Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

How to convert LDO output noise to PSRR and ripple on LDO input

Status
Not open for further replies.

vovan76

Junior Member level 3
Joined
Aug 19, 2009
Messages
26
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,484
Hi
I have the Noise Spec on LDO output in nV/√Hz
And from this Spec I have to separate to LDO self noise and noise on LDO input with rejection --> PSSR
For Example
10MHz noise =100nV/√Hz
50nV/√Hz--> self noise
50nV/√Hz--> noise from LDO input after PSRR rejection
Now I want to know how to convert 50nV/√Hz to the max ripple on LDO input
 

10MHz noise =100nV/√Hz
50nV/√Hz--> self noise
50nV/√Hz--> noise from LDO input after PSRR rejection
Now I want to know how to convert 50nV/√Hz to the max ripple on LDO input

Hi vovan,
  • nV/√Hz is a spot noise measure, i.e. it states a certain noise voltage at a given frequency - and this value is valid only at this very frequency. So I take it as given that this value of 100nV/√Hz is valid only at 10MHz.
  • Unrelated noise voltages - and those two are unrelated - add quadratically, i.e. 2 unrelated noise voltages of 50nV/√Hz each add up to 70.7nV/√Hz - not to 100nV/√Hz. 2 unrelated noise voltages of 70.7nV/√Hz each would add up to 100nV/√Hz.
  • In order to convert spot noise to total noise, one would have to integrate the squared spot noise curve vs. frequency over the interesting frequency range, then take the root of this result.

    If the spot noise were flat between 0Hz and 10MHz - which is never true! - a flat 70.7nV/√Hz spot noise spectrum would yield an integrated rms noise voltage of √{(70.7nV/√Hz)² \times {10}^{7}Hz} = 224µV.
However, PSRR noise from power supply never has a flat noise spectrum; it much more contains (possibly a lot of) noise spikes dedicated to the frequency (and their overtones) which the DC is rectified from (line voltage frequency, SMPS switching frequency/cies).
 

Hi Erikl
First of all thank you for fast response
The example of 50nV/√Hz+50nV/√Hz=100nV/√Hz was only for example for noise buget separation (and of couse you right for unrelated noises)
acuttaly I have noise Spec for each decad and yes PSRR ids different per Frequency.

But from you answer I didn't understand how to convert noise - density or integraetd form to LDO input voiltgae ripple
Lets take the example

Fom 1MHz -->10MHz I have Noise density 50nV/√Hz integrated noise is

√{(50nV/√Hz)² \times 9*10^{7}Hz} = 0.15mV --> This is the RMS voltage
And PSRR in this Freq range for example -40Db --> rejection is about factor 100
 

Lets take the example

From 1MHz -->10MHz I have Noise density 50nV/√Hz integrated noise is

√{(50nV/√Hz)² \times 9*10^{7}Hz} = 0.15mV --> This is the RMS voltage
√{(50nV/√Hz)² \times 9*10^{6}Hz} = 0.15mV --> This is the RMS voltage. Correct.

And PSRR in this Freq range for example -40Db --> rejection is about factor 100
With a PSRR = -40dB , you could afford a power supply RMS noise of 15mV .
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top