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.

TPS62097-Q1: StepDown DC/DC feedback loop divider: mismatch between datasheet calculation and practical implementation

Part Number: TPS62097-Q1

Dear TI colleagues,

TPS62097QWRGTRQ1 is a part of power supply chain of TimeOfFlight reference design with Melexis camera sensor. This means the customers of Melexis can see our validated design and became also TI customers by using TI parts.

The DC/DC based on TPS62097QWRGTRQ1 converts from 2v7 to 1v2:

The question is on the picture about R5 in FB loop. According to the datasheet here should be 10k up and 20k down, but voltage is wrong - 1v6 instead 1v2. I need to decrease upper resistor twice - and it works good. But since we would like to have reference design in big quantity - we would like to clarify it with TI not to risk.

Inductance is big enough to provide good ripple and not to saturate - that I can guarantee. Also different inductance tried - effect is stable, no change.

Here how our layout looks:

Oscilloscope plot on hot side of the FB divider (filter cap) provides me near 12mV rms noise. This is a digital supply and this noise is acceptable. I do not see any instability, uncompensated pole or so.


In theory formula in datasheet should be:

So, 0.8*(1+10k/20k) = 0.8*1.5 = 1.2V. But in fact 1v6 which is not acceptable. Same, if I use 5k as upper resistor, it gives in theory 0.8*(1 + 5/20)=1V but in practice 1v2 which I need.

Thank you for any ideas of this mismatch, as it looks like reference is not 0.8V but near 1.0V.

  • Hi,

    In schematic there are two Ferrite beads on output. Can you remove it and do measumrents again?

    Generally, ferrite bead add inductance and resonant tank to device and require additional damping circuit. I would suggest to remove both ferrite beads and observe output voltage.

    Best regards,


  • Dear Tanvee, sorry for some delay - needed time for investigation. Ferrite beads are separating two power domains - both 1v2. I have a capacitor at the output of DC/DC, as you may see on the shared schematic, then ferrite beads and another caps close to the chip which is supplied. Ferrite has a dissipation properties and should dump parasitic oscillations - also I do not see any oscillations on the scope. Nevertheless, I removed the ferrite beads and made a direct connection as you suggested. No change - voltage is stable and behavior of the feedback loop is the same. Second - I tried different capacitor values (ferrite still removed) like additional 100uF tantalum. Same. Voltage is stable, DC/DC behavior is stable. Do not think it is something related to frequency uncompensated feedback loop like uncompensated pole. Still looks like something with a reference of the DC/DC which requires me to have another voltage divider like mentioned in the thread.

    Still no idea what could be a reason. Do you have, for example, a reference design evaluation kit which you can share with me (Melexis company) to study this effect?



  • Hi,

    Thanks for update.

    Sorry for confusion my request is to isolate the device from out by removing the ferrite bead. Please measure output voltage without load connected to output. Remove L4 and other ferrite bead and do not short the path so that buck converter is isolated from output load and measure the output voltage.

    What is the expected load on output? is mode pin connected high or low ? how often you observe this behavior in application ?

    Best regards,


  • Hi Tanvee,

    please see below result of my study.

    First of all - operation without load completely. HW was modified - no load. Voltage is stable, not noisy - same effect. I have 1v2 on modified resistor value set.

    MODE pin usually is floating - forced PWM mode, but I've also connected it to ground (yes, I know that I need to re-power the DC/DC to apply new configuration) - same effect.

    Again - chip is working fine but looks like reference is playing here a role which is different from the datasheet one. Reference board would be very useful for us, as it still looks like miracle.



  • Hi Volodymyr.,

    Normally, higher voltage than expected is depending on output connection. Two factors you can check is the feedback connection and the reverse voltage from output. I would not expect device will not work with standard resistor divider.

    I would recommended to order EVM board. It will help you as reference board to see feedback voltage behavior. 

    I will send you the request and if you want to take it offline for more investigation.

    Best regards,