Ask the Experts Nugget No. 3: Power Supply Rejection in an LDO Regulator

More tutorial nuggets courtesy of Greg Waterfall of Texas Instruments, one of our experts for Planet Analog’s Ask the Experts session on Wednesday, April 23 at 1:00 p.m. EDT.

PSR vs. noise in LDOs
Expressed in dB, Power Supply Rejection Ratio (PSRR) denotes the low drop-out (LDO) regulator’s ability to reject noise from the input of the power supply. The higher the absolute value, the better equipped the LDO is to attenuate ripple from the input source, prohibiting it from affecting the output rail.

Output noise voltage (µVRMS) represents the noise generated from the LDO itself. The lower this value, the less impact the LDO will have on the integrity of the desired output rail.

Why do we want high PSR?
The output of many DC/DC converters can be very noisy for some applications due to the high frequency switching pulse width modulator (PWM) in the switching regulator inside the DC/DC.

By placing the correct LDO after a switcher, most of the noise can be filtered.

Of course, one could try to filter out the noise by means of a discrete LC filter solution, but :

  • LC filters take up lots of space and are typically more expensive.
  • The LC filter won’t be able to clean up lower frequencies as a performance LDO can.
  • Compensating for the DC/DC converter is more difficult when using an LC filter.
  • There’s a resonance associated with any LC circuit.

So we have the input to our LDO (which is the output of the noisy DC/DC converter)…

…vs. the LDO “filter” effect on that noise due to the PSRR of that device (We use the Texas Instruments TP717 as an LDO example here).

Let’s take a closer look at PSRR

PSRR is the measurement of a power supply’s ability to reject/filter noise on the input bus. Historically, LDOs have had poor high-frequency PSRR performance, but today LDOs have PSRR > 40dB @ 5MHz.

Most LDOs have high PSRR at lower frequencies. Having high PSRR over a large bandwidth is what distinguishes high-performance LDOs from the lower performing ones.

Switching frequency noise/spikes may show up at the output of an RF VCO, which after mixing will affect the PA performance. Also, in audio, noise may fold back into the audio band and create noise in an audio application.

4 comments on “Ask the Experts Nugget No. 3: Power Supply Rejection in an LDO Regulator

  1. amrutah
    April 22, 2014

    I have a basic question.

    We have DC-DC switching regulators so that we want to get very high efficiency as compared to LDO's.

    Now, in order to improve the PSRR we introduce the LDO back taking the hit on the efficiency? The total efficiency has taken a toll?

    This is going to be very interesting topic.

    Looking forward for the discussion. Thanks.

  2. amrutah
    April 22, 2014

    Are also discussing about the snubber circuits (LC circuits to avoid transients)?

    A good material on the TI site about snubber circuits

  3. Steve Taranovich
    April 22, 2014

    @Amrutah—pleae do ask these and other questions on tomorrow's “Ask the experts” session. You can even post them right now by clicking on the link I have provided in previous blogs

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.