parallel capacitors in "special situations"?

[KarenColumbo]

dear gentlemen, a power supply I am in the process of building specifies a 10n/2kV cap in series with a resistor from the raw tranny secondary to gnd (right after the fuse). my specialty dealer round the corner only has 4n7. Will it work if I snap in 2 of those in parallel? will they still withstand those 2 kV? will 1 4n7 suffice if I double up the resistance I mentioned?

[EBK]

Is the 2kV spec arbitrary, or are you making something scary?

Two 2kV caps in parallel make a 2kV cap.

[KarenColumbo]

Looks pretty scary to me - and sorry for the mention of "ground" in the post before. Wrote this wiothout looking at the schematic:

[Steben]

No one really notes a 50V or 100V cap in a 9V pedal.
2000 V is "only" about 4 times the working conditions.

[KarenColumbo]

No one really notes a 50V or 100V cap in a 9V pedal.
2000 V is "only" about 4 times the working conditions.

Yeah, it's not strictly and directly stompbox-related. More of a stompbox presentation platform I'm building.

[merlinb]

4n7 will work well enough on its own, it's not a critical part. Leave the resistance the same, or tune it if you have the tools.

[KarenColumbo]

Thanks! I had a bit of a read about snubbering ... don't get it, as usual. But I thought there was some intricate Ohm's Law/particle physics going on with the maximum voltage when paralleling capacitors, like halving or doubling this particular spec.

[bluebunny]

The voltage across one capacitor is the same whether or not it's on its own, or if it's in parallel with others.

[Rob Strand]

Thanks! I had a bit of a read about snubbering ... don't get it, as usual. But I thought there was some intricate Ohm's Law/particle physics going on with the maximum voltage when paralleling capacitors, like halving or doubling this particular spec.

You use subbing to limit the voltage.  Usually when you have an inductor (in this case the transformer winding) and something turning off (in this case the diode).   In your circuit each arm of the transformer supplies current on alternative cycles when one winding is working other is not so there is an on-off mechanism going on.  When the diodes turns off you get little glitches and the snubber is trying to soften them and limit the high voltages that are produced.  The designer has either seen problems in the past or wants to increase reliability.

When you interrupt the current through an inductor the current wants to keep flowing; that's the way inductors work.  The way a circuit does that is the inductor voltage keeps increasing (on it's own) until something breaksdown and allows the current to flow.   If there's nowhere for the current it often ends up going through parasitic capacitances like the winding capacitance, and since they are small the voltage goes up to crazy high values.

A snubber is a circuit which provides a path for that current to go.   The size of the capacitor is chosen the limit the voltage.  A larger capacitance will keep the voltage low.   The resistor is there to dissipate the energy.  If the resistance is too high the voltage will rise-up too much.    (If you put a snubber across switch contacts + the resistor limits the current through the switch when the contacts are closed, otherwise the switch will short out the cap and the high current can damage the contacts.)

There's a similar circuit called a clamp which has an added diode as part of the clamp circuit.
See D2, C2, R1, on this switchmode,
http://www.industrial-electronics.com/images/switchmode_2-3-1.jpg

[antonis]

When the diodes turns off you get little glitches and the snubber is trying to soften them and limit the high voltages that are produced.

Same with "brute" solution of cap across each rectifier diode..

[Rob Strand]

Same with "brute" solution of cap across each rectifier diode..

Pretty much.   Across the diode is probably best solution for switch-modes with inductors on the output side.

Interesting how a lot of commercial devices (eg. clock radios, cassette players) had caps across the diodes.

[bool]

Dio+cap is usually much cheaper than a fast-recto or a schottky.

[Rob Strand]

Dio+cap is usually much cheaper than a fast-recto or a schottky.

Back in the days before switch-modes fast-rectifiers and Schottkys were not common at all.   There's no way they would put them in cheap consumer products.  For the same reason of cost I was intrigued why they even put the caps in.  I've noticed on noisy transformers the caps help reduce the high-frequency (acoustic) fizz from the transformer.

[bool]

Designing gear that works smoothly "as expected"; with common parts only is an art in itself. If "throw-away-after use" model is expected; there is no point to use specialty parts from the start. And if "prolonged use" model is expected; there is strong point to use common parts for better serviceability. Consumer/prosumer, SMD or THT; it dosent matter.

Specialty parts; use these in bowteek stuff!

BTW caps-across rectos; I've seen these as early as in late 70's, early 80's "made in Japan" consumer stuff for Europe/UK export.

[Rob Strand]

BTW caps-across rectos; I've seen these as early as in late 70's, early 80's "made in Japan" consumer stuff for Europe/UK export.

Same here. 

[amptramp]

There's a similar circuit called a clamp which has an added diode as part of the clamp circuit.
See D2, C2, R1, on this switchmode,
http://www.industrial-electronics.com/images/switchmode_2-3-1.jpg

The topology here is a forward converter which needs the D2-C2-R1 network to carry the demagnetization current.

With a flyback converter, the transistor conducts and stores the energy in the inductor or transformer and when current stops, the inductor disgorges all its stored energy through the diode.  The core magnetization is reset during every cycle.  This works well up to about 50 watts but the transformer gets too large.

In a forward converter, the output diode conducts while the transistor is on, so the transformer doesn't have to store the energy - it is delivered as soon as it arrives, so this is better for higher power.  But there is a magnetization that has to be eliminated or the core will gradually get magnetized until you reach the end of B-H curve.  The reset winding plus D2-C2-R1 ensures the transformer recovers from magnetization.  The reset winding can be thinner than the other windings because it carries only enough current to reset the core.

This is different from what the OP posted about.  In this case, as soon as the voltage gets low enough for the diodes to stop conducting, the power transformer is unloaded and rings at some frequency, often as high as 100 KHz.  Adding a snubber to the highest voltage winding is one of the tricks to reduce noise in FM tuners and there is a whole culture of us who modify tuners for best noise performance as well as mods to improve fidelity and left-right separation.  Snubbers are not needed with tube rectifiers because the diode sections never shut off - they just increase in resistance near zero voltage.