Caps vs. regulated PSU

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

This topic has been discussed here before.
Anyway I would like to post it again.
I run my (revisited) Zen with about 30000uF
capacitance in each channel.
Recently I replaced some of the caps with
a (semi)regulated power supply from
the Power Follower99 by A.Ciuffoli
(with 10000uF in front of the circuit)

I should note that with the caps there is some hum
(audible at about 20-25cm (maybe 30) from the speakers).
With the virtual battery (the name comes from A.C.
page) there is no hum whatsoever.
The sound is pleasant, but misses body and air; there is less bass as well. I don't understand this.
As Zens draw constant current for the PSU, it shouldn't
matter how fast the PSU is. Maybe the (semi)regulation circuit reduces the current (substantially).
Can anybody write a few words?


Thanks a million for your personal interest and encouragement of DIY. Words cannot describe the appreciation of DIYers, when they know someone of your calibre is right behind them. Thanks again sir, on behalf of the DIY community.


The virtual battery supply adapted by you is a good idea, reducing ripple but with the penalty of sonic degradation as you have mentioned. Now that you already have 10,000uF after the active filter, why not use a choke and then substantial capacitance; perhaps, the sound may be better than the original.

An alternative is to replace the MOSFET based virtual battery with a transistor based capacitance multiplier, like the one described by Rod Elliot ( I myself haven't tried that exact circuit but have used something similar with a zener tied to the base of the lower transistor, (basically a series pass regulator with a darlington arrangment which doubles as a capacitance multiplier), which gave excellent results.

If the aim was to cut down on costs and therefore, reduce the physical capacitance, the latter method would yield substantial results.

Please keep the rest of us posted on further developments.
Hi all, Greg, your hum findings puzzle me - I have completed a Zen Revisited using twin paralleled gain stages (for 4ohm load - 90dB/W) - **no hum or noise audible at all even at the spkr cone**. My PSU is remote - Pi filter a la NP - 20,000uF on i/p, 2mH, then 10,000uF at PSU and another 10,000uF local to each gain stage. (all repeated for other channel) This runs at 50v, 4amp per gain stage from big toroid. Is there another reason for your hum - PSU too close to amp or sg like that? BTW I found PSU Designer at really helpful.

Switch on surge hasn't been mentioned - I experience a very gentle 'shove' only (0.5Hz?) - was expecting real problems. Switch off no problem either. I put this down to the thermistors - 48ohm (cold) of "Surge Guard" on the toroid mains i/p and the amount of cap involved. Certainly beats delayed relays and such for simplicity!

Keep the comments flowing...


In the Andrea's circuit the cap after the virtual battery seems pretty small. Your intentions to increase it sound good. Please report the results here.



PS - I think that Mr. Pass hadn't understood the topology of your PSU. You've taken away the 30000 uF, right?
Hi Greg,

(I apologize for mispelling and others : my native language is french...)

You wrote "As Zens draw constant current for the PSU"

I'm not sure that you are right. The *mean* current is constant, but the *instantaneous* is not. For this reason, you must connect a big capacitor at the output of the regulated PSU. Otherwise, the regulator becomes an active element of the amplification process, causing distortions and/or restricted bandwidth. The alternative is an exceptionally fast and stable regulator, which is not a simple thing to do...

Best regards, PL.
Hi all,

thanks for your responce.
To make things clear my setup with caps
only (30000uF, made up with a few 4700uF
and some smaller ones)
and doesn't include any choke.
The setup with the regulator has
10000uF before the regulator and about
5000uF after.
With the regulator there is no hum, but the
sound is not as nice as with the caps (it sounds
like my Aura80SE amp)
With the caps there is hum, but the sound is fuller.

Thanks to P.Lacombe for explaining why there is
sound degradtion. I did not understand that point.

I am surprised at what peted has written. My speakers
are certainly not as sensitive as his. Maybe the psu
is to close to the amp. Anyway adding additional caps
still helps. The major improvement was when I passed
20000uF mark. Thanks also for the info on thermistors.
I could not find anywhere the values that should be used.

I will add some more caps after the regulator.
I will report (this may take some time - have to go
for holidays in a few days).

I haven't tried the capacitance multiplier yet.
It also looks interesting.

The one and only
Joined 2001
Paid Member
Nevertheless it is a constant current source.
In practice, the volt of ripple or so on the
supply results in about 2 mV at the output,
so we can imagine making a better, possibly
more complex current source. This does not
significantly influence the current draw of the
amp as seen by the power supply, which is constant.


I was intrigued by what peted has written about hum
with his 30000uF capacitance. I examined my circuit.
It turns out that I have some grounding problem with
the BOZ. Zen connected directly to the CD player produces
some hum at the cone of the speaker. Connecting the BOZ
inbetween raises the hum (regardless if the BOZ is on or not; and yes, I made sure that not powered BOZ is not running on its oversized caps). I don't understand why this disappears when I use regulated psu for the Zen.

I watch exchange of thoughts between NP and P.Lacombe.
I still don't understand the theory behind "air" and "body"
of the sound produced by an A class amp. I understand that
more current = sweeter sound, but... I thought if the amp draws some (substantial) current it shoud not matter whether this current comes from a regulator or caps.

I still haven't increased the capacitance after the regulator. I want to make all the other experiments before
I possibly damage the regulator. I also believe that if
the thing survives than this indeed should be what I have been looking for.

Mr Pass,

In fact, the problem is audible hum : a voltage as low as
2 mv can produce a hum of 30 db SPL, which is quite
inaudible at 100 hz. I presume that it is no necessity to
make better current source.

It's possible that incorrect wiring of power supply
return and/or input ground causes hum, and others troubles.
This is difficult to fix without having the possibility to
examinate carefully the equipement...

Regards, P.Lacombe.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.