Capacitor-only snubber for flyback?

Hello,

In offline, isolated flyback converters (43W) in Discontinuous mode, and using an Ultra-fast secondary diode............

...Is it wise to use a simple diode snubber comprising just a 470p capacitor?

I thought that the diode snubber should always be R in series with C.?

After all, if its just a capacitor used, then a large current spike flows through this capacitor, and through the FET, every time the FET switches ON.

Surely the simple Capacitor-only snubber is no more efficient than an RC snubber?........after all, if the R is omitted, then the energy just gets burned up elsewhere in the circuit?

...after all, the capacitor-only snubber is not a specialist lossless snubber, which is a far more complicated animal?

I've never heard of a capacitor only snubber, except on the DC output (which basically means you have a nice filter cap on your output, not really a snubber). Putting a capacitor alone across the FET wouldn't even snub the spikes, just make them lower in resonant frequency. Plus the issues with extra power dissipation. Definitely a bad idea.

The term snubber is used to designate different circuit functions, so we should better ask, what's the intended function respectively observable effect of the said "capacitor-only" snubber? In fact, you need to convince me, that it's bringing any benefit to add capacitance to a switching node (unless you're designing a resonant converter).

I am working for a LCD TV company.

In one of their 43W flybacks, they have a capacitor across the secondary diode (its a UF diode)

Flyback specs
Vin = 390V
Vout = 16V
Pout = 43W
F(sw) = 67KHz
DCM

I cannot find any details , anywhere on the internet about such a "snubber".

All the diode snubbers that i can find information about are RC snubbers.

I am wondering if its a mis-design?

I daresn't tell them its wrong unles i can absolutely beyond all doubt state why it is wrong.

They have 20 years of experience here, and i have only 8.


I have been scratching my head for a long time, but i cannot see any benefit.

...other than perhaps they found it easier to pass the EMC radiated emissions test with caps across the output diode?

But i wouldnt know, because i never get to work in the EMC section.
-and its only my boss that talks to them in EMC.

So , i am wondering if you have any info on the use of a capacitor (470p) across the flyback secondary diode?

The capacitor will surely reduce interferences generated by the diode's reverse recovery, but also slow down the rectifier operation and establish lower frequency resonances with transformer leakage inductance. It's hard to predict the overall effect and to claim, it's totally wrong. The capacitor may still serve it's purpose.

You can try to extract a transformer equivalent circuit and play around with different snubbers in a simulation.

I feel the specified capacitor might be a low quality one.Capacitor will obviously reduce the voltage stress across the secondary diode during the secondary resonance.

If it is low q factor one then it also might provide a R effect,thus making it RC snubber.Not sure really.

Is it true that the capacitor-only snubber for the flyback diode will mean more losses in the primary FET?

This is because when the FET switches on, a large spike current will flow through the snubber capacitor.......and this spike will also be referred to the primary, where it will flow through the FET as its in its linear region just as it is about to turn on............

so as such, will more losses be existent in the FET due to the fact that we are using a C snubber instead of an RC snubber for the secondary diode?

Right, even if the capacitor is across the output diode, it will be reflected across the transformer and appear as extra drain capacitance. I would have a really hard time believing that its effects are overall positive. If it's a high D cap, then maybe its okay, but that's still a very bizarre way to implement a snubber. Any idea what kind of cap or what value it is?

I found capacitors in a several 100 pF range across secondary rectifiers with some older TV switch mode power supplies. Apparently, designers have found it useful at those days. My assumption is that low EMI rather than high efficiency intentions have driven the decision.

The doubts are remaining, but you have to find your own way.

Diode junction capacitance and inter-winding capacitance in xformer are in parallel, putting an extra cap across diode in a bad idea unless, as stated by others, it's a low Q type, i.e. low Q at the ringing frequency, which is generally in the 10MHz region. You use a single, low cost part, instead of a RC.
Optimum snubber would almost never be suitable because of dissipation, though.