Half bridge SMPS capacitor capacitance question

[Salvador12]

I wanted to add a better quality capacitor to my working half bridge SMPS but I was wondering given frequency of switching stays the same and is controlled by the IGBT/mosfet driver IC.
What would change if I changed the capacitance value of the half bridge transformer capacitor ? Say i increased it from the existing 1uF to say 2.2uF.

Would that cause a larger current on each half cycle as the cap gets charged and discharged through the switches?

I attached the schematic.

 

[cupoftea]

Thanks, the book definetely says this, this subject has come up a number of times in the past in various places, and i bring up the Mack Book. Its in my lockup 90 miles away at the moment, so i cant photo it (the page) at the moment. The Mack book definetely says that a series cap is needed in half bridge to balance the rail split caps. What Mack believes is the mechanism of how the series cap does the balancing, is not stated in the book.

 

[cupoftea]

The series capacitor can develop a voltage across it, and that voltage can subtract from the voltage available to the primary, so choose a cap that can keep that voltage as low as poss, in line with your other requireements for this capacitor.
Im sure all would agree that the series cap gives help to the case of short cct output.

Mack says it aids in balancing of the rail split caps.........in order to do this, it would have to be the voltage developed across the series capcaitor that "opposes" the voltage imbalance on the rail split caps........but this seems not possible.

I hope we can help the OP to find what the series cap is totally for.....then we can help OP to size it correctly.

My bet it that its just for short cct protection. So in that case, think of the following..
1.....resonant frequency with pri shoudl be reasonable
2.....Shoudl not overly distort primary current shape
3.....should not take too much voltage off the primary coil
4.....resonant frequency with primary leakage inductance ok
5......i^2.ESR not too much

 

[Salvador12]

Thanks folks for such a nice discussion.
So it turns out it's not so easy to assess the "value" and purpose of this series capacitor.

My bet it that its just for short cct protection. So in that case, think of the following..
1.....resonant frequency with pri shoudl be reasonable
2.....Shoudl not overly distort primary current shape
3.....should not take too much voltage off the primary coil
4.....resonant frequency with primary leakage inductance ok
5......i^2.ESR not too much

ESR, you mean the ESR of the cap so that the current through it is not as high as to cause serious heating and destruction of the capacitor?
As for number 2. as long as the capacitance is enough so that the cap doesn't get fully charged halfway through each cycle I suppose the current shape should be ok.

One thing I thought about, let's say the cap value is such that it forms a series LC at resonance with the primary coil, but the capacitor is also in series with the mains rectifier rail splitting caps, so any current passing through the primary even at resonance will also have to pass through the rail caps, so that makes two series caps at any given time , isn't that the main current limiting factor through such a circuit?
So at resonance the primary circuit would behave as if the small cap is not there and instead would simply switch as if the primary coil was directly attached to the midpoint of the rail split caps correct?

 

[cupoftea]

Yes thats right...I^2.ESR heating.
The rail splitter caps will be big enough for you to consider them a constant voltage source from the point of view of resonances.

There is a resonance between primary and series cap, but it should be so far below the f(sw) that its insignificant.

With plain half bridge, resonances involving the series cap and primary L should not come into it.

Also ensure the resonance between series cap and leakage L is below f(sw)

 

[KlausST]

Hi,

Yes thats right...I^2.ESR heating.

I RMS to be exact.

***

But when you charge a capacitor ... and ESR dominates the resistance, then the power dissipaton is independent of ESR.

The energy per charging cycle is 0.5 * C * V * V
multiply it by the frequency to get the dissipated power.

(the same is ture for the discharge cylcles)

Klaus