• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

The Truth About HiFi Amplifier Power Supplies

I have one of these in another room...

The main power supply is not complicated.

Acurus A-250

1663021586138.png


Transformer, Bridge Rectifier, a pair of big Capacitors for the positive and negative DC supply..

About 90 volts if I remember correctly, on the caps.

Works fine.

It continues to play for 30 seconds or so (depending on the volume/load) after turning it off.

I should drag it out and measure it someday.

1663023113026.png
 
Last edited:
Chokes... don't forget about chokes in linear power supplies!
Marantz-8B-63-version-ed.png


st-70_psimageclip-jpg.564492

Dynaco ST-70

JELSimple45-2A3schematic.jpg

JE Labs/Angela Labs Simple 2A3/45
Some of those old chokes used to swing!
 
It could be noted that the capacitors in a linear supply charge only near the peaks of the input voltage from the transformer secondary.

View attachment 230483

I suspect this negates much of the concern about "mains purity".

Only true when they are already charged and experiencing a low current draw.

A discharged (switch on) capacitor pulls current from the entire waveform minus the rectifier drops (0.6-1.2V). Hence enormous switch-on currents as the rate of change (slope) of the waveform is greatest near 0V.
 
Only true when they are already charged and experiencing a low current draw.

Yes.

A discharged (switch on) capacitor pulls current from the entire waveform minus the rectifier drops (0.6-1.2V). Hence enormous switch-on currents as the rate of change (slope) of the waveform is greatest near 0V.

Not the entire waveform.

It will pull after the diode is forward biased; .7V or so, and stop charging on the backside of the wave when the voltage of the wave from the secondary falls below the cap voltage plus the turn-on bias voltage of the diode.

The next charge cycle picks up when the transformer secondary reaches the cap voltage plus .7V to turn on the diode, and repeat.

Maybe something like this for a 4 or 5 cycle initial chargeup period, red being the cap voltage:

1663023706103.png


(Nonetheless, I bow to your near infinite knowledge of stuff I know nothing about)
 
Ray, you did read what I wrote? I discussed the forward voltage drop already.

And, you are showing a full wave rectified waveform (peaks at twice mains frequency), which means by definition you have two diode voltage drops in your Vmax. Silicon diodes are around 0.6V in practice, hence my comment of 0.6-1.2V. (depending on whether the PSU is half wave or full wave rectification). A full bridge has four diodes and at any given time the current is flowing through two (2) diodes = 2 forward V drops...

Those voltage drops are typical only for a low load. Pull a heap of current through your bridge and expect the diode drops to get a lot worse and the bridge to get rather hot- hence the way your amp bolts it to the chassis...
 
Last edited:
I'm old... I retreat.

That's funny. :)

I'm old too. Well, not as old as you, but all the hair on the barber's shop floor yesterday was grey. And I was his first customer of the day.
 
I have one of these in another room...

The main power supply is not complicated.

Acurus A-250
Nice straight forward amp design, very similar to my ole Adcom 545 II's.
I understand Class D is the future and all that, but when you look inside them they look like toys.
Where's all the missing parts? LOL
I sure will miss the beauty I see in the when you look inside
Class A/B of the past like these or your monster Krell's.


2385956807_3f315b5240_b.jpg
 
That's funny. :)

I'm old too. Well, not as old as you, but all the hair on the barber's shop floor yesterday was grey. And I was his first customer of the day.
You have hair?

Oh, yeah, the avatar says it all... :D
 
Don't forget the C-core or the double C-core- a very popular transformer in the past and well into the 1980/90s. Still seen in gear. A bunch of advantages (and disadvantages).

View attachment 227585
My first audiophile amplifier was with two this kind of transformer. It was a 50 watt amplifier that drive my ML well. Later I had a 100 watt class D power amplifier that I need to push the gain much higher than my 50 watt amplifier. That power amp was heavier than my 50 watt amplifier. My reference became double C core. Later I changed to studio monitor due to shift in interest.
 
For a more generic middle of the road amp example, this is the psu of the very modest Marantz PM5004 i own (taken when i was cleaning it). The PSU is on the right side of the cooling fins, the amp and preamp part is left of it. The picture is not the most clear i know, but it was taken long time ago, and in the evening with the led spot on it that i used to see when cleaning it.

It does not have the big caps that most have, but it has voltage regulators (not very clear on this picture) which makes the need for those big caps much less as they take away the ripple largely.
DSC_0022e.jpg
 
Chokes and resistors can be seen mostly in tube amplifiers, where PSU supplies high voltage and low current.
On class A/B solid state amps, where sudden surges of power may be needed, chokes and resistors are in the way.
Also a 100W SS amp works on +/- 50V. That is low enough to use very large capacity capacitors, negating the need for chokes and resistors for smoothing.
But in a tube amp, where one might see hundreds of volts, large capacitors become very expensive. But Current drawn is low, so a choke or a resistor can help with smoothing, using multiple smaller capacity capacitors.
 
You lot will like this then, real overkill.

Phonostage, twin 25va dual secondary traffos, into two rectifier boards, each sporting 4 diodes, 6x 8200uf 50v caps, arranged in a CLCRC with a 300mA Hammond choke and a 5ohm 10w Caddock resistor per rail. Feeding +/-27v into a bipolar shunt reg, feeding separate current mirrors for input stage and riaa stage, topped with quite a fancy discrete output stage, cascode fed sziklai pair hybrid buffer.

All that for about 120mA per rail, but it's so worth it to hear all that groove noise, ;)
 
You lot will like this then, real overkill.
Phonostage, twin 25va dual secondary traffos, into two rectifier boards, each sporting 4 diodes, 6x 8200uf 50v caps, arranged in a CLCRC with a 300mA Hammond choke and a 5ohm 10w Caddock resistor per rail. Feeding +/-27v into a bipolar shunt reg, feeding separate current mirrors for input stage and riaa stage, topped with quite a fancy discrete output stage, cascode fed sziklai pair hybrid buffer.
All that for about 120mA per rail, but it's so worth it to hear all that groove noise, ;)
If it was me, I would replace the whole shebang with rechargeable batteries and be done with it.
 
Back
Top Bottom