Power transformers versus amplifier output power..what is your option?

[destroyer X]

I have this problem to find a solution.... i have three possibilities.

Do you want to cooperate.

This is to beginners, novice into DIY..people that is learning electronics..this is not to enginners as they will just perceive non precise things and wrong numbers...the idea is to show you tips and tricks and the decisions you face in your DIY life.

I hope young folks will learn some..and some experienced, skilled folks may help me to decide the best option.

.................................................. .................................................. ....
My transformer...how it works..to beginners...novice do it yourself young friends

The audio amplifier power depends on the transformer power… we cannot have
more power output than the power produced by your transformer, so, it is a very
important component that must be selected with care, and tested into your bench

This big black one was rewounded, was made once again..let’s say refurbished.

Now it can works into 220 volts and the output voltage goes to 80 plus 80 volts
after rectification and filtering…the output AC voltage is 57 plus 57 volts…and
this…multiplied to 1.41 will give you the resultant DC voltage after rectification
and filtering.

Then I have loaded the transformer… I have connected 7 ohms power resistance
(70 watts big one!) from positive to ground and from negative to ground… this
way I was testing the whole system, not only transformer, but also rectifiers and
filter condensers…. The transformer performance was not the one I have dreamed.
but it is reasonable… and result 48 volts plus 48 volts DC when loaded…. Now
you see… the open voltage, without load..nothing connected into the transformer…
only diode rectifiers and the filter electrolitic condensers…. The voltage was 80
plus 80 volts… and dropped this way when hardly loaded… this means I can have
(because was tested!) 48 volts DC into 7 ohms resistance (6.85 Amperes)…this
means I can have 336 watts each rail…the transformer can produce 672 watts
of power and will give me 48 plus 48 volts when hardly loaded…means…
that an amplifier full power will be draining all that power… and 7 amperes
can be supplied by each one of the voltage rails…but the voltage will be 48 plus
48 volts DC.

Well…now we have to think about……………

My transformer...how it works..to beginners...novice do it yourself young friends

The audio amplifier power depends on the transformer power… we cannot have
more power output than the power produced by your transformer, so, it is a very
important component that must be selected with care, and tested into your bench

This big black one was rewounded, was made once again..let’s say refurbished.

Now it can works into 220 volts and the output voltage goes to 80 plus 80 volts
after rectification and filtering…the output AC voltage is 57 plus 57 volts…and
this…multiplied to 1.41 will give you the resultant DC voltage after rectification
and filtering.

Then I have loaded the transformer… I have connected 7 ohms power resistance
(70 watts big one!) from positive to ground and from negative to ground… this
way I was testing the whole system, not only transformer, but also rectifiers and
filter condensers…. The transformer performance was not the one I have dreamed.
but it is reasonable… and result 48 volts plus 48 volts DC when loaded…. Now
you see… the open voltage, without load..nothing connected into the transformer…
only diode rectifiers and the filter electrolitic condensers…. The voltage was 80
plus 80 volts… and dropped this way when hardly loaded… this means I can have
(because was tested!) 48 volts DC into 7 ohms resistance (6.85 Amperes)…this
means I can have 336 watts each rail…the transformer can produce 672 watts
of power and will give me 48 plus 48 volts when hardly loaded…means…
that an amplifier full power will be draining all that power… and 7 amperes
can be supplied by each one of the voltage rails…but the voltage will be 48 plus
48 volts DC.

Well…now we have to think about……………

[destroyer X]

After thinking a little bit, i have perceive that a 48 plus 48 volts amplifier
can be a nice idea to this huge transformer.

I will have something very close to 700 watts… and this means that an
amplifier “sucking” all that energy from the supply will transfer to speaker
something alike 60 to 65 percent of this power..because Class AB efficiency
is around 60 to 65 percent (sometimes 50 percent).. this way..audio power
can be, at least….420 watts RMS.

It is interesting… an amplifier with 45 plus 45 volts can produce this power
into 4 ohms loads…. More than 200 each channel… this way you can have
two amplifiers… total of more than 400 watts RMS.

But the voltage is not 45 plus 45!... it will be 45 plus 45 volts at maximum
consumption..when the amplifier will be entering the clipping.

What to do?..... to produce a supply with expensive condensers… 83 to 100
volts condensers are a hell expensive…or to use only few condensers and
to enter a voltage regulator..an electronic supply to reduce this voltage to
45 volts?

What do you think?

I would be happy to remember you that an amplifier with 80 plus 80 volts will
be able to produce enormous power…but this if the transformer keep the
voltage stable into 80 volts (rectified and filter…means to keep the 56 volts
AC plus 56 volts AC stable)..this do not happens..never happens…some losses
happens usually..and will depends iron core saturation, iron core material, type
of transformer construction, type of enameled wire used and wire diameter…
coil windings too… ratio of windings from primary and secondary and so on.

We gonna have an amplifier able to produce, or transfer to the speaker, more
than 600 watts each channel..because of voltage…. Into 4 ohms speakers..this
may be something near the limits…. More or less..i have no intention to be
precise….i want only to give you the tips and tricks.

[destroyer X]

Imaginary power...only into our imagination... this is not possible
as I could see measuring the transformer… the first bass tone will
make this voltage drop..and if we sustain something around 400
Watts RMS of power, the voltage gonna drop to 48 plus 48 volts.

So… this is the realistic sittuation…of course I can be wrong..my
mathematics not precise..if wrong, someone can fix and inform you
a more exact numbers…. As to show you tips and tricks are more
important the numeric precision into my point of view.

The amplifier using that 80 plus 80 volts will have to use 160 volts
transistors… rail plus rail is always the most safety margin..if
something burns into the amplifier, then the survivers, the non
burned transistors, will be facing rail to rail voltage..if they can
hold that without burn will be good… this way, less transistors
will blow up.

