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

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
05/11/06 14:29
Read: times


 
#115988 - pi-filters
Responding to: ???'s previous message
Erik said:
stating the often overlooked obvious: do not use a tantal at the input to the pi filter, I just use a 100nF at that place.

Of course!
Anyway, I never use tantals at places, where you don't know what can feed the tantal, what surges, what bursts and what lightning overvoltages can arrive. So, I always omit tantals at locations like this.

In such a pi-filter I only use aluminium electrolytics of 47...220µF each, depending on load current, mostly paralleled by 100nF/X7R. The choke is a soft ferrite bead showing an equivalent inductance of about 1...10µH, depending on load current, which was made of a 6hole-core in earlier days or is one of the now available superb SMD versions.

This arrangement does not need any additional series resistance (in series to the choke, of ourse), because the unavoidable equivalent series resistance of aluminium electrolytics of some 1/10 of Ohms helps to suppress ringing. Because of that the seeming unadequate aluminium electrolytics are just the right choose, also the rather small inductance of soft ferrite beads, according to the formula

R > SQRT(2 x L / C)

which is the criterium for the avoiding of ringing.

A series resistor can be added nevertheless, tremendeously enhancing the filtering performance at lower frequencies. Then, only the additional voltage drop and the current rating must be taken into consideration.

Such pi-filters do a wunderful job in my applications and you will never see a circuit of mine neglecting such a pi-filter!

Of course, additional measures can be added to the pi-filter, like overvoltage protection, current fusing or else. But that's a different story...

Kai

List of 13 messages in thread
TopicAuthorDate
Tantalum on power supply filter - Blowup            01/01/70 00:00      
   does it get (luke)warm ?            01/01/70 00:00      
      Nope            01/01/70 00:00      
   Symbol is marking the "+"-pole!            01/01/70 00:00      
      Verified            01/01/70 00:00      
   Surge current            01/01/70 00:00      
      Some interesting reading            01/01/70 00:00      
      Much thanks!            01/01/70 00:00      
         Pardon the pun....            01/01/70 00:00      
         Tantals aren't unreliable!            01/01/70 00:00      
            not a tantalum pi            01/01/70 00:00      
               pi-filters            01/01/70 00:00      
            Guilty            01/01/70 00:00      

Back to Subject List