diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Power Supplies (http://www.diyaudio.com/forums/power-supplies/)
-   -   Reduced diode switching by placing a cap across (http://www.diyaudio.com/forums/power-supplies/160589-reduced-diode-switching-placing-cap-across.html)

Goetz 3rd February 2010 08:23 PM

Reduced diode switching by placing a cap across
 
Hi,

meanwhile I read twice about ways to lessen or change the effect of diode switching (i.e. placing a capacitor over the rectifier diodes, chosing fast recovery diodes, etc.) concerning PSU for solid state amp.
I have the following questions:
1.Do you have experience with placing a cap over the rectifier diodes and do you have circuit plan or sketch where to exactly to place the cap between the diodes?
I guess it shall be a small, let's say 10nf fast cap like silver-mica or polystyrene?
2.Schottky diodes are preferred against bridge rectifier due to higher switching speed. Now there is a producer IXYS with GaAs Schottky Diodes
They shall have
"Faster Reverse Recovery Time than UltraFast and Silicon Carbide
Much lower forward Voltage than Silicon Carbide
Isolated, low inductance, surface mount,..."
Do you know have experience with this "new" diodes ?
Have a nice day and thank for your efforts in advance.

Iain McNeill 4th February 2010 12:14 AM

Taking an ultra-fast recovery diode and slowing it down with a parallel cap seems rather counter intuitive to me. Especially considering the slow moving 50/60Hz sinewave being applied. As the mains voltage reaches the capacitor voltage, the diode gradually becomes forward biased and a current starts to flow. As the mains voltage starts falling, the current drops to zero and then the diode is reverse biased.

Why do you need nanosecond diodes in a traditional PSU?

SMPS - that's a completely different case.

Joachim Gerhard 4th February 2010 12:24 AM

I think there are "zerro recovering" diodes fron Cree.

jackinnj 4th February 2010 12:28 AM

You want to damp the energy of the switching diode -- this can be in the low tens of kHz to hundreds of kHz in the linear power supplies we use -- the problem isn't that we can hear the switching noise -- it's that the high frequency noise propagates easily and may cause the PN junction of all the semi's in your DUT to conduct.


A capacitor alone isn't sufficient -- google "diode snubber"

Eva 4th February 2010 02:11 AM

A capacitor in parallel with a diode just moves the characteristic LC resonance down in frequency (from diode C and circuit series L). The purpose of this is to move the resonance down to frequencies that are not easily radiated.

The resonance is triggered by the abrupt fall of reverse recovery current after the recovery process finishes and condutcion actually ends. Schottky diodes don't exhibit reverse recovery so there is much less energy to trigger the resonance.

The more ideal approach used in SMPS is to add a RC (or RLC) snubber across the diode tuned to the resonant frequency, this really prevents any ringing, but in mains transformer applications circuit L depends somewhat on mains supply L., so the high value capacitor solution is preferred. In this case, the type of diode is quite irrelevant, but standard diodes suffer from less conduction losses, so they should be preferred.

mlloyd1 4th February 2010 05:37 AM

for "regular" power supplies (i.e. not smps), i would think you want soft recovery rectifiers, not necessarily high speed or ultrafast rectifiers.
on semi and fairchild make some parts worth considering.

mlloyd1

jackinnj 4th February 2010 02:27 PM

Quote:

Originally Posted by Eva (Post 2072949)
but in mains transformer applications circuit L depends somewhat on mains supply L., so the high value capacitor solution is preferred. In this case, the type of diode is quite irrelevant, but standard diodes suffer from less conduction losses, so they should be preferred.

And the "L" which Eva refers to is the "leakage inductance" -- you can measure it with an impedance bridge -- and they vary all over the lot depending upon the type of transformer etc., etc.

But most of the time, if it is an issue (you can look at the waveform on a scope and determine if there is ringing) I wind up using a few nano-Farads and a 330 ohm resistor.

There was no ringing on the last couple power supplies I built using Antek transformers and diode bridge -- on a power supply for a preamp which I cobbled out of an HP465a there was very noticeable ringing which had to be damped.

jackinnj 5th February 2010 02:23 PM

Erno opines: Power Supplies

Scott594 5th February 2010 09:47 PM

Quote:

Originally Posted by mlloyd1 (Post 2073115)
for "regular" power supplies (i.e. not smps), i would think you want soft recovery rectifiers, not necessarily high speed or ultrafast rectifiers.
on semi and fairchild make some parts worth considering.

mlloyd1


I agree. IR HEXFREDS or equal. A tad pricy, but quiet.

tvrgeek 19th December 2011 12:20 AM

OK, I basically know what a snubber is. Several authors, Mr. Pass not the least, recommend a snubber across the bridge diodes. Makes sense. Looking at my DH120 supply output, I see harmonics from 120 past 40K. Any recommendations on where I should start? .1u and 330Ohms? Is that a "small" resistor, which is all I see mentioned in the reading I can find.

Second , any part numbers for a nice 25 or 35A HEXFRED bridge?


All times are GMT. The time now is 06:23 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2