Designer’s Guide Community

‹ Previous topic | Next topic ›

Pages: 1

Power supply noise: is it the ratio that counts? (Read 2087 times)

Visjnoe

Senior Member

Offline

Posts: 233

Power supply noise: is it the ratio that counts?
Mar 16^th, 2006, 11:18am

Hello,

I have a question concerning the verification of analog/mixed-signal building blocks with regard to power supply noise.
Let's assume for simplicity the block's performance is to be assessed using an AC simulation.

Now, one can extract two metrics:

1. Power Supply Rejection (being the gain of the power supply 'disturbance' towards the output)
2. Power Supply Rejection Ratio (being the ratio of (1) and the nominal signal-path gain.

Which one is more relevant? Does it depend on the application? In my firm, people tend to simulate/report (1), but almost all textbooks/papers
mention (2). It got me wondering, if -just like noise- it is only meaningful to take about a disturbance with regard to the wanted/intented signal?

Example: Say we have an OTA with a nominal gain of 60dB. Assume we simulate (in a certain frequency range) a PSR of 0dB. This means the power supply noise directly appears at the output (seems bad). However, when we look at PSRR, we have 60dB (seems good). Is this OTA ok with respect to power supply noise?

Does anyone have an opinion on this?

Kind Regards

Peter

IP Logged

ywguo

Community Fellow

Offline

Posts: 943
Shanghai, PRC

Re: Power supply noise: is it the ratio that count
Reply #1 - Mar 19^th, 2006, 6:34pm

Hello, Peter,

About the two metrics that you mentioned here, I think both are useful. Of course the selection depends on applications. Most textbooks/papers use PSRR instead of PSR so that make many readers confused. For eg., when we characterize bandgap reference, we want to know the gain of the power supply 'disturbance' towards the output. That is power supply rejection as you mentioned. Personally, I would like to use the term power supply rejection here, but many textbooks/papers use the term PSRR, although it has different definition for bandgap from that for amplifiers.

When we characterize amplifiers, we need know the ratio of power supply rejection and the nominal signal-path gain. That is power supply rejection ratio as you mentioned.

How much PSR or PSRR is OK depends on the applications.

Best regards,
Yawei

IP Logged

Pages: 1

‹ Previous topic | Next topic ›