Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
05/18/06 04:06
Modified:
  05/18/06 04:06

Read: times


 
#116457 - Ripple filter
Responding to: ???'s previous message
Bruce said:
I'm working on a new design now that requires a 1.1Mhz switching regulator at 600mA. I'm going to use an LC filter on the output and this is my first time using SMT inductors. I'm not very familiar with what is commonly used, but I found this basically going through a digikey catalog.

http://www.sumida.com/en/products/pdf/CDRH74.pdf

My requirements are basically:
22uH
600mA
SMT



Ok, you need this choke for the ripple filter like it's shown in this application, right?




Well, for this ripple filter it's not needed to exactly hit the 22µH. The given value can be altered within a large range. To be honestly, the only requirement for this ripple filter choke is, that its value is smaller than that of the storage choke, which is 150µH in the above example.

The other point is, that the dampening of ripple also highly depends on the quality of filter cap. At 1MHz (the ripple frequency of your switcher) it's mainly the equivalent series resistance (ESR) of your filter cap, which determines the dampening performance. A 10µ/16V aluminium electrolytic shows an ESR of about 5Ohm (extreme deviations possible). A 100µF/25V aluminium electrolytic shows an ESR of about 0.5Ohm. A 10µF/25V tantal electrolytic shows an ESR of 0.2Ohm and a 22µF/6.3V/X5R ceramic highcap (Taiyo Yuden) shows an ESR of only 0.003Ohm!

So, depending on which cap you use the same choke can give you a dampening performance varying by a factor of 5/0.003 = 1700!
Means, using this superb ceramic high cap allows you to drastically decrease the value of filter choke while still yielding excellent ripple dampening.

The choke you have chosen suffers from the interwinding capacitance of the many turns of wire bend on the core. This interwinding capacitance short circuits the inductance at high frequencies, with the effect, that there's only poor dampening performance at the higher harmonics of your 1MHz ripple!

A much better approach is to use a choke with as few windings as possible. The optimum choice would finally be the ferrite bead, where a piece of wire is fed through a bead made of soft ferrite material. "Soft" means, that heavy ohmic losses occur above about 10MHz, improving the dampening performance at the ripple harmonics.

A good compromise between increasing of interwinding capacitance and improving of dampening when increasing the turns of a choke shows the 6-hole-choke. Such a choke can offer up to 10...22µH of inductivity while at the same time offering highest impedances at up to 1GHz, which is totally out of reach with the choke you have choosen!
There are also SMD ferrite beads available, which offer about 1...10µH of inductivity. Mostly a piece of wire surrounded by a rectangular ferrite bead, if a high current version is choosen. Yes, even 6-hole-chokes in SMD style are available.

So, don't focuse too much on this choke. It's the cap which determines the dampening performance!

Kai

List of 7 messages in thread
TopicAuthorDate
Inductor types for a switching regulator            01/01/70 00:00      
   Core type critical            01/01/70 00:00      
   I did not specify, I left that to someon            01/01/70 00:00      
   Where is the inductor located?            01/01/70 00:00      
      I see no way this can be anything but "f            01/01/70 00:00      
         Erik is correct            01/01/70 00:00      
   Ripple filter            01/01/70 00:00      

Back to Subject List