What does .01 cap on bridge rectifiers do?

[reiver]

Per an old thread, I'm been playing around with using four IXYS FRED rectifers to create a 25A bridge rectifier for use in my modded Hafler amplifier.

The new bridge works fine but I have a question:

- All stock Hafler amps use a .01mf capacitor across the AC leads on the stock 25A 'square' bridge rectifier.

- What is the purpose of putting a tiny value cap on a bridge rectifier? Is it necessary when using IXYS and/or Schottky rectifiers/bridges?

I've read some people say not use a cap on Ixys/Schottky, but I"m interested in what their purpose of the cap is is before I clip it off.

[KBK]

They are there to shunt switching noise. If the capacitors are of the wrong type, they can create more noise than they attempt to remove. The size of the cap vs the expected 'load' is key to their effectiveness.

With a linear Power Supply on a 'standard' power amp design, the rectifier switching noise can come through in a subtle fashion as the switching diode IS connected directly to the given PS rail and going from the charge state to the 'store-release' state and thus modulating the output--when the amplifier's given devices are 'active'. Now, the rail can be modulated by the active device, and thus make the diode junction do more than one single on-off cycle as the cascading voltage/current load is moving toward it's on/off point, and therefore, a well or properly 'shunted' diode junction is considered to do less harm to the output signal.

Once again, though..the loading of the junction is different at all levels and thus the shunting capacitor can have more or less effectiveness. The situation is not static due to music signals never being repetitive, so it can even be done (shunted/value-type-quality-loading change/no shunting) to personal taste, to some degree. The more local the capacitor to the diode itself, the more effective the capacitor. Right on the leads itself is the most effective. Fast switching diodes are not necessarily less noisy (to the human ear) than a regular type diode. Schottky diodes are considered to be fast but 'soft switchers' (as compared to regular 'ultra fast' diodes-which are considered to be 'hard sounding' and noisy switchers by many audio folk) so are far less likely to need the shunting caps under a greater variety of loading scenarios.

However, the saving grace is that the circuit itself has standardized static values of components and thus the system can be predicted on one level.

Every diode type seems to have its give and takes. If the given circuit has a voltage to be utilizing a diode where the possible peak reverse values can accept a Schottky, it is generally considered these days that this is the best scenario. However, schottky's tend to not be used in critical spots in commercially released designs as the failsafe levels are not great enough to cover the larger possible situations that the given audio gear may be subject to. A combination of costs vs. reliability in the field causes most commercial audio gear to use regular and ultra fast diodes in critical spots. Thus the shunting capacitors found at many a rectifier in the audio world.

Thus, for us, the major use of schottky diodes tends to be in hobby and self built circuits, when it comes to marginal reliability issues.

[reiver]

So I'm guessing it's a good idea to always have a cap of some sort on a bridge rectifier to cut potential noise?

From what I've read a cap is not needed on soft recovery diodes/bridges because they are low noise, and that putting a cap on can somehow affect performance. Can any enlightened soul verify this?

-Bryan

[andy_c]

Originally posted by reiver
I've read some people say not use a cap on Ixys/Schottky, but I"m interested in what their purpose of the cap is is before I clip it off.

Hi,
The best info I'm aware of on this subject comes from this excellent article by Jim Hagerman.

[KBK]

Standard bridge: high grade cap - on the pins if at all possible.

Soft recovery: not needed.

Ultra fast: High grade cap - On the pins if at all possible.

This is with regard to bridge rectification of PS rails, only.

The values are low as it is peak switching noise shunting here, that's all.

[reiver]

Interesting stuff...thanks for the quick replies.

From Hagerman's article it appears that a cap only reduces the frequency of the ringing, not dampens it out. A resistor/cap snubber looks like the way to go but calculating the values could prove tricky.

I think I would feel right in simply upgrading to a soft recovery bridge rectifier such as IXYS which is available up to 68A.

"Keep it simple"

-Bryan

[KBK]

It is possible to damp the noise out with the right kind of circuit- but far more effective and cheaper to use the correct diode type in the first place.

For example, half the capacitance (two caps) and a 500kohm between the two caps, tied to ground -this, as a three piece arrangement- across the junction. Vary the resistor-tune to suit. (CRC arrangement)

Drops like a rock.

I just looked at the problem, and fixed it. I dunno if anyone's ever done that particular trick before.

[Eva]

Bypass capacitors on diodes?

Diode capacitance changes abruptly depending on reverse voltage and after reverse recovery.

The purpose of these capacitors is to provide a much smoother overall capacitance and prevent voltage spikes.

This applies to all diodes. RC snubbers are routinely used with schottky and ultrafast diodes in SMPS.

[Nico Ras]

Originally posted by andy_c


Hi,
The best info I'm aware of on this subject comes from this excellent article by Jim Hagerman.

Thanks Andy, that is well written paper.

Nico

[Eva]

In the following thread, the same principle described in Jim Hagerman's paper is applied to a real circuit. There are plenty of real oscilloscope captures showing the problem and the improvements.

http://www.diyaudio.com/forums/showt...664#post753664

Only two capacitors and one resistor are required for each diode bridge or transformer winding.