rectifier switching noise

[trem]

I have never actually heard it, just read about it.
But.....when a diode sees the polarity swing, and its Anode goes from Positive to Negative (from conducting to nonconducting) how does that cause noise exactly.?
I am not even sure what really happens.
The instantaneous polarity of the transformer swings faster than the diode can turn off.? So this will let electrons flow the other direction, at a time when that particular circuit plans for no current.?
Or does it still let electrons flow in the same direction, against a now Negative swing of the transformer.?
Thank You

[loudthud]

When the polarity reverses, electrons near the junction have to reverse direction. Once the junction is evacuated, electron flow stops rather abruptly (nano seconds). This sudden change excites any LC parasitic circuit that will ring for several cycles.


See discussion here: http://music-electronics-forum.com/t34483/

[trem]

OK.....Thanks for the link.

[Jazz P Bass]

I have a book here by William Sloot.

To quote:
" The low value capacitor across the rectifier diode is to minimize so called 'silicone radiation'.
The phenomenon occurs because a silicone diode is such a good rectifier.
When the ac sine wave reverse biases the diode, it cuts off very abruptly.
The abrupt cutoff causes rapid current changes which shock associated inductive circuitry into momentary oscillation.
The result of this oscillation is high frequency harmonics modulated at a 60-Hz rate."

He goes on to say, in radio circuits, the silicone radiation may become evident as 60- or 120-Hz hum.
Tv circuits are also affected.

What I found intriguing is this statement:
"The capacitor across the diode should be considered a 'tuning' capacitor and not a 'suppression' capacitor or 'swamping device'.
For this reason, a small value capacitor can often eliminate silicone radiation, wheras a larger value may intensify the problem."

William Sloot Solid State Servicing 1972

[Chuck H]

I have a book here by William Sloot.

To quote:
" The low value capacitor across the rectifier diode is to minimize so called 'silicone radiation'.
The phenomenon occurs because a silicone diode is such a good rectifier.
When the ac sine wave reverse biases the diode, it cuts off very abruptly.
The abrupt cutoff causes rapid current changes which shock associated inductive circuitry into momentary oscillation.
The result of this oscillation is high frequency harmonics modulated at a 60-Hz rate."

He goes on to say, in radio circuits, the silicone radiation may become evident as 60- or 120-Hz hum.
Tv circuits are also affected.

What I found intriguing is this statement:
"The capacitor across the diode should be considered a 'tuning' capacitor and not a 'suppression' capacitor or 'swamping device'.
For this reason, a small value capacitor can often eliminate silicone radiation, wheras a larger value may intensify the problem."

William Sloot Solid State Servicing 1972

This is the first I'm reading that this phenomenon can manifest as hum. And the idea of tuning the circuit (or de-tuning?) rather than snubbing is contrary to every other treatise I've read on the matter. But this is from a guy in the 70's that seems to have practical experience dealing with it rather than just regurgitating a paper theory. Very interesting.

[Dave H]

This is the first I'm reading that this phenomenon can manifest as hum.

It sounds more like a buzz than a smooth 120Hz hum. Its a very short burst of HF ringing at 120Hz repetition rate which is radiated so you hear more of the higher harmonics of 120Hz. Soft recovery diodes can reduce the buzz.



Edit
I forgot to say, when viewed on a scope the buzz looks like a series of spikes at 10mS intervals (8.33mS in USA).

[trem]

Have any of you guys actually experienced this.?
Maybe heard a noise, had a hard time diagnosing the cause, and then fixed it by using one of the methods mentioned.?
To my knowledge, I have never heard this.
I know Marshall and Vox used snubbers across the HT Secondary to avoid the switching noise.
But I do not think Fender "ever" used snubbers, or fast switching diodes. Did Fender (tweed and bf/sf) amps have a problem with rectifier noise.?
Thanks

[Jazz P Bass]

From what I can gather, the problem manifests itself the most on very small signals & in radio & TV applications.

Although I have seen them a lot on HiFi equipment like Yamaha.

[nashvillebill]

I have installed snubbers on the diodes, as well as making power supplies without the snubbers, and I can't hear a difference. Of course, perhaps I have tin ears. But there sure seem to be a lot of good-sounding amplifiers (both hi-fi as well as music instrument) that don't have snubbers. So is this a "fix" for a problem that mostly exists in theory and/or viewed on the 'scope? I say mostly because perhaps there are some amp designs where the diode switching noise manifests itself.

[trobbins]

For typical valve amp B+ supplies, the relatively slow dV/dt of the mains sine wave at the time of diode turn-off, and the relatively low load current, means that the dI/dt experienced by the diode is way below what any ss dataheet shows for a reverse recovery test conditions. So the 'gross' reverse recovery datasheet simulated, or actual rr waveforms are not experienced in a typical amp.

If however, you are attempting to have a very stiff B+ (or dc heater supply), with a low source resistance transformer, and large first filter cap, and a poor ss diode, then yes the rev recovery glitch gets worse. It's easy to see a diode's rev recovery performance on a cro by driving a diode with a signal generator at say 10kHz - old ss diodes, and 1N4007, and higher current modern grunt ss diodes have a very discernible reverse glitch - whereas the UF4001 is clean.

The 'problem' for ss diode rr is normally related to Mhz splatter affecting EMI radiated compliance in other equipment (especially smps). In audio equipment it is a 2f mains repetitive glitch, so with poor power supply wiring layout and filter management, can couple in to the amp as a 2f mains harmonic rich signal.

[Chuck H]

But I do not think Fender "ever" used snubbers, or fast switching diodes.

A whole lot of Fenders from the mid 70's forward used snubbers on a diode bridge rectifier. Pretty much all the models where the schem reads Fender/Rogahn/Rhodes, and probably a lot of the Randall/Fenders as well.

[nashvillebill]

Yet the Fender Princeton Chorus and Ultimate Chorus of the late 80's/ early 90's didn't use snubbers, and they used 1N5402 diodes. Both SS amps with what appears to be reasonably stiff power supplies.

edit: I'm curious, is the 1N5402 any faster recovery than a 1N4007?

[trobbins]

Using the signal generator test (5kHz, 10Vpk-pk, 100 ohm source so 50mA peak diode current) , a 1N5408 was about as bad as a 1N4007 and vintage devices like HR25. A larger P600K was worse.

[nashvillebill]

Again out of curiousity, how about the old selenium rectifiers? Not that I'd want to go back to using them! (I've got a couple of unused ones still in their boxes, I believe)

[trobbins]

Selenium diode is metal junction so no minority carrier RR effect. Like schottky or valve diode. Exhibits leakage i think, and certainly low voltage withstand.