how much C is too much C?

[DavesNotHere]

it just means you have a better current capacity, but it not really needed. In reality it doesn't need to be that stiff. Some designers do this on purpose because they can't figure out how to clean the heater power properly.

[Pano]

I've been wondering about this for years - "How much is enough?"

Does it come down to Joules per Watt? I.E., energy stored in the caps vs the rated output of amp. How many Watt/Seconds are needed per output Watt? Do different ratios sound different? Obviously tube amps may get away with smaller cap values simply because the voltages are higher and thus more energy is stored. What do you think?

[DavesNotHere]

if you really want to get into the math of it, its in 3 dimensional terms, not a strait linear formula.

output power impedence vs current demand over operational voltage.

Recovery time of watt per joules in seconds under full load.

and

recovery time of watt per joules per watt at idle (no signal)

[DavesNotHere]

now that is just the dc side, the ac side has another set of three dimensional terms.

now, when you go too far with capacitance, the circuit will behave more as if you had a battery B+ supply.

[DF96]

Its partly a matter of PSRR, and partly a matter of source impedance - a wimpy transformer can't recharge the caps quickly enough after a loud noise. Beyond a certain point bigger caps don't help, they just create start-up problems and safety problems.

[DavesNotHere]

inrush current can be a concern. in both SS and tube rectification (the latter will arc). but can be overcome by a current limiting resistor

[Pano]

Originally Posted by DavesNotHere

now, when you go too far with capacitance, the circuit will behave more as if you had a battery B+ supply.

That's what I was thinking, tho it might be lower impedance than the battery supply, at least at some frequencies.

I'm still interested in "how much is enough?" and what would need to be taken into account; Full power, output impedance, idle current, etc.

[ChrisA]

Originally Posted by Pano

That's what I was thinking, tho it might be lower impedance than the battery supply, at least at some frequencies.

I'm still interested in "how much is enough?" and what would need to be taken into account; Full power, output impedance, idle current, etc.

How much is enough and how much is to much are different questions and you need to answer each one.

Enough: The ripple voltage on B+ is reduced such that it is no longer a noise source. I think you can specify a desired noise floor. How low should the power supply impedance be? Just so that it is "low" compared to the output transformer. I mead, if the B+ is connoted in series to the center tap of a transformer with 100R of DC resistance.

To Much: At some point serge current at turn on becomes an issue also the total energy to charge the cap and what does it have to go through. Parts like a rectifier tube or choke or power resister may cause you to limit the capacitance unless you design in some kind of current limiter.

[Pano]

True enough, for C connected directly to the rectifier. But what about C after a choke or resistor? It could be pretty large and not have a drastic turn on pulse.

For example: A SET monoblock with a B+ of 425V and a 10uF cap after the choke would have about 900 Joules on tap for the output tube. Good, bad, or not important? How much C would you need for a solid state amp running at +/-50V to have the same energy on tap?