Linear regulator Question.

[kvresto]

Hi Everyone. A previous post here on the forum has prompted this question as I'm in the middle of a similar project.

I have a project fitted with an LM2593 150khz switcher, generally it could be any type of DCDC switching supply, and I want to use this 5V output as the input to a LDO for a 3.3V supply.

I have noticed some linear voltage regulators are listed as good for "Post-Regulation for Switching Supplies". I'm not too familiar with the details, so what makes a regulator better than another for switching supplies, and what criteria do I look for in a data sheet (my biggest issue) to help me establish if the part I'm looking at is suitable for Post-Regulation for Switching Supplies.

Thanks for your help.
kvresto.

[kizzap]

You are looking for the devices PSRR, the Power Supply Rejection Ratio. It is the relationship of the input ripple to output ripple. The higher the number, the better it is at blocking ripple.

For a post-switcher regulator, you ideally need a regulator with a high PSRR at and around the frequency of the switcher.

[yramgu]

And pay attention to what the datasheet says: sometimes manufacturers will advertise very high PSRR but at very low frequencies (~10's of Hz), when what you need is high PSRR in the 10's kHz range

[Kleinstein]

PSRR is important in the range from about 10 kHz (about the bandwidth of the switcher's regulator loop) up to something like 5 times the switching frequency. AT the higher frequencies an extra LC filter in front of the LDO may help. Also layout can be important.

Generally with LDOs have a close look at the output capacitor specifications. A regulator that can use a ceramic output capacitor can make things easier.

[cellularmitosis]

Are capacitor multipliers fast enough to have any effect on switching spikes?



http://www.radio-electronics.com/info/circuits/transistor/capacitance-multiplier-circuit.php

Perhaps I should break out the scope and a switcher this weekend!

[Alex Eisenhut]

Are capacitor multipliers fast enough to have any effect on switching spikes?



http://www.radio-electronics.com/info/circuits/transistor/capacitance-multiplier-circuit.php

Perhaps I should break out the scope and a switcher this weekend!

...or LTSpice...

[Alex Eisenhut]

Hi Everyone. A previous post here on the forum has prompted this question as I'm in the middle of a similar project.

I have a project fitted with an LM2593 150khz switcher, generally it could be any type of DCDC switching supply, and I want to use this 5V output as the input to a LDO for a 3.3V supply.

I have noticed some linear voltage regulators are listed as good for "Post-Regulation for Switching Supplies". I'm not too familiar with the details, so what makes a regulator better than another for switching supplies, and what criteria do I look for in a data sheet (my biggest issue) to help me establish if the part I'm looking at is suitable for Post-Regulation for Switching Supplies.

Thanks for your help.
kvresto.

Extra points for using a buck to get from 5 to 3.8-3.7, then use a LDO to get 3.3. Big power savings!

[Alex Eisenhut]

And pay attention to what the datasheet says: sometimes manufacturers will advertise very high PSRR but at very low frequencies (~10's of Hz), when what you need is high PSRR in the 10's kHz range

..or in unrealistic test conditions. Switchers now run at 1MHz+, so it looks like it's going to be a combination of passives and active circuits.

The balance is set by cost and PCB real estate.

Sure there are powerful LDOs but if they cost 5x what an LC costs... But the L can drop 10s if not 100s of millivolts.

The best approach is to simulate but if there's no model, well what can you do? And what happens if you put the switcher's voltage feedback point after the LC filter?

[kvresto]

Wow, great, just what I need to know!!

And what happens if you put the switcher's voltage feedback point after the LC filter?

OK, what happens? my test cct has the feed back before the LC filter. I suppose I could change it a see.

kv


EDIT:

Extra points for using a buck to get from 5 to 3.8-3.7, then use a LDO to get 3.3. Big power savings!

Actually The buck brings down a vehicles batt voltage from max ~ 14.2V to 5V. Then a linear reg gives me the 3.3V. The 5V is used for comms stuff as well, but the 3.3v is mainly for amps, ADC, CPU. At least that's the design at the moment on paper, I've only started doing the power supply research now.

[Alex Eisenhut]

Wow, great, just what I need to know!!

And what happens if you put the switcher's voltage feedback point after the LC filter?

OK, what happens? my test cct has the feed back before the LC filter. I suppose I could change it a see.

kv

Unknown, but you're adding phase to a control loop, your LDO might become a Large Dynamic Oscillator.

[kvresto]

That's bad for sure.