This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS54040A 24V Application

Other Parts Discussed in Thread: TPS54040A

Input Voltage: 23.5V to 32V with occasional transients to 42V

Nominal output voltage: 24V, dropout to 22V acceptable

Load: 0mA to 250mA

I have an application requiring an output with the conditions stated above.  I am in the process of determining if the TPS54040A is an acceptable solution.  I have a TPS54040EVM that I am using along with SwicherPro for my calculations.  Below is my current circuit on the EVM.

These values are open to modification, although getting much larger input/output capacitance with 1206 and 1210 ceramics at 24V may prove difficult (extra capacitors can be added if needed).  The capacitor values shown are nominal values and do not take into consideration DC bias effects.

I have a few questions/concerns.  

What issues does the circuit have with DCM?  Because my load gets very light, it simply is not possible to stay in CCM.  If there are none I will likely lower my inductance.  The recommended 30mA ripple will be hard with minimum input voltage (especially with a large inductor).  Is 30mA always required, especially near dropout?

When my input voltage enters the dropout range I can audibly hear the inductor (with and without load, although the sounds are different).  I assume this is from the pseudo-100% duty cycle during moderate to heavy load and ECO-Mode during light load.  Or is this a stability issue?  Will operation like this cause long-term issues?

VIN also has some minor ripple due to switch ringing.  To prevent EMI/EMC issues, is it better to more heavily filter the input or use a snubber across D1 if overall efficiency isn't the biggest concern?

Is there a better solution or IC for this application?  I have considered a SEPIC but have been wary about the increased complexity and stability issues.

Also, my load occasionally has self-contained PWM circuitry in the low kHz range.  Will this cause any stability interactions with my control loop?

Thanks!

  • Please see my comments below.

    It is ok to reduce the inductance so it operates in DCM. Make sure the peak current doesn't reach the current limit at full load.

    The noise you are hearing may be due to BOOT UVLO being triggered. At light load and low VIN you may see the output voltage drop down to ~VIN - 2.1V worst case. Do you have some oscilloscope waveforms you can show of the condition when you hear audible ringing?

    For EMC I recommend both. The snubber will help reduce higher frequency emissions due to the fast slew rates and ringing at the switching node. The input filter typically reduces lower frequency emissions closer to the switching frequency and its harmonics.

    The SEPIC is a good option if you want to provide better regulation of the output. The loop complexity only makes it harder to model which is the challenge for compensation. However a transient response can be used to judge the stability. Other things to consider for a SEPIC solution are that the switch needs to be rated for VIN + VOUT and the current in the switch is IIN+ IIOUT.

    I don't expect any stability issues with this.

  • Hello Anthony,

    Thanks for the feedback. If my max output RMS current is 250mA, is there any reason besides capacitor ripple rating (and the overcurrent trip point you mentioned above) to avoid running in DCM over the entire load range?

    After further investigation it does appear that BOOT UVLO is being triggered in both situations that audible ringing occurs.

    Note on scope measurements below: Red = Input Voltage, Blue = Output Voltage, Black = BOOT-PH Voltage

    Low input voltage, no load

    Low input voltage, 167mA load

    Inadequate BOOT-PH Voltage appears to be the problem.

    Does TI have any solutions that would enter a true 100% duty mode (i.e. internal charge pump)?  I may still consider a SEPIC.

  • There are a couple solutions suggested in the below application note.

    http://www.ti.com/lit/an/slva547a/slva547a.pdf

    You may want to consider the diode and resistor at PH solution. It is one of the more effective solutions. The higher output voltage and lower output current will not be affected as much with the added diode.

  • I am going to try the LM(2)5085 instead.  Since this is a PFET controller, I should be able to achieve true LDO operation near dropout.  This should eliminate any audible frequency shift due to BOOT UVLO.