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.

Dual Supply System - TPS6040X & TPS72301 - Noise Observed

Other Parts Discussed in Thread: TPS7A80, TPS7A30, TPS63700

Hello all,

I am designing a system that takes power from an input supply and generates a negative and positive supply voltage for biasing op amps. I created a test setup, but I am getting some odd results.

Design Brief:

- input power regulated switching power supply (will eventually be USB bus power)

- 5V in, regulated to + and - 3.3V

- Must be able to handle current of 800mA (maybe 1A max) through positive regulator

- Must be able to handle current of 40 mA through negative regulator

- noise level of less than 1 mVpp

 

See the following schematic:

When loaded with active devices (analog to digital converter, four op amps, etc.) and drawing about 91 mA current, the waveforms observed are as follows (labels on schematic correspond to waveforms below):

Figure 1

Figure 2

Figure 3 (inverter supplying about 17 mA)

Figure 4

 

These noisy waveforms feed into the signal stage and cause a great deal of noise on the signal (see above figure). I have an instrumentation application, so I need to figure out how to get rid of this noise completely. My analysis indicates that the inverting switched capacitor circuit is a very problematic element. I have heard that the 250 kHz fixed version of the TPS6040X has better noise performance under load.

Does anyone have suggestions on how to go about fixing this problem? I would like two low noise LDOs, one for positive and one for negative voltage regulation, as well as to generate the negative voltage from the switching capacitor IC.

Thanks,

Alex

 

  • Alex,

    When using switching converters, as you know, you will get a significant voltage ripple. My recommendation is to use a High PSRR LDO to clean your signals.

    For the positive rail, consider TPS7A80.

    For the negative rail, consider TPS7A30.

    Noise is not the biggest concern here... it is PSRR. The internal noise of the LDOs is of only a few uV, hardly noticeable when you see ripples in the other of mV.

    Finally, maybe this document will give you some ideas for future projects: http://www.how2power.com/newsletters/1009/products/H2PToday1009_products_TI.pdf

     

                                    Best Regards,

     

                                             Victor.

  • BTW, you will need to set the output voltages of your DC/DCs to a higher value to give the LDOs enough headroom to clean your signals optimally.

    Or you could get rid of the positive DC/DC completely and just use TPS7A80.

    for the negative rail, replace the LDO for TPS7A30.

  • Victor,

    Many thanks for the quick reply. The voltage headroom seems to be perfectly okay (+5 -> 3.3V). The odd thing about the whole setup is that the DC/DC inverter is the root of all that noise. When I use the power supply to do the + and - voltages directly to the LDOs, the problem is not as pronounced.

    Perhaps since my power supply is off USB's +5V, I should just have a positive voltage regulator from the USB Vbus rail and have an inverting DC/DC inductor based solution followed by your proposed negative LDO.

    Do you have any suggestions about this in this light?

    Thanks,

    Alex

  • Alex,

    The TPS72301 will have higher PSRR at lower frequencies so switching to a lower, fixed frequency version of the TPS6040x will likely reduce the ripple on the output.

    Or, you could try the TPS63700 inductive based inverter.

    Regards,

    Jeff

  • Alex,

             You can double check that the switching frequency of whichever inverter you use (charge pump or DC/DC) is within the bandwidth of the LDO to ensure that this noise gets effectively rejected.

             I agree that it will be better to use the positive LDO from the USB supply and just an inverter plus the negative ldo for the negative rail.

            Hope this helps,

                        Victor.