question about snubber on a flyback

[Walkura]

After a couple of days of fruitless searching on google, I would like to ask a question.

I encountered RC snubbers over the secondary windings of a flyback converter in a project i worked on.
Its pretty much standard that You will see them over the mosfet on the primary side or on the secondary side over the diodes, but this one was over the windings before the diodes.

The only reference (that slightly resembled it) i could  find was in the book of Marty Brown, the so called parallel snubber and it was placed over the primary of a flyback.

How is this type of snubber called and where can i find any more information on it.

My gut feeling told me it was there to create a lossy resonance tank circuit at the source of the problem (being the charge stored in the leakage inductance) and basing my calculations on this it worked pretty good.

But I would be very grateful if anyone of You could point me in the right way to get a better understanding of its workings and the theoretical background of it.

Regards,
Walkura.

[zab]

I think you are referring to RCD snubber  used across the main winding. Normally two ways are used in this respect. First not allow to generate the high voltage spike.second bypass the spike. so that it may not damage the semiconductor devices. the snubber under discussion is of second type.
some very useful informations are available at http://www.fairchildsemi.com/an/AN/AN-4147.pdf
as Design Guidelines for RCD Snubber of Flyback Converters

[wellnerson1]

Please refer: http://en.globtek.com/controlling-emi-in-power-supply-design.php may be useful for you.
Also refer the book: High Frequency Switching Power Supplies: Theory and Design- George Chryssis.

[Walkura]

@Zab,

ofcourse we use some method of clamping on the primary side as well,
this is a standard RC snubber which is over the secondary side of the transformer before the diodes.
By placing this we do not require any snubbers placed over the diodes.
I just can't find any data or reference to a method like this.

@wellnerson1,

I will have a look at Chryssis book,
I did not look through that book as of yet, although i did look through the 3th edition of Pressman of which he is co-editor.
So far the closest i found that resembled this was Marty Brown's parallel snubber which was placed on the primary winding.

[zab]

Snubber on secondary in parallel is still confusing. I think the snubber you are referring is not for High frequency switching supply.
In low frequency supply we do have snubber before rectifier diode,just to suppress High frequency,high voltage transient pulse created by spark etc.

[Walkura]

Well, i think You start realizing this isn't exactly any conventional snubber.
The power supply i talk about run's at 100+ kHz and i am 100% sure it is on the secondary side over the winding.
I am familiar with the standard snubbers and clamping methods, but i never seen any placed like this.
Basing my calculation on the leakage induction of the winding i have a good result with it, but this was gut feeling and I would like some theory behind it.

[zab]

you make me curious about it. can you share basic circuitry. so that we can collective analyze it.
As for as fly back topology concerns we normally 2 type of snubber, both on primary,
1. parallel to switching transistor
2. parallel to winding ,
the purpose of these to protect  switching transistor,not the diode on secondary.for diode we add rc snubber across diode as well.

Now coming to your riddle I think a Capacitor is attached parallel to secondary winding to create a LC tank circuit.This tank circuit do not let high voltage peak to appear.High volt spike is curtailed because generated voltage created by coil has to charge the attached capacitor. the value of the capacitor will be selected according to the operating frequency and value of the secondary winding.

[Walkura]

Forgive me if it is going to be a bit long story.

Its a flyback based on TNY276, its an old design which was "transferred" to a newer version of the same appliance.
I'll start describing the original (old) powersupply (this was designed by a very experienced engineer which is pensioned now).
(that X2 capacitor is the only EMI-filtering for it)
It starts with a varistor & (an optional) X2 capacitor, single diode rectifier followed by a pi-filter.
This powers the transformer & TNY276, over the primairy winding is a clamp in the form of a diode and a zener.
Secondary i have 3 windings, 2 non-isolated & 1 isolated, the rest is relative standard except that there are RC snubbers over the (non-isolated) secondary winding.

Now to the new situation,
the transformer is changed, the layout is of course different and it had problems with emissions.
First i placed a 100nF X2 capacitor, this lowered the overall emissions but i had harmonics of the switching frequency very visible & above 3 MHz it slowly climbed to 10dBuV above the limits around 5 MHz.
I looked with the scope at the switcher and there was an oscillation of 5 MHz, I measured the primary leakage induction calculated the snubber and that was the end of this peak.
A new peak was visible at 3 MHZ, below the limits but to close for comfort.
I removed the "old" snubbers that where over the secondary windings, I devided my primary leakage inductance by the ratio of the windings calculated my snubbers (which are only over the non-isolated windings) and this got my emissions 10 dBuV below the limits.
The results were actually a few DBuV better then with placing snubbers over the secondary diodes itself.
The harmonics were easily solved by replacing the input diode with a slow recovery type.
So far it all seems relative straight forward.

When i asked around with my co-workers nobody ever seen a snubber placed like this, trying to find info on the web i found nothing like it.
I searched through books on switch mode supply's, application notes and again found nothing.
My assumption was that it created a "lossy" tank circuit to dampen the oscillation between the L being the leakage induction of the transformer and the C i calculated, with R having the same resistance as the impedance of the oscillation circuit (L & C)
But why i can't find any information on this method when it obviously works.
I didn't require any snubbing over secondary diodes, and i do have a (relative) clean waveform now (primary and secondary).

But how does it exactly work and did i mathematically approach this correct.
How can I calculate the losses this snubber has, I know it survives but how can i get it mathematically "waterproof"

Regards,
Walkura.

[zab]

The detail of (old) portrays a known picture. but detail of actual problem(new) is still ambiguous. please further clarify it.
No doubt the purpose of the snubber is stop high frequency ringing as well. So by adding a cap. will bring down the frequency and damp the oscillation.

please study  thse links http://www.hagtech.com/pdf/snubber.pdf
http://space.dianyuan.com/blog/u/50/1170381240.pdf  Snubber Circuits: Theory , Design and Application Philip C. Todd
It is Very informative and has a detailed description of different snubbers.
Hopefully Fig 13 of this book resemble your circuitry. its detail will answer your query.

Snubber circuit always waste some power now SiC components are being in action to get rid of this dissipation /power wastage.

[Walkura]

@Zab,

thank You for those application notes.
It actually is the first reference to a snubber like the one from my project.
But give me some time to read it through a bit thorough, from what i seen so far at illustration 13 looks like what i have here.
The "new" supply i spoke about is in electronic parts identical to the "old" one, only the layout changed & the transformer was changed.
To be honest I am not even sure if the power supply isn't "borrowed" from an even older project.

It really is a great help that i at last have something to start with.

Regards,
Walkura.