Some questions about ferrite selection

[00940]

Hi,

I'm in the midst of finalizing a PCB and its BOM before sending it to oshpark.

And of course a question came... what ferrites to pick ? The offer is quite wide, with ferrites ranging from a few dozens of ohms to 2K @100mhz, some peaking sharply, some whose impedance is rising significantly even at 10mhz (Laird LF1206 for example).

Would someone have some info on ferrite beads selection ? An article to link to maybe ?

The IC to decouple is a PCM5102 but I also need ferrites for a USB input filter.

Thanks for anything on the topic.

PS: Schematic and layout of the DAC attached if it helps.

 

[29285]

Decoupling with ferrites is a good idea. I recommend Ferrites with a broad damping curve, not the peaking ones. Otherwise a high Q-factor bead, in conjunction with an ultra low ESR MLCC blocking capacitor form a resonant tank with high Q-factor. Any pulsating current may excite that tank and thus produce overvoltage ringing that easily can destroy the IC.
A safe way to avoid this is parallelling a low ESR electrolytic cap to the 100nF Blocker. This tunes the circuit to a lower resonant frequency and lowers Q-factor thus damping resonant overvoltages.
The Z-value of beads is not critical, 100Ohm/100MHz or more should do the job.

Concerning usb filtering: you need special common mode chokes for usb to guarantee signal integrity.

There exists a library of Wuerth-Electronic ferrite beads suitable for LTSpice that comes in handy on selecting a choke.

 

[Avro Arrow]

How about something like BLM21AG471SN1D?

 

[00940]

Decoupling with ferrites is a good idea. I recommend Ferrites with a broad damping curve, not the peaking ones. Otherwise a high Q-factor bead, in conjunction with an ultra low ESR MLCC blocking capacitor form a resonant tank with high Q-factor. Any pulsating current may excite that tank and thus produce overvoltage ringing that easily can destroy the IC.
A safe way to avoid this is parallelling a low ESR electrolytic cap to the 100nF Blocker. This tunes the circuit to a lower resonant frequency and lowers Q-factor thus damping resonant overvoltages.
The Z-value of beads is not critical, 100Ohm/100MHz or more should do the job.

Concerning usb filtering: you need special common mode chokes for usb to guarantee signal integrity.

Thanks, that's very helpful. It helps a lot to make sense of this app note from Altera and of the spice analysis circuit they offer.

TDK and Laird are both offering ltspice models btw, which is quite practical.

Wrt USB: yes, I found special offerings from Murata on CM chokes for data lines. I was more looking for the ferrites on the power and gnd lines.

 

[Avro Arrow]

Murata has a filter simulator.
You can get it here.