PI-Filter

Hi there,

I´m thinking about modifying the powersupply of my power amp with a PI filter. It now consists of a 450VA 2x30Volt transformer and 2 x 50 x 1000 microF per channel. I would like to change this to 2 x 15 x 1000 microF, 2 x 2mH, 2 x 50 x 1000 microF per channel.

Would this be a usefull size or should the first capacitor be bigger? (The amp is biased at 1 amp but can draw up to 15A eff.)

I´m thinking of winding 4 air chokes of about 70mm diameter,20mm thick.

Is it neccesary to shield these from the rest of the amp?
What material would be the best to do the shielding (if neccesary)?

william
 

Nelson Pass

The one and only
Paid Member
2001-03-29 12:38 am
You only have to watch the ripple current
rating of the primary capacitors - it needs to
be about 2-3 times the DC use you expect.

By the way, I've really come to like the ERSA
inductors available from Zalytron, as they have
low resistance and can take a lot of DC before
they saturate.

I have not had to shield the chokes, but a little
distance from the low level circuitry, say 6 inches
or more, is always a good idea.
 
Thanks for the quick answer!

Since I build the casing for this amp about 15 years ago (when still at university with all the beautifull milling and drilling maschines) and it has undergone quite a lot of changes inside it will be very difficult to keep the chokes 6" away from the low level circuitry (the casing´s only 440 x 400 x 150 ) More like 3" maximum, so I´m afraid I´ll have to shield them.
The question is with what.

For the capacitors I will use Panasonic FC types wich are rated at 2.2A @100kHz for a 1000 microF/50V capacitor. I´ve tried them in my preamp and cd and they sound really good while being less expensive as an ELNA Cerafine or Silmic.

william
 
Hi All,
I use two 10000 µF and two 22000µF caps and two 15 mH chokes in my Ampzilla. The 10000 's are at the rectifier side.
I started with the 2.5 mH coil from the Klipschcrossover but with the 15 mH the sound is better. The inductor I use is from www.intertechnik.de part#1340279 , has a DC resistance of 0.57 Ohm, ferrite core.
The ripple is almost too low to measure on my scope.
Hope this helps.
 
William,

An inductor of 2 mH represents an impedance of 1.25 ohms at 100 Hz. This value is too low.

In order to obtain a good efficiency with your filter, you have to use inductors of 10 mH and more. But air inductors of such value are very big and expensive. That is why iron-core inductors are used for this purpose. Iron-core inductors don't requires shielding, and can be diposed anywhere in the amplifier cabinet.

Regards, P.Lacombe.
 

Nelson Pass

The one and only
Paid Member
2001-03-29 12:38 am
Actually 2 mH works fine. The lowest frequency it will encounter
will be 120 Hz, and the impedance there will be about 1.5 ohm.

( Z = 6.3 * L * Freq )

The cap which follows the inductor needs to be low impedance
at this frequency ( Z = 1 / ( 6.3 * C * Freq ). A 10,000 uF cap
will be .13 ohms, so we are already seeing about 20 dB
reduction at 120 Hz.

Keep in mind that much of the noise in the ripple is at even
high frequencies, and they are attenuated even more.
 
Mr Pass,

I'm sorry, I cann't agree with you. Attenuation of 20 dB can be insufficient to avoid audible hum in the vicinity of a good loudspeaker : 20 dB is only 1/10 in voltage, and, depending of the supply rejection of the amplifier, significative residual signal at 120 (or 100) Hz can be measured at the output...

Regards, P.Lacombe.
 
AudioFreak said:
Yes Keith,

But ... the solid state amps such as the Zen have really poor PSRR so a clean supply is critical.

Accepted. Does this mean that the Pass amps are poorly designed? :)

The design setup for a PI filter between a PSU and an amplifier is NOT intuitive. The filter is just that... a filter and a low pass one to boot. This means that it'll have all the electrical characteristics of a filter and, if not properly handled, may be worse for the amplifier than a capacitive power supply.

Again, see the Radiotron Designer's Handbook for all the gory details.

regards all, and have a great musical new year.
Keith
 
For those who don't remember a certain thread from last summer, Keith likes to argue more than anything else in the world.
As for intuitive...PI filters seem quite reasonable to me; make perfect sense. Low pass filter? Of course it is. But so's a simple cap. The cap just has a lower slope: 6 dB vs 18 dB. Caps work fine. Inductors work fine. Caps and inductors work fine together. For that matter, if inductors bother you, you can use a resistor between the caps in a PI filter, albeit with some loss of power.
I think that you'll find that many circuit elements have poor PSRR. Does that mean that every stage in the circuit has to be a differential? Of course not. Don't be silly (or should I say argumentative?). Even portions of circuits with good CMRR benefit from a quiet power supply, even if the residual is, say, 80 dB down. Why have unnecessary racket?
(The Radiotron Designer's Handbook aka Radio Designer's Handbook is the third book down on my shelf even as we speak.)
Please, Keith, if you're going to start arguing purely for the sake of arguing again, at least refrain from foolish statements such as the PI filter being a low pass filter, whilst ignoring the fact that a simple cap is also a low pass filter. It reflects poorly on your comprehension. You did enough damage to your credibility last summer without chipping further at it now.

Grey
 
Hi,
Whether the PI-filter is a lowpass filter or not does not bother me, but most power amplifiers have <B>un</B>regulated supplies.
The PI-filter did improve the <B><U>SOUND</U></B> of my GAS Ampzilla amplifier and that is interesting.
I also experimented with larger values for the inductor up to 22 mH but got no further improvement.:)
 
Power supply can be critical, even with a good *apparent* PSRR. Extensive experiments have demonstrated that strange behaviours can occur in some circumstances, in class AB amplifiers, specially BJT, whose the idle current is (necessarly) low.

Hum at output can be inaudible, but sound quality can be poor when the listenig level is high : increase of power consumption causes more hum, which is inaudible because of the loud music, but the amplifier act in the same way than a balanced mixer or ring modulator, causing intermodulation products at +/- 120 (100) Hz of each music tone. This can be clearly seen with a spectrum analyser, or a good sound-card in a PC compatible, and the suitable software.

This is why I always use a very good filter design in PS filters. Iron-core inductors don't causes any distortion, even saturated, because audio signal don't flow across them, don't need shielding, and are less expensive than coreless ones.

Regards, P.Lacombe.