• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 63 subscribers
  • Views 205 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

LM27762EVM: Unusually high noise on the output

Youssef El-khatib
Prodigy 30 points
Part Number: LM27762EVM
Other Parts Discussed in Thread: LM27762

Tool/software:

While using the evaluation board, I'm noticing very high noise on both the positive and negative outputs above 100KHz. I used the board as is straight out of the box, with 50Ohms load on each output. The total current draw of the eval board under load is 77mA.

The input supply is a linear Lab power supply set to 5V, and I'm measuring the outputs using coaxial SMB cables to BNC inputs of the Oscilloscope and Spectrum Analyser.

Measuring the total RMS noise from 10Hz to 100KHz, I get 30uVRMS, close enough to the spec of 22uVRMS. The difference is due to integration errors and noise floor level. However above 100KHz I can see massive switching noise on both outputs and ripple noise on the positive output :

In the time domain I'm getting : 
Vpp + = 51.382mV
Vpp - = 111.97mV ! 

These measurements are also seen in the spectrum analyser capture, see attached. The magenta trace is Vout- and the yellow trace is Vout+ in all screen captures

Is this the expected behaviour? How can I get around it? Would a simple LC filter on the output do it ? I'd like to the output below 10uVpp or approx 2uVRMS for Freq>100KHz.

  • 0
    TI__Mastermind 22400 points

    Hello Yousef,

    thanks for using E2E.

    I will check it and come back tomorrow.

    Best regards,

    Andreas.

  • 0
    TI__Mastermind 22400 points

    Hello Yousef,

    I think this level of switch noise is not normal.

    Please check the input voltage connection. For best performance, connect the power supply to the EVM with short low-inductance cables. Alternatively, an additional capacitor can be placed to the VIN node.

    If necessary then you can add an LC filter at the output to further reduce switch noise.

    Let me know if the issue persists.

    Best regards,

    Andreas.

  • 0 in reply to Andreas Mueller
    Prodigy 30 points

    Hello Andreas,

    I added an electrolytic capacitor on the input which helped with the ripple, but no change on the switching noise. When you say that the switch noise is not normal, what would you expect? Because I do not consider this to be "Low noise" like stated in the datasheet

    thanks 

  • +1
    TI__Mastermind 22400 points

    Hello Yousef,

    I quickly checked -Vout on our EVM in the lab and I measured 52mVpp switch noise. But these short voltage spikes are occurring with a frequency of ca. 4MHz, which is the double of the switching frequency of the device (2 switch events per cycle). This noise is far above the specified 100kHz range. Unfortunately this switch noise can not be avoided as it is a switching charge-pump device. You could add an LC filter to filter out the noise. But please be aware that the LM27762 is reducing the switching frequency for light load conditions, which could lead to switch noise below 100kHz. In this case you could add some permanent load to keep the switching frequency in MHz range.

    Best regards,

    Andreas.

  • 0 in reply to Andreas Mueller
    Prodigy 30 points

    Hello Andreas,

    Thank you for confirming the results. I also see this on my tests when varying the load. I do agree that it fits within the Datasheet specs up to 100Khz, so I would say this is good for applications requiring a lower BW that 100KHz.

    I will add and LC filter to see how much noise suppression I can get.

    Thanks again !