Those transistors are expensive..this is not a good idea.

Also you electrolitic condensers will be very expensive, as price
increases not only because the capacitance, but also because
insulation voltage..you gonna need high voltage electrolitc
condensers to use into your pcboard…enormous monsters there
wasting a lot of physical space.

So..if the supply can give you something around 700 watts..then
will be a better idea to produce amplifiers that can drain all that
power… and this is a 45 plus 45 volts amplifier.

The good idea is to produce a voltage regulator… a step down
electronic voltage stabilizer to face that sittuation…to enter those
80 plus 80 volts… that will not be really 80 plus 80 volts and will
be dropping, swinging down into the 48 plus 48 volts ……

So… the awfull calculation we use to make… entering 80 volts and
beeing 45 volts the ouput keeps 35 volts into the series pass power
regulator transistor…and 7 amperes means each electronic supply
series pass regulator may face 245 watts of power…. But this is not
the real thing…. When sucking 7 amperes, each rail, from this supply,
we know the voltage gonna drop to 48 volts..then only 3 volts will be
into the voltage regulator input terminal (colector) and the output
terminal will be 45 or sligtly less..maybe 43 volts plus 43 volts.

This way..your power into the voltage regulators will be 35 watts
to each side… a single transistor and 4 square fins, each one 4 by
4 inches will hold the job into the voltage stabilizer..to each rail of
course… 8 fins will be used…. Or 800 squared centimeters of
aluminium (calculate using only one face of your plates and fins)

[destroyer X]

Sorry if not clear.

.................................................. .................................................. ....

Is this a good idea?.... to use this hell transformer... very weak…loosing
all that voltage we gonna have to produce dual power regulators..series
pass transistors and so on…. This is costy and we gonna loose the peak
power… during miliseconds we could have enormou power…more than
1 kilowatt (two channels) of transient, small time, power surge into the
speaker….this will give you a sense of realism..but will force to produce
an amplifier with a very stabilized voltage to the input circuits..as the
power amplifier will face enormous decay, drop of voltage..this forces
you to use another transformer, or another winding, having higher voltage,
to rectify and filter that other voltage and to enter a voltage regulator, a
stabilizer to keep voltage stable to the input circuits that cannot suffer all
that voltage drop.

What gonna be your decision?

I will be happy to know?..... and how and whies about that subject?

Will you use 80 plus 80 volts supply..only rectifiers and condenser .. and
using big voltage into your amplifiers..or will you use electronic stabilized
supply?... with all the aditional circuitry and heatsinks.

Or will you trash that hell transformer..or to use it as paper weigth, as hammer,
or to use it to hold your door opened during the windy days..using it as door
locker.

Ahahahahaha!

[megajocke]

How much capacitance did you use in your first measurment? Maybe it was too little. Both rails loaded at the same time or just one?

48V seems very low, but you didn't say what kind of measurement it is. Is it valley, average or peak? DMM:s usually show average.

Who rewound the transformer? Are you sure the windings you use are the main high power ones?

[destroyer X]

Voltage measured was average.... RMS standard electronic DMM.

regards,

Carlos

[PMA]

Hi Carlos,

for my hi-biased (1A idle current) 2 x 35W/8 ohm, 2 x 60W/4 ohm, 2 x 110W/2 ohm I use 300VA transformer, if it helps. It would be better to use a 500VA unit, though.

Take care,
Pavel

[pacificblue]

I wonder, if this thread should not be in the Other Stuff > Power Supply Design section.

This is interesting, because the normal way is to decide, which output power you want from an amplifier into which load. Then you calculate the necessary output swing, add the losses through transistors and find the voltage your rails need to have at nominal output power. Add diode drop in the rectifiers and decide about the ripple you can accept. That leads you to the capacitance and the transformer secondary voltage. The transformer rating in VA should be 1x to 3x the amplifier output power. More doesn't hurt the performance, but your purse. Less is also possible, if you design the power supply for the average power and use big capacitors to deliver the peaks. An example is Linkwitz's Pluto, where one 50 VA transformer is used for three chipamps with an output power around 45 W each.

You go the other direction, take a transformer and build an amplifier around it. What makes it more difficult is, we know the secondary no-load voltage, but not the VA rating. Now we can try to deduce that from the voltage drop, when you connect a 7 Ohm load to it. 57 V x 1,41 = 80 V. If that drops by 40 % to 48 V, it looks, as if either the transformer was severely overloaded or the smoothing capacitors were much too small for that load. 48 V into 7 Ohm translates to ~330 VA, or 660 VA for both rails, but we don't know, whether the transformer is already driven into saturation and how far.

Ripple could help you in your decision about regulated or unregulated power supply. If the amplifier has a good power supply rejection ratio you will probably decide for the higher voltage of the unregulated supply. If the amplifier has a bad PSRR, the regulated version might have an edge over adding big amounts of capacitance or L-C or R-C filters to the power supply.

[destroyer X]

if this signal enters the power stage.... even modulating the power stage, then i would have some miliwatts of noise (rectified mains) there.

The earlier stages uses electronic regulation, so, ripple will not bother the input stages, nor the VAS and nor the bootstrapp.

We use to do the way you told Pacific Blue.... now i have a refurbished transformer and deciding where and how to use it...something very DIY..we find some junk and we decide how to use in a better way.... it is opposite to the traditional design direction.... where you decide the transformer you need.

This way i am deciding the amplifier style i need... the average power and so on..based into transformer bench testing.... as you have understood very well.

regards,

Carlos

[PMA]

Hi Carlos,

I also have a high constant PSU caps ripple, due to high, 1A idle current per channel. There are higher demands on PSRR of the amp then, and also for PCB design and wiring. But it can be solved to obtain no audible hum from speakers.