Somebody help me understand the principles behind this DC blocker?

[454Casull]

Mains DC and Transformers

So I understand that the capacitor(s) are supposed to block the DC, and that the diodes are supposed to limit the voltage across the capacitor(s).

Questions:

1) How do the diodes limit the voltage across the cap(s)?
2) What does the +/- shorted bridge rectifier offer that two diodes in anti-parallel don't, aside from higher current handling?
3) The capacitance affects the impedance, and the impedance affects what exactly? The voltage drop and therefore the power loss (heat) in the cap(s)?
4) Why does the voltage drop have to be so low across the cap?
5) The ripple current of the cap(s) need to be rated at no less than (VA/V) amps, no?

[benb]

1. Each diode drops about 0.7 volts. Two in series drops about 1.4 volts, limiting the voltage across the capacitors to 1.4 volts in either direction. Apparently, DC offsets of the power line are less than this. If it's more, at least this circuit reduces the offset by 1.4 volts.
2. Maybe two diode drops instead of one? Current handling depends on how big your diodes are. A bridge rectifier is just a convenient package of four rectifier diodes.
3. The power loss and heat in the caps is from the ESR, Effective Series Resistance of the caps. Larger capacitors may have lower ESR, but "better" caps DO have lower ESR.
4. One reason is so the electrolytic capacitors don't have too much reverse-polarity voltage across them, which would damage them. Having any reverse voltage on an electrolytic seems dubious, thought they're supposed to take a volt or so of the wrong polarity without being damaged.
5. That seems right, but I'm not sure.

[bear]

Otoh, a pair of diodes in series will pass a half wave of AC just fine, yes?
And there are two pairs of diodes in series, one pair in each direction providing a direct path for the AC, minus the Vdrop through the diodes... so I find that this works at all to remove DC to be confusing to me. Of course I am easily confused, and this may be the case here...

_-_-bear

[Ron AKA]

This would seem to suggest avoiding the use of toroidal transformers is the simplest solution.

[454Casull]

Originally Posted by benb

1. Each diode drops about 0.7 volts. Two in series drops about 1.4 volts, limiting the voltage across the capacitors to 1.4 volts in either direction. Apparently, DC offsets of the power line are less than this. If it's more, at least this circuit reduces the offset by 1.4 volts.
2. Maybe two diode drops instead of one? Current handling depends on how big your diodes are. A bridge rectifier is just a convenient package of four rectifier diodes.
3. The power loss and heat in the caps is from the ESR, Effective Series Resistance of the caps. Larger capacitors may have lower ESR, but "better" caps DO have lower ESR.
4. One reason is so the electrolytic capacitors don't have too much reverse-polarity voltage across them, which would damage them. Having any reverse voltage on an electrolytic seems dubious, thought they're supposed to take a volt or so of the wrong polarity without being damaged.
5. That seems right, but I'm not sure.

I am sorry but 1) still confuses me - the diodes limit the voltage across the caps by what mechanism? Even if the maximum voltage across the diodes was 0.7V/1.4V, what's to prevent the AC to go through the caps at max line voltage?

4) One of the solutions proposed is two caps in series, wired in reverse polarity... but even this I do not see able to prevent voltage going the wrong way.

[bear]

It is supposed to prevent DC from passing... pass the AC... that's the plan, keep DC off the core of the xfrmrs - toroids do not like DC either...

_-_-bear

[bear]

I think the idea is that if the Vdrop across the caps starts to exceed the conduction voltage of the diodes, then the diodes conduct, discharging the caps in effect. So as I see it, it is a clipper of sorts and protects the caps (although I am not sure why that is needed if you use the appropriate caps - maybe this keeps it cheap and small?).

Somewhere I seem to have seen a better implementation of the idea... and there is a thread referenced in Elliot's text here on DIY, so a search is in order?

_-_-bear

[benb]

Originally Posted by 454Casull

I am sorry but 1) still confuses me - the diodes limit the voltage across the caps by what mechanism? Even if the maximum voltage across the diodes was 0.7V/1.4V, what's to prevent the AC to go through the caps at max line voltage?

The caps have a much lower impedance at the line frequency than does the transformer, so the caps PASS the AC voltage to the transformer (very little AC voltage is dropped across the caps), while blocking the (less than 1.4 volt) DC offset voltage from the mains.

Actually reading through the text tells me he used an example DC offset of 0.264V, well below the forward "on" voltage of a standard rectifier diode. This DC voltage will appear across the capacitors in his circuit, rather than across the very low DC resistance of the transformer's primary winding.

Originally Posted by 454Casull

4) One of the solutions proposed is two caps in series, wired in reverse polarity... but even this I do not see able to prevent voltage going the wrong way.

The AC voltage or the DC voltage? Or does my previous paragraph help?

[454Casull]

Originally Posted by benb

The AC voltage or the DC voltage? Or does my previous paragraph help?

Thanks - I understand the first two paragraphs.

The AC voltage reverses every cycle, so don't both caps see voltage the "wrong" way?

[benb]

Originally Posted by 454Casull

Thanks - I understand the first two paragraphs.

The AC voltage reverses every cycle, so don't both caps see voltage the "wrong" way?

Yes, maybe - the AC voltage across the caps will be pretty small, as a SWAG maybe only 0.1V. Knowing the capacitance, one could calculate reactance at line frequency, and from that and the current pulled by the transformer calculate the voltage across the caps.