Ever thought about ... ?

I can tell you why that capacitor is there, but I share your distrust of lightweight designs, and much prefer it when it is moved to between the windings.

(some  'good' manufactures may use two caps in series to get double fault to danger, which is sort of cheating..)

You cannot do  a widely split bobbin at tens or hundreds of KHz in the way you would at 50Hz, as the leakage of magnetic  flux is not insignificant (the ur of ferrite is some hundreds to thousands not  the tens of thousands of the traditional nickel Iron transformer steels,or the hundred thousand or so of the copper/molybdenum steels. )

In effect the magnetic field induced by the primary breaks out of the core a bit, and some sneaks back without enclosing the secondary  - so in the circuit equivalent  you have a transformer with a rather indeterminate transformation that is not quite the turns ratio any more , and an inductance you cannot get rid of in series with the primary - so when you put a load on the secondary that inductance causes the volts to droop badly.


So to get back to nice transformer behaviour, the primary and secondary windings have to overlay, or on a large transformer interleave either as alternating layers, or as alternating  disk like piles of winding.

So the primary to secondary  capacitance is much higher than for the split bobbin counterpart. - so the 400V p-p square wave on the switching transistors gets to the output - very badly, as not only is the frequency that much higher, so for a given capacitance the displacement current is higher, but the higher  capacitance also increases the leakage.

However  we can return the switching frequency back to the source by forming a voltage divider between the winding capacitance and  that capacitor from output to input  leaving us with a predominantly mains input  frequency signal on the output instead of one at tens or hundreds of KHz. (tingles may be 50Hz or 100Hz depending if the return is made to the DC bus, or pre-rectifier  mains)


The far nicer solution used in stealth SMPS that must not disclose their location by radiating like mad and jamming sensitive receivers is to accept a larger transformer, and to interpose foils between primary and secondary windings, and then to connect the foils, and the cores to one side of the DC bus, but to leave the secondary floating. Such designs can be made to have very low mains leakage and to exceed the highest surge requirements

. They do however cost a bit more, and may not fit in the size folk expect a phone charger to be, and whenever I have got involved in such things the idea you value non-radiating over size and weight  seems an alien language to so-called power supply experts..


Smallest, Cheapest and high performance -  you can only have any two at most, and some times only one.


regards Mike

I can tell you why that capacitor is there, but I share your distrust of lightweight designs, and much prefer it when it is moved to between the windings.

(some  'good' manufactures may use two caps in series to get double fault to danger, which is sort of cheating..)

You cannot do  a widely split bobbin at tens or hundreds of KHz in the way you would at 50Hz, as the leakage of magnetic  flux is not insignificant (the ur of ferrite is some hundreds to thousands not  the tens of thousands of the traditional nickel Iron transformer steels,or the hundred thousand or so of the copper/molybdenum steels. )

In effect the magnetic field induced by the primary breaks out of the core a bit, and some sneaks back without enclosing the secondary  - so in the circuit equivalent  you have a transformer with a rather indeterminate transformation that is not quite the turns ratio any more , and an inductance you cannot get rid of in series with the primary - so when you put a load on the secondary that inductance causes the volts to droop badly.


So to get back to nice transformer behaviour, the primary and secondary windings have to overlay, or on a large transformer interleave either as alternating layers, or as alternating  disk like piles of winding.

So the primary to secondary  capacitance is much higher than for the split bobbin counterpart. - so the 400V p-p square wave on the switching transistors gets to the output - very badly, as not only is the frequency that much higher, so for a given capacitance the displacement current is higher, but the higher  capacitance also increases the leakage.

However  we can return the switching frequency back to the source by forming a voltage divider between the winding capacitance and  that capacitor from output to input  leaving us with a predominantly mains input  frequency signal on the output instead of one at tens or hundreds of KHz. (tingles may be 50Hz or 100Hz depending if the return is made to the DC bus, or pre-rectifier  mains)


The far nicer solution used in stealth SMPS that must not disclose their location by radiating like mad and jamming sensitive receivers is to accept a larger transformer, and to interpose foils between primary and secondary windings, and then to connect the foils, and the cores to one side of the DC bus, but to leave the secondary floating. Such designs can be made to have very low mains leakage and to exceed the highest surge requirements

. They do however cost a bit more, and may not fit in the size folk expect a phone charger to be, and whenever I have got involved in such things the idea you value non-radiating over size and weight  seems an alien language to so-called power supply experts..


Smallest, Cheapest and high performance -  you can only have any two at most, and some times only one.


regards Mike