Hipro HP-D3057F3H Review

[Wester547]

If capacitors on the output of a flyback only take up to half the output current, then do capacitors on the output of other topologies only take a fraction since the toroid handles so much?

[LongRunner]

I suppose so, basically.

[Wester547]

One more thing I'm curious about... if the voltage drop is higher (the forward voltage through the schottkys if they're less overspec'd), does that mean the ripple voltage would be ultimately worse (by a significant amount)?

[LongRunner]

Not that I'm aware of.

[Wester547]

What about the quality of the main switching MOSFETs or BJTs in say, a forward or half bridge topology? Would the ON resistance (if undesirable) of the primary switchers have a negative effect on the ripple voltage, for example?

[LongRunner]

No, that only affects the efficiency.

[Wester547]

However, if there's only one capacitor on a rail, it has to deal with the full ripple current from the mag-amp whether or not there's a ferrite coil before it.

Why is this? Just curious... does it have to do with the low reactance of the ferrite coil? Does the ferrite coil not remove any ripple voltage at all even if it's before the capacitor on an output filter, or would it just be too little to matter?

[LongRunner]

Because in that case, there's only one way for the ripple current to go - through the ferrite coil and capacitor.

The reactance of the ferrite coil is indeed much lower than that of the mag-amp. As for its effect on ripple voltage, I can't say.

[LongRunner]

Ripple in / ripple out = (Z+X)/Z

Okay, I stand corrected on that formula — combining resistance and reactance doesn't work that way (as stated in this article). Instead, they add together like this:

sqrt(<resistance>^2+<reactance>^2)

Replace <resistance> (ESR in this case) and <reactance> (of the inductor, here) with the values. In that equation, it doesn't actually matter which way around they are, as they are added together. Take that and divide by the ESR to get your filtering ratio. (The high ratio I gave in that post was not far off, though.)

What that means is that if resistance and reactance are the same, ripple voltage will only be divided by root-2 (1.414…) instead of 2. Another thing about reactance is is that, if you put a capacitance and an inductance in series, the smaller of the two reactances will subtract from the larger one, instead of the two adding together…

(The internet here is performing extra-terribly now. Almost makes me wonder if an actual dial-up connection would be faster.)