Created on 2011-03-21 22:16:00
This article describes measurement of PSRR using the capacitor coupling method. An alternate method, using transformer coupling, is described in Technote 106: Measuring Power Supply Rejection Ratio (PSRR).
Power Supply Rejection Ratio (PSRR) is a measure of a device’s ability to reject noise from the supply used to power it. It is defined as the ratio of the change in supply voltage to the corresponding change in output voltage of the device. It is often desirable to measure PSRR over a range of frequencies and to produce a spectrum plot of PSRR versus test signal frequency. PSRR is often expressed in dB, where ΔVin is the change in voltage input and ΔVout is the change in voltage output (see equation below). However, due to lack of standardization, the ratio is sometimes inverted, and the value in dB is sometimes expressed as a negative number. Some datasheets also refer to AC (as opposed to DC) PSRR measurements as kSVR.
PSRR measurements are typically made for ICs and other functional assemblies. To measure such a device’s power supply rejection, we need to insert an AC signal (ΔVin) onto the DC voltage from a power supply and examine the device’s output (ΔVout) for the presence of the signal. The AC signal generator, in this case an AP 2700 Series instrument, can be coupled to the power supply rail with either a capacitor or a transformer.
Capacitive coupling is simpler to do, and is discussed below. Its major limitation is that because the generator is loaded down by the output impedance of the DC power supply, it is necessary to insert a series resistor into the power rail. You need to be careful not to increase the output voltage or decrease the load so much that the maximum output current of the generator is exceeded. You also need to verify that excessive load is not causing distortion to increase, which would invalidate the readings. If you need to supply larger amounts of current, you can use the transformer coupling method instead, which is discussed in Technote 106: Measuring Power Supply Rejection Ratio (PSRR).
The graph below shows PSRR versus frequency for a class D amplifier used in mobile devices, with appended traces for different rail voltages. You may download the .at27 test file and sweep file that we used to make it from our website.
Schematic
RSVR is chosen to maintain an adequate voltage at the node with CSVR, and to prevent excessive load on the generator. However, if it is too high, it will reduce the “stiffness” of the DC supply (its ability to maintain a constant voltage) at the device under test. A good compromise is 20 Ω—this will cause a 6 dB drop in generator output when its source impedance is set to 20 Ω.
DC blocking capacitor CSVR prevents the AP 2722’s generator from loading down the power supply. Since CSVR and RSVR form an RC filter, CSVR should be chosen so that the lowest frequency of interest is not rolled off too much. Choosing a 330 µF electrolytic capacitor for CSVR, along with a 20 Ω resistor for RSVR, creates a high-pass filter with a 24 Hz corner frequency. The regulation feature in AP2700 will compensate for this roll-off when we make the measurements.
The Analog Generator is set up as shown below:
Analyzer channel A is connected across the power-supply pins of the DUT, to measure the incoming AC stimulus. Channel B is connected across the DUT’s speaker terminals, to measure any change in output. Since the DUT in this case is a class D amplifier, an AP AUX-0025 Switching Amplifier Measurement Filter is inserted before the analyzer inputs to minimize the high-frequency switching frequency.
The Analog Analyzer Bandpass function meter is used to measure the amplitude of the ripple created by the analog generator.
The Digital Analyzer Crosstalk function is used to find the ratio between the Channel A stimulus and the Channel B DUT output. The crosstalk measurement utilizes a 1/13th octave bandpass filter to eliminate broadband noise from the reading.
Regulation is used at each step to keep the ripple magnitude constant at all frequencies.
This automatically compensates for the low frequency roll-off of the coupling capacitor. Set the regulation panel Anlr.Bandpass Target Value to the rms value of ripple that you desire.
Set the Sweep panel as shown. In this example, we use a 1/6th octave sweep table, but you can use whatever interval (number of steps) that you want.
Ground the inputs to the DUT. Turn on the DC power supply and set it for the correct DC voltage. Then, run the sweep and watch the plots being created.