John Curl's Blowtorch preamplifier part II

[pooge]

Originally Posted by john curl

Lastly, a large common mode choke MIGHT be a rational choice for this component, as it would be significantly smaller and lighter, due to the fact that the DC components of the positive and negative supplies could cancel and keep a relatively small choke from saturating. However, there may be hidden problems, that I, at least, have never looked at, when it comes to a large power amp that should be noted.

I see a potential problem with a common mode choke if the star ground it connected to the transformer center tap. If trying choke, I would favor a bridge on each secondary, and not ground PS until after the final caps. But I would do this, anyway, whether or not choke is used.

[john curl]

AH, pooge, you are discovering the secrets of the Krell, I mean Blowtorch. ;-)

[Joachim Gerhard]

I work with Sowter and they are very flexible. When i have the spec. i could ask Mr.Sowter if he can design a choke for poweramps.
I also heard that putting a small resistor ( say 0.5Ohm) between two fiter caps improves ripple rejection much compared to a bigger cap with the same value then the two combined.

[bocka]

Originally Posted by Joachim Gerhard

I also heard that putting a small resistor ( say 0.5Ohm) between two fiter caps improves ripple rejection much compared to a bigger cap with the same value then the two combined.

Yes, that forms a really good RC filter with a very low cutoff frequency of about 10Hz. In conjunction with high speed rectifiers probably the best (and most simple) solution to get rid of hum and noise from the power supply.

[simon7000]

Originally Posted by Joachim Gerhard

I work with Sowter and they are very flexible. When i have the spec. i could ask Mr.Sowter if he can design a choke for poweramps.
I also heard that putting a small resistor ( say 0.5Ohm) between two fiter caps improves ripple rejection much compared to a bigger cap with the same value then the two combined.

Putting two identical capacitors in parallel would halve the ripple.

A typical 22,000uf capacitor at 120 hertz would have an impedance of .06 ohms.

20Log (.06/.5)=-18.4db. A big improvement

[Joshua_G]

Originally Posted by simon7000

Putting two identical capacitors in parallel would halve the ripple.

A typical 22,000uf capacitor at 120 hertz would have an impedance of .06 ohms.

20Log (.06/.5)=-18.4db. A big improvement

Using 2 capacitors of the same make in parallel, say 2 x 10,000uF instead of 1 x 20,000uF will improve also the ESR.

[pooge]

Originally Posted by Joachim Gerhard

I also heard that putting a small resistor ( say 0.5Ohm) between two fiter caps improves ripple rejection much compared to a bigger cap with the same value then the two combined.

RC filter is a single-pole filter.
LC filter is a two-pole filter. Sharper cutoff than RC filter.

[pooge]

I'm wondering how drastic the emission from a choke would be inside the chassis. Them charging currents are pretty high and short. The chokes would help smooth the spikes, but I would think they would have to be isolated from the amp circuitry in some way, and not just distance or orientation.

[martin clark]

Originally Posted by pooge

LC filter is a two-pole filter. Sharper cutoff than RC filter.

IME, CLC like that works fine providing L is picked so that ESR of the caps make the whole at least critically damped. Air-core inductors of around 10uH are easily handwound; Z= SQRT(L/C) = SQRT(10uH/10000uF) = ~30mR, comparable with the ESR of such caps...

Even such tiny L puts the filter knee at just 500Hz, low enough to be quite useful.

[pooge]

Yes, damping an LC filter must be considered. Something not to worry about with RC filter.