Proper RC Snubbering

[mikolaj612]

Hello all.
At the beginning sorry for my bad english (not native speaker).

I know that this thread was mentioned so many times, but I'm little confused about it.
So I need a good clarification and some answers.

I'll use AC to DC power supply with transformer on board (toroidal or R-Core).
How should I properly add RC Snubber to my circuit?

I'll use no center tapped secondaries, with classical rectifier bridge (4 schottky diodes will be used).

I saw two options mentioned above :

RC Snubber in parallel to AC secondaries (sometimes with additional capacitor across AC secondaries), like here :

https://diyaudioheaven.wordpress.com/tutorials/power-supplies/snubbering/

I saw each diode "snubbered" with RC or sometimes with C, like here :

Which option is better? Or maybe both should be used?
Another question is if I use 4 schottky diodes as a rectifier should I use a ceramic/film capacitor right after them? Or maybe it's not worth it after all?
I saw some designs with soft recovery diodes without hi freq capacitor installed after diodes (because of soft recovery emitting less RF energy, am I right?)

According to SNAA057B (TI document).
They use RC Snubber circuit across reservoir (big electrolytic capacitor), if I use RCR power supply should I use it with every big capacitor or maybe with only one of them? That RC snubber do the same in AC and DC circuits?

Thnak you for answers and your time.
Have a nice day.

 

[Mark Johnson]

I have had good success focusing exclusively on the "L" part of RLC resonant circuit. I attach external components directly across the two terminals of the transformer secondary. The external components are directly in parallel with the resonant "L", namely, the secondary leakage inductance.

(Here is a link) and some oscilloscope overlays are attached below.

 

[nigelwright7557]

I havent found any problems with low voltage power supplies needing snubbers.

I have seen problems with high voltage valve power supplies.
But the I found just using a decent fast recovery diode fixed the problem.
HER158G are good for high voltage circuits.

 

[mikolaj612]

My power supply is designed for powering OPAMP's and other rather low power (low current and voltage) devices/chips like buffers, DAC IC's etc.
Right now I want to use some schottky SMA diodes.
AC power line will never be higher than 24V.
I want to use Murata PLY10 hybrid choke between AC input and rectifier.
So should I use any RC snubber like mentioned in frist post?

 

[nigelwright7557]

I have designed numerous items with low voltage power supplies with snubbers on the diodes and not had any problems.
Just make sure the op amp supply is well smoothed, regulated and decoupled and it should be OK.

 

[mikolaj612]

Could you tell me what is the main diffrence (if any) between RC snubber placed in parallel to AC line to placed in parallel to each diode / bridge?

 

[nigelwright7557]

Could you tell me what is the main diffrence (if any) between RC snubber placed in parallel to AC line to placed in parallel to each diode / bridge?

One will soak up switching spikes from the diode.
The other will soak up noise on the input from the transformer.

 

[mikolaj612]

Ok.
So technically both should be used for best performance?
But how about snubbering soft recovery diodes or even better schottky?
nigelwright7557 do you place some ceramic/film capacitor after rectifier to filter hi freq noise?

 

[nigelwright7557]

You can use both if you want.
Personally I dont use either.
I jus tmake sure the decoupling and smoothing is good.

 

[Mark Johnson]

Here's a low voltage power supply delivering a low current (4 mA). The oscilloscope is measuring the voltage directly across the transformer secondary. Secondary AC voltage, with all-too-common flat top distortion found in California, is plus or minus 13 volts. RLC resonant ringing is present, even at this low voltage and low current. Some people would make no changes (leaving it in their power supply), others would snub out the ringing. It's completely a matter of individual choice.

Does the ringing get worse when you increase the load current? Of course it does!

Photos are from post #652 of the Quasimodo thread.

Edit- there is an article in the current v.10 issue of Linear Audio magazine, "Soft Recovery Diodes Lower Transformer Ringing by 10-20X", which might be of interest to some readers. (Here is a link to one of the figures) .

_

 

[mikolaj612]

Ok guys.
Shoud I use RC or CRC snubbering if I'll use PLY10 hybrid choke to remove common mode and differential mode noise from AC line?
In PLY10 datasheet there is typical application with placed capacitors after choke, one connected between AC line and two placed between AC and ground.

Schottky diodes required RC snubbers or maybe not? I always saw capacitors across ultrafast (non soft recovery) diodes like BYV27 or 1N4007 etc.

 

[Mark Johnson]

You should do what your design intuition tells you to do. Map out the equivalent circuit for s-domain analysis and sketch a couple of plots. Calculate the energy balance (recall that Q = 2*pi* energy_stored/energy_dissipated) and decide which among your alternative design concepts makes you the happiest. Maybe circuit simulation could assist with some of these steps, assuming you believe in it. Some people, usually autodidacts, do not.

 

[zigzagflux]

RC across transformer secondary, tuned by oscilloscope, not calculations. There are a number of threads with instructions on how to do this.
Regardless of current/voltage ratings, all transformers have leakage inductance, and I have never found a design I built that did not benefit from snubbing.

 

[Mark Johnson]

Here is the rest of the drawing from the Murata datasheet (post#11 only presented a small portion). The PLY seems to sell for USD 3.26 at Mouser. It's a shame they don't show data for a cascade of N=2, 3, 4, ... PLY modules in series, which some hobby enthusiasts may consider using, in light of the mantra "If some is good then more is better"

_

 

[mikolaj612]

Guys could you tell me one thing?
If I use that PLY10 in series with my AC line.
Should I use RC snubber placed before or after coil? Or maybe it should be on both sides.
My AC schematic looks like in Murata datasheet.

 

[Mark Johnson]

If the PCBoard has got the room, is there any reason not to include two full (2C+1R snubber)s, one in each position?? Then when you start measuring the real thing including the real transformer and the real Murata PHY choke, you can carefully explore all three options (left_only, right_only, both) on the real thing. Now you are putting your trust in objective, measured data. Not the questionable opinions of random strangers in online chat forums.

You WILL need to measure the real thing, to select component values. You WILL need an oscilloscope to perform the measurements, as mentioned in post #13 above. In addition to an oscilloscope, you will also need some kind of stimulus generator, either a CheapoModo that you build yourself, or a signal generator with square wave output {search for "Quasimodo ExtraLight" to learn more}.

Your only other option is to hire somebody to make these measurements for you, and to ship them your PCB + components. This costs money and takes time.

 

[nigelwright7557]

Not the questionable opinions of random strangers in online chat forums.

If that was true then whats the point of a DIY forum ?

There are some members on here with many years of design experience behind them. I have 35 years experience.

The only time I ever needed a snubber was on a valve power supply in my early years and then later I found a better fast switching diode that didnt give off switching noise.

On low voltage stuff I have never needed a snubber.
But I have always been careful with decoupling capacitors on power supplies.

 

[jean-paul]

Thanks for bringing PLY10 to the attention. I was not aware of hybrid common mode/differential mode filters. Seems best of both worlds in a compact form.

 

[29285]

Every common choke has some stray inductance. Thus I considered this "hybrid" since it appeared as some kind of marketing gag. Anyway - avoid toroids - their common mode impedance is poor, and stray inductance is low due to tight coupling even with sectional windings. Best results were obtained with o-core ferrite chokes on a 4 chamber bobbin.
something like this:
http://www.digikey.de/product-detail/de/EV24-1.0-02-10M/817-1040-ND/1928619

 

[jean-paul]

Of course I know of stray inductance, it can help in many cases 😉 It seems a dedicated/enhanced function in the PLY10 series so it makes me curious how well they work.