Hi,

I use different buck-boost and boost voltage regulators to generate typically 3.3V and 5V from a 3.7V Li-Ion battery in my designs. I notice that the ripple is a bit high; typically 40-65mV. I have added 2-3 22uF ceramic capacitors and a 100nF capacitor on the input and output of the regulators, but I still have the high voltage ripple.

How can I get rid of/minimize the voltage ripple?

 

Several things may be considered, one is to increase the switching frequency keeping the capacitors the same, another is to use tantalum electrolytic capacitors instead of aluminium electrolytes.

The tantalum capacitors have much lower ESR than the Aluminium capacitors which should decrease the peak to peak ripple voltage.

Let me know how it goes. email - nalamgir118@gmail.com

 

Schematic please?

What kind or ripple; Spikey or wavey? Switching frequency or load current frequency?

It sounds like you don't have any inductors beyond the one involved in the switching. Various kinds of inductors are helpful in reducing various kinds of noise.

 

I think it is mainly switching frequency I have the most of.

That is correct; I don't have any inductors besides the switching inductor. I think I will try to add a ferrite bead and a capacitor to form a LC filter and get rid of some of the ripple.

 

I use different buck-boost and boost voltage regulators to generate typically 3.3V and 5V from a 3.7V Li-Ion battery in my designs. I notice that the ripple is a bit high; typically 40-65mV.

Noise is one of the disadvantages of switching regulators, but it's usually at a frequency that's easier to filter. What frequency is the noise? Is it under no load conditions? If not, state load.

I have added 2-3 22uF ceramic capacitors and a 100nF capacitor on the input and output of the regulators, but I still have the high voltage ripple.

22uF ceramic capacitors?? Those would be huge... Try a 100uF or larger electrolytic.

A schematic would be helpful, as would specs on the inductor being used and the desired load current.

 

The 22uF capacitors are 0805 package.

I have used the LM3668SD-2833 in a 3.3V setup. I see that with light loads, <100mA, the ripple is a bit too high.

 

The 22uF capacitors are 0805 package...

And in fact, if your supply is 3.3V, you can get 6-volt ceramic capacitors of 100 uF in 1206. Ceramics are good for getting rid of high frequency noise because of their extremely low equivalent series resistance. You may find that you need a much smaller capacitance ceramic than aluminum electrolytic.

65 mV ripple doesn't sound so bad, unless you're doing audio or instrumentation of some kind. I've had bad luck trying to filter a switching supply to use with audio, and in my limited experience, always ended up using a battery with a linear regulator, or line transformer plus a linear regulator. The noise just seems to get in there no matter what, even if I try to do a separate supply for the analog, but it has a common ground.

 

The TI datasheet shows 150mV of ripple for 3.7V in/5V out.

What capacitor type (material) are you using? The datasheet recommends X7R or X5R.

 

Here is a nice read:
http://www.ti.com/lit/an/sbva012/sbva012.pdf

 

I have used ceramic capacitors, X5R, 22uF, 6.3V, 20%, 0805.

 

I have used ceramic capacitors, X5R, 22uF, 6.3V, 20%, 0805.

It will take some trial and error to find a combination of output capacitors that will give more noise reduction; though what you're seeing is better than what the datasheet suggests you'll get. The datasheet also said more capacitance isn't necessarily better; and for the regulator in question that may be true.

I had a step up regulator using MC33063A breadboarded, so I did some experiments. When I changed the output cap (aluminum electrolytic) from 470uF to 680uF, I saw no change in ripple. However, when I removed the 0.1uF ceramic cap I had in parallel with it noise increased significantly. The regulator switching frequency is around 25KHz, stepping up 5V to 12.5V; so I have more ripple than you're seeing...