Linear Regulator output ripple voltage

Thread Starter

illusive

Joined Jul 9, 2015
46
Hi,
I'm building a power supply based on lm317/350 voltage regulators and i'm trying to figure out using the datasheet what output voltage regulation/ripple to expect based on the input voltage. The input voltage comes from bridge rectifier and filter caps and i expect about 1.2V p-p ripple on that voltage (50Hz before the rectifier so 100Hz after it) at full load (about 2.5A). The input voltage will be about 34 volts filtered. The output voltage will be 1.25-24V 2.5A max. What parameter should i look for in the datasheet and whats the basic procedure for some rough estimates?
 

dl324

Joined Mar 30, 2015
16,845
You don't need to check the datasheet. At 24V out, you have about 10V of headroom for the regulator; so you could have several volts of ripple and it won't appear in the output. Assuming a reasonably sized filter cap.
 

Thread Starter

illusive

Joined Jul 9, 2015
46
Yes, i am aware that i have big headroom, i was wondering more of the millivolts output noise, i.e. how much of that 100Hz signal will appear as noise at the output. I'm trying to figure out which of the regulators i have will give me the smoothest output, the LM317 or the LM2676 switcher (with secondary output filter, which according to it's datasheet should give me 2-3 milivolts p-p noise).
 

dl324

Joined Mar 30, 2015
16,845
In general, switching regulators have more "noise" than linear regulators because the whole idea with switchers is for the pass element not to conduct continuously. It's higher frequency, so easier to filter. Why do you want/need low output ripple? Most circuits will tolerate much more than that without impact.
 

Thread Starter

illusive

Joined Jul 9, 2015
46
Thanks crutschow!
The datasheet says that the LM350 has a ripple rejection of 80dB (with a cap at the adj. pin). So if my calculations are correct the regulator will reduce the input ripple by a 10 000 times, and the 1.2V p-p input ripple will be attenuated to 0.00012V at the output as noise. 0.12mV is very good.

The formula i used is the first one in this document:
http://www.ti.com/lit/an/slyt202/slyt202.pdf
 
Top