PSU for power amp

[akis]

When selecting capacitors for a power amp PSU, how much ripple voltage do you allow or aim for? For example doubling the capacitance ought to halve the ripple voltage.

As a practical example, using a 4700uF cap and pulling 1.35A out of a test PSU showed a ripple voltage of 2V p2p on the scope (with a funny shape). Doubling the capacitor to 9400uF produced 1.2 Volts p2p ripple voltage, almost half, assuming I measured correctly on the scope (funny shape trace).

With the small cap (4700uF) the % of ripple voltage to DC voltage (10.66 Volts) is 9%, with the big cap it is 5.6%.

So assuming the above were a real PSU to power an audio power amp, what ripple would you allow for?

Or to ask it otherwise, can I just increase the caps to lower the ripple, or am I in danger of harming the transformer, eg when I power up or ?

[nigelwright7557]

Some people tend to go a bit mad with capacitance.

The more capacitance you have the lower the impedance faced by the bridge rectifier, transformer and fuses on power up.

With some amps you even need a soft start to stop the capacitors blowing the fuses on power up.

For a 450WRMS amp I usually get away with 10,000uF per rail.
However, some enthusiasts would say this was far too small.

To me it sounds good even at full power so that is good enough for me.

[cbdb]

At some point the transformer becomes the limiting device and any more capacitance is wasted as far as RMS power is concerned, but it will still help for PEAK power. The amount of ripple that a power amp can handle before its audible depends completely on design and is called power supply rejection ratio ( PSRR ) . Which can vary enormously.

[pacificblue]

Textbooks say, ripple should remain below 10 %, but that is not a dedicated value for audio power supplies.

It depends also on the listening level. Most of the time people listen at less than 1 W output power. Ripple voltage should not be an issue at that level, even with small capacitors.

If you like it loud, add some more capacitance.

Originally posted by nigelwright7557
With some amps you even need a soft start to stop the capacitors blowing the fuses on power up.

The soft-start is there to limit the transformer inrush current. Even big amounts of capacitance will not have a significant impact on the primary side.

[nigelwright7557]

Originally posted by pacificblue

The soft-start is there to limit the transformer inrush current. Even big amounts of capacitance will not have a significant impact on the primary side. [/B]

What inrush current ?

A transformer is an inductor and so rejects changes in current.

Surely its the capacitors being a short circuit on power up that kills the fuses ?

My 900w amp kills 5amp mains fuses on power up but doesnt with the capacitors disconnected. The transformer is 1000VA.

[akis]

OK, using 4 * 4700 uF on each rail, pulling 4.3 Amps will produce a ripple of 2.3 V p2p with a VDC of 35 Volts, so the % will be 3.22% which is very low (compared to the 10%) :-)

I will post the schematic and PCB soon for comments.

[pacificblue]

Originally posted by nigelwright7557
A transformer is an inductor and so rejects changes in current.

from http://en.wikipedia.org/wiki/Inrush_current
When a transformer is first energized a transient current up to 10 to 50 times larger than the rated transformer current can flow for several cycles. This happens when the primary winding is connected around the zero-crossing of the primary voltage. For large transformers, inrush current can last for several seconds. Toroid transformers can have up to 80 times larger inrush, because the remanent magnetism is nearly as high as the saturation magnetism at the "knee" of the hysteresis loop. This is caused because the transformer will always have some residual flux density and when the transformer is reenergized the incoming flux will add to the already existing flux which will cause the transformer to move into saturation. Then only the resistance of the primary side windings and the power line are limiting the current.

Originally posted by nigelwright7557
Surely its the capacitors being a short circuit on power up that kills the fuses ?

Capacitors hold a few As of charge. The current they draw is divided by the transformer ratio. E. g. a capacitor charge current of 80 A after a 240/24 V transformer will appear as 8 A on the primary. The charging takes place so fast, that a 5 A slow-blow fuse would not react to that, if the transformer's inrush current was not present in the first place.

Originally posted by nigelwright7557
My 900w amp kills 5amp mains fuses on power up but doesnt with the capacitors disconnected. The transformer is 1000VA.

Without a soft-start it is surprising that the 5 A fuses hold even without capacitors. Must be really on the edge there.

[akis]

Why not put a resistor inline with the main winding to guarantee a minimum resistance presented to the mains ?

For example a 300 VA toroidal draws 1.25 A, if we put 10 Ohm resistor it will drop 12.5 Volts at max power, which is tiny compared to the 240 V supply. If the transformer primary goes short circuit (melts) then the resistor will blow, that's an additional advantage?

Another idea would be to place a choke in series with the primary, so that the transient spike (which has a very high frequency) is choked but ordinary 50 Hz goes through with no resistance at all?

[pacificblue]

A soft-start circuit is a resistor that is shorted with a relay contact after 3-5 mains cycles to avoid the voltage drop, once the inrush current has passed.

Chokes are not practical. Those with iron cores suffer the same effects at first connection as the transformer. Coreless inductors would have to be huge and expensive to pose little resistance and stand the high currents involved.

PTCs are also used as inrush current limiters.

[AndrewT]

Originally posted by nigelwright7557
With some amps you even need a soft start to stop the capacitors blowing the fuses on power up

NO!!!!

The transformer draws a large start up current, before it tries to supply the load.
This start up current can peak in the many tens of Amps and even hit 100Apk.
A soft start is needed to allow close rated fusing of the transformer.

Once the transformer is working, then it supplies secondary current to the load/capacitors.
This is the charging current.
It too peaks at initial charging. This peak charging current can stress both the rectifier and the capacitors. This can be reduced by adding a slow charge circuit.


These two mechanism are quite different. They require a different solution and are used for completely different reasons.

Don't confuse others, just because you are confused.