Quality of filter caps in class A or AB designs.

[beppe61]

Dear Sirs,

I have a question on filter caps.
I think that the quality of them is more important in class AB design than in full class A.
In class AB, as the current draw is variable rapidly while in full class A the current draw is constant, the caps have to be very fast in charging and discharging (very low ESR).

Opinion are very welcome and appreciated.
Kind regards,

beppe

[sreten]

Hmmm......

In a class aB amplifier at low power (most of the time)
the supply is hardly stressed.

In class A the supply is permanently fully stressed.

I'd say quality is equally important in both cases,
but optimum parameters are different.

/sreten.

[beppe61]

Originally posted by sreten
Hmmm......
1) In a class aB amplifier at low power (most of the time)
the supply is hardly stressed.
In class A the supply is permanently fully stressed.
2) I'd say quality is equally important in both cases, but optimum parameters are different.
/sreten.

Thank you Sir for your kind and valuable reply.
1) Perfectly clear.
2) Very interesting. Could you please elaborate about the different requirements for caps to be used in A and/or AB design?

I have to say that discussions on caps interest me very much.
For instance I have to decide if recapping an old vintage amp (an AB class design made in 1980) I have at hand.
I am pretty sure that the caps are gone even if I do not hear any particular hum or noise.
What is your opinion on caps lifespan?

Thank you so much
Kind regards,

beppe

[sreten]

Hi,

for class A you would be more interested in continuous ratings,
whilst for class aB I think you'd be more interested in peak ratings.
e.g. ripple current.

I think its always worthwhile replacing aged electrolytics.

I'd replace old power supply electrolytics with a new electrolytic
and a smaller electrolytic (designed for switch mode supplies)
in parallel.

The SMPS capacitors are usually red and described as high temperature.

/sreten.

[beppe61]

Originally posted by sreten
Hi,
for class A you would be more interested in continuous ratings,
whilst for class aB I think you'd be more interested in peak ratings.
e.g. ripple current.
I think its always worthwhile replacing aged electrolytics.
I'd replace old power supply electrolytics with a new electrolytic
and a smaller electrolytic (designed for switch mode supplies)
in parallel.
The SMPS capacitors are usually red and described as high temperature.
/sreten.

Thank you so much Sir for the very kind and precious advice.
Kind regards,

beppe

[janneman]

Originally posted by beppe61
[snip]In class AB, as the current draw is variable rapidly while in full class A the current draw is constant, the caps have to be very fast in charging and discharging (very low ESR).

Opinion are very welcome and appreciated.
Kind regards,

beppe

In class A the current draw is not constant. The difference is that in class (A)B the current has a half-wave shape, and in class A it has a full wave shape. I don't see any difference in cap requirements.

Jan Didden

[beppe61]

Originally posted by janneman

In class A the current draw is not constant.
...
Jan Didden

Thank you very much Sir.
Honestly I believed so (wrongly) !

Kind regards,

beppe

[AndrewT]

Hi,
shame on you all except Janneman. No one else noticed the blunder re. ClassA current draw.

There are two problems.
1. the change from zero current into the classAB output stage transitioning to a half wave current and then from half wave to zero causes severe reactance from the PSU. High frequency ripples, spikes etc. That instant change is effectively an infinite series of harmonics of the music frequency.

2. the continuous draw of variable current in the ClassA output stage severely stresses the smoothing caps, asking a lot of their ripple capacity. In addition the ripple content increases massively compared to a ClassAB output stage when near quiescent conditions and of necessity quiet music and this gives rise to hum from the ClassA voltage amp stage when run off the same Vrails as the ClassA output stage. To improve the PSU ripple, the capacitor ripple capacity and reduce the volt amp hum you need a big increase in total smoothing capacitance. Some suggest three times as much as a competent ClassAB amp others even more.

I would use 2mF to 3mF per peak Amp of output current from a ClassAB stage into the lowest load. ie. 10mF to 15mF for 100W into 8ohm

for ClassA I would use 6mF to 8mF per Amp of peak output into the lowest load. i.e. 21mF to 28mF for 50W into 8ohm. But this ClassA stage could easily drive 4ohm so I would adopt 42mF to 56mF for 100W into 4ohm as my design.

Due to the softer nature of the varying demands from the ClassA stage you may manage with lower quality caps or fewer very big caps and even less decoupling off board but similar on board decoupling. The ClassAB stage would require impecable caps and both off and on board decoupling.