensing pin at the bulk elcos without an issue, but i'm, unsure how to implement this with the LM5123.

thanks a lot for tips and tricks here!

Best regards,

Lukas

  • Hi Lukas,

    Thanks for using E2E.

    1. Sure, this is possible an can improve the efficiency.

    2. Please connect the Vout sensing pin at the output capacitors like you would do it with LM5122. I would not recommend to use a LC filter. This would change the transfer function and the compensation would need to be changed. So stability might be worse when doing this.

    Best regards

    Moritz

  • Hi Moritz,

    thanks alot for the answer.

    Regarding 2. Usually i do set up the compensation network to include all the filter and parasitics, with analog devides Parts LTSpice is quite handy here. So, is the function of the V_SENSE pin ONLY the Voltage Sensing, or are there more functions multiplexed onto this pin which could be compromised by an additional LC Filter?

    Best regards,

    Lukas

  • Hi Lukas,

    The additional function of the Vout sense pin is the zero-current-detection. 

    Thats why it should be connected not too far away from the drain (At the ceramic caps that are vlosest to the output is ok, but not too far away at the elkos). Please refer to the connection of this pin on our EVM for this device. 

    We do not recommend to have an additional filter there also because of the ZCD. This signal should not be filtered or delayed.

    Best regards

    Moritz

  • Hey Moritz, good to know. The reason why i need the LC Filter is a reduced output ripple for an audio application. Hence - it would be OK to connect the V_Sense directly on the ceramics/Output of the Fet and just put the LC Filter behind it? Is there a specific lower limit on the allowable ripple  at the V_SENSE Pin?

    By putting the ceramics in front of the choke and the electrolytics behind it, the actual ripple on the V_SENSE pin would be several volts. Do you see an issue there? I put a screenshot of the schematics here to visualize the arrangement.

    Thanks alot!

  • Hi Lukas,

    1. When the gate driver current is low, people typically do not use a buck converter because the loss due to LDO is very low. But the cost is high.

    2. Can you share more information of your application? Waht is the Vin, Vout and Iout?

    You might be interested in LM5125-Q1 which is designed for audio application.

    For the schematic, you should connect VOUT/SENSE to the ceramic caps.

    Best regards

    Feng

  • Hi Feng,

    1. i do have a "spare" Buck Converter set to 5.75V on the board anyway (preregulator for the 5V_LDO for the analog Supply), so it comes very handy. It is a battery powered application, so the few 100mW the LDO inside the LM5123 for the gatedriver shall be omitted if possible. Basically, the system power draw during near 0 Load conditions has to be optimized a lot, so the actual gate drive currents to play a major role. 

    2. It is a dual Rail System, so:

    V_in:12- 20V (Liion, 4 - 5S)

    V_Out (15 - 42V & 15 - 21V, TPA3223 for Subwoofers and 3x TPA3221s for Mid/Highs are used)

    I_Out (Peak 12A, cont. 4A. for both Boost Converters.

    Meanwhile i worked out a configuration where either - the LC-Filter can be positioned after the electrolytics, or it can be omitted. It requires a little more Parts than i would have hoped in order to handle the continous ripple currents, but it is manageable.

    I already found the LM5125-Q1 already, it is indeed a suitable candidate to consider, yet it adds more complexity than the single phase counterparts. Still - it has to be evaluated, since the board space and manufacturing technologies and choice of parts are very limited.  

  • Hi Lukas,

    LM5123 is not suitable in Class-H system, you should consider LM51231 for single phase version.

    We also have LM5126-Q1, it is the single-phase version of LM5125-Q1. LM5126 will be released soon.

    I would suggest to consider LM5125 for 42V output and LM5126 for 21V output.

    The deadtime of LM5125 and LM5126 is adjustable compared to LM51231 which gives you more flexibility.

    Best Regards,
    Feng

  • Hi Feng,

    thanks alot for all the Input. Is there any datasheet or something available for the LM5126 yet? 

    Best regards,

    Lukas

  • Hi Lukas,

    Please contact TI local sales office for datasheet, samples and EVM.

    Best Regards,
    Feng