What are your subjective experience with decoupling the rectifier diodes? I dont mean a snubber across the secondary but rather a cap across each diode. With 1nf mkt caps across the silicon diodes im using to feed my 1875 amp, i notice a significant difference in the sound. With the decoupling in place the sound is heavier, darker (less treble energy) and the sound stage more closed in. Without the decoupling i get most of the lost ambience back but knowing that much of its sound character could be due to the unfiltered harmonics of the diodes i cannot decide which configuration is a superior
I cant imagine why it would make a difference, they are usually there to kill switching spikes. Any subsequent smoothing will kill any changes made by the small caps.
Try searching for diode snubber circuits, you’ll find a bunch here on the forum.
Probably can get better results with less parts.
Probably can get better results with less parts.
Snubbers are redundant these days with fast switching diodes.
I had a valve mixer that buzzed badly and it was switching spikes from rectifier diodes.
Snubbers didnt work but HER158G fast diodes did.
I had a valve mixer that buzzed badly and it was switching spikes from rectifier diodes.
Snubbers didnt work but HER158G fast diodes did.
A cap across the diodes will conduct HF noise on the power line across them when they are in blocking mode. No need for the caps at all.
Caps across all four diodes will conduct HF noise and short circuit it. It is usually not needed in audio but is a must in a radio.
10nF + 10ohm over each diode. I also always place an inductor in series with the bridge before the big bulk cap.
It doesn't matter whether you put individual caps on each diode or one cap across the transformer secondary. The purpose of the cap(s) is to eliminate the ringing of the parasitic inductor in the secondary circuit when the diodes turn off. A C-only snubber will reduce the frequency of the ringing to the point where it no longer couples effectively into sensitive audio circuits. To actually dampen out the ringing, you need an RC snubber.
There are other considerations as well, including selecting the right diode type for rectification.
I write quite a bit about rectification and snubbers here: Taming the LM3886 Chip Amplifier: Rectification and Snubbers – Neurochrome
Tom
There are other considerations as well, including selecting the right diode type for rectification.
I write quite a bit about rectification and snubbers here: Taming the LM3886 Chip Amplifier: Rectification and Snubbers – Neurochrome
Tom
No, it does matter, if we are talking not about only audio range of signal frequencies.
I'm aware of that. Did you read the article I linked to in Post #9?
The issue is that the parasitics in the secondary circuit of the power transformer will resonate when the diodes turn off. You can add an RC anywhere in the secondary circuit as long as it ends up in parallel with the secondary winding of the transformer. The parasitics are distributed, not lumped components.
Tom
The issue is that the parasitics in the secondary circuit of the power transformer will resonate when the diodes turn off. You can add an RC anywhere in the secondary circuit as long as it ends up in parallel with the secondary winding of the transformer. The parasitics are distributed, not lumped components.
Tom
Actually, the ringing at HF/RF can cause "un-intended conduction" of PN junctions and all sorts of issues with BJT's and opamps, such as unexplainable DC offset, surprising distortion.
Discussed in an Analog Devices application note.
Yes, of cause. I just want to add that parasitic ringing described in your article is not the the only process going in a power lines circuit.
I'm talking about another fact that power lines act as antenna (at HF). And we don't want this antenna to be 'blinking' (to be modulated) by diodes when they turning on and off. That's why we may put a path for HF frequency with small caps so 'antenna' was always 'connected' through caps (but not 50/100 Hz blinking). Or we may use common mode filters (ferrite chokes) to get rid of this 'antenna' (or to lower an effect).
Here is a good collection (Post #15) of all the possible answers.
Last edited: