Hi,

I'd like to build a reasonably high quality amp and pre-amp. I'm not up
to the skill level needed to build one of Nelson Pass' from his passdiy.com
site (although I like them), and kits like Velleman don't really do it for
me. I'm looking for something in the middle - that provides more direction
and instruction than Pass DIY...

Does anyone here have any suggestions?

--
Cordially,

Sonam Dasara
dovekeeper+at+electric-ink+dot+com

In article ,
"Sonam Dasara" wrote:

Hi,

I'd like to build a reasonably high quality amp and pre-amp. I'm not up
to the skill level needed to build one of Nelson Pass' from his passdiy.com
site (although I like them), and kits like Velleman don't really do it for
me. I'm looking for something in the middle - that provides more direction
and instruction than Pass DIY...

Does anyone here have any suggestions?

Get a couple of LM-12's from National Semiconductor. Don't laugh until
you examine the specs.

Isaac

"Sonam Dasara" skrev i en meddelelse
...
Hi,

I'd like to build a reasonably high quality amp and pre-amp. I'm not
up
to the skill level needed to build one of Nelson Pass' from his
passdiy.com
site (although I like them), and kits like Velleman don't really do it for
me. I'm looking for something in the middle - that provides more direction
and instruction than Pass DIY...

Does anyone here have any suggestions?

--
Cordially,

What about: http://users.ece.gatech.edu/~mleach/lowtim/

Ole

Check out this

http://www.lcaudio.dk/com/index.php?page=7

The Zapsolute amplifier has received pretty good reviews in the hifi press.



"Sonam Dasara" wrote in message
...
Hi,

I'd like to build a reasonably high quality amp and pre-amp. I'm not
up
to the skill level needed to build one of Nelson Pass' from his
passdiy.com
site (although I like them), and kits like Velleman don't really do it for
me. I'm looking for something in the middle - that provides more direction
and instruction than Pass DIY...

Does anyone here have any suggestions?

--
Cordially,

Sonam Dasara
dovekeeper+at+electric-ink+dot+com

"Bernt Rønningsbakk" nospam wrote in message

Check out this

http://www.lcaudio.dk/com/index.php?page=7

The Zapsolute amplifier has received pretty good reviews in the hifi
press.

I have a little experience with switchmode amplifiers, and therefore
seriously question how long the module pictured on the web page would last
were it to actually be obliged to deliver 580 watts into a 4 ohm load in
practical use.

Contrary to popular belief, switchmode amps don't have 99.99999999999%
efficiency. In fact, 90-95% efficiency can be optimistic and far worse is
not unlikely. I figure that dissipating more than a few watts into the
flimsy circuitry real-estate pictured would result in some pretty
destructive temperature rises, post haste!

The vendor's own numbers suggest that about 22 watts will be dissipated,
raising critical parts to 100C.

The only way temps would be limited to 100C would be for that little
aluminum bar to be attached to something far larger, about the size and
weight of a boat anchor for a nice boat.. IOW, a complete amplifier is not
being shown or sold. The claim that "Even the bottom plate of a normal 1 mm
sheet metal enclosure will do fine. " seems highly optimistic.

The specs themselves are misleading. IME switchmode amps have alot more
problems with high frequency IM at large fractions of full power than 1 KHz
THD at a few percent of full power.

The output filters seem to be very tiny, and can reasonably expected to
either melt quite quickly or have negligible filtering action, or both. They
will also cause serious rise in output impedance within the audio range, a
very common failing of switchmode amplifiers for general use.

A report of questionable performance and technical support of one of their
other products can be found he

http://www.diyaudio.com/forums/showt...?postid=203757

Sonam Dasara wrote:
: Hi,

: I'd like to build a reasonably high quality amp and pre-amp. I'm not up
: to the skill level needed to build one of Nelson Pass' from his passdiy.com
: site (although I like them), and kits like Velleman don't really do it for
: me. I'm looking for something in the middle - that provides more direction
: and instruction than Pass DIY...

: Does anyone here have any suggestions?




Checkout www.diyaudio.com lots of good info there.




: --
: Cordially,

: Sonam Dasara
: dovekeeper+at+electric-ink+dot+com

"Arny Krueger" wrote in message
...
The Zapsolute amplifier has received pretty good reviews in the hifi
press.

I have a little experience with switchmode amplifiers, and therefore
seriously question how long the module pictured on the web page would last
were it to actually be obliged to deliver 580 watts into a 4 ohm load in
practical use.


I second the sceptisism about the switchmode amps. There are some
testemonials out there that indicate that these are not state of the art and
I have chosen to stay away from them. But I was referring to the Zapsolute
amp. It is a conventional class A amp that has been available for several
years. It was a lot cheaper some years ago and is maybe not as good a
bargain as it used to be.

As for other products from LC audio, I have mixed experiences. I have built
a preamp kit that they sold some years ago - and it was not a good sounding
device for the price. But I have also installed a digital clock and the
analog output stage that Arnolds link referred to - and I can only say that
the Denon DCD s10 still is a very good CD player with those two
modifications.

I didn't see the preamp request the first time around.

For exellen DIY preamp modules: Try DACT.com. Niels Larsen have designed the
devices that are offered here (buffer, attenuator, RIAA and lately power
supply) and they are all exellent and very affordable. You will only find
good to exellent reviews of any of these products (and the technical specs
are also outstanding). There is also a remote control kit available
somewhere on the www for input switching and potmeter control on these
modules for those who want it all. Niels Larsen is btw av very friendly
person that you can call and ask stupid/difficult questions an get a serious
reply. I redesigned the not-so-successful afore mentioned preamp from LC
Audio with Niels Larsens Buffer. It had to be implemented in the middle of
an attenuator relay network and NL was very helpful with necessary details
regarding gain and resistor values and thanks to this I now have a very
transparent 4 unbalanced/2 channel balanced preamp available.

Best regards,

Bernt

Isaac Wingfield wrote in message ...

Get a couple of LM-12's from National Semiconductor. Don't laugh until
you examine the specs.

Isaac

I got a load of LM12's as cast-offs from a canceled project, and used
them to build an amp for bass guitar. They worked great. Due to their
relatively low power supply voltage, I had to build a bridge amp with
two LM12's, but it was a piece of cake.

Nowadays, I would consider National's LM3886 chips, which can easily
give you 100 Watts in bridge mode. Spreading such a modest thermal
load over two of these chips should make heat sinking pretty easy.

My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously? Under those conditions, it seems
like the VA rating of the transformer has to be greater than the RMS
sinewave output rating, by a factor of 3 or more. It would surprise me
if commercial amplifiers (save for really high end products) were
overdesigned to this extent.

The only other dilemma was that the LM12 (and LM3886) do not have
adjustable current limiting, so my amp effectively had no current
limiting. But it was wired to a predictable load -- the speaker built
into the amplifier case. For a typical stand-alone amplifier, you need
current limiting in order to handle the various load impedances that
can be plugged into the speaker jack.

"Detector195" wrote in message
om

Nowadays, I would consider National's LM3886 chips, which can easily
give you 100 Watts in bridge mode. Spreading such a modest thermal
load over two of these chips should make heat sinking pretty easy.


My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously? Under those conditions, it seems
like the VA rating of the transformer has to be greater than the RMS
sinewave output rating, by a factor of 3 or more.

The question there would be RMS sinewave rating into what impedance.

It would surprise me
if commercial amplifiers (save for really high end products) were
overdesigned to this extent.

Usually, commercial amplifier power supplies are far lossier when working
into lower impedance loads.

The only other dilemma was that the LM12 (and LM3886) do not have
adjustable current limiting, so my amp effectively had no current
limiting.

Actually, sophisiticated SOA limiting is feature that National makes heavy
claims about on their data sheets for these parts. Please see
http://www.national.com/ds/LM/LM3886.pdf .

"The performance of the LM3886, utilizing its Self Peak Instantaneous
Temperature (°Ke) (SPiKeT) protection circuitry,
puts it in a class above discrete and hybrid amplifiers by providing an
inherently, dynamically protected Safe Operating Area (SOA). SPiKe
protection means that these parts are completely safeguarded at the output
against overvoltage, undervoltage, overloads, including shorts to the
supplies, thermal runaway, and instantaneous temperature peaks.

If you check out the Safe Area curve on page 9, you find the combinations of
voltage and current that the power supply need not exceed.

"Arny Krueger" wrote in message ...
"Detector195" wrote in message
om

Nowadays, I would consider National's LM3886 chips, which can easily
give you 100 Watts in bridge mode. Spreading such a modest thermal
load over two of these chips should make heat sinking pretty easy.


My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously? Under those conditions, it seems
like the VA rating of the transformer has to be greater than the RMS
sinewave output rating, by a factor of 3 or more.

The question there would be RMS sinewave rating into what impedance.

It would surprise me
if commercial amplifiers (save for really high end products) were
overdesigned to this extent.

Usually, commercial amplifier power supplies are far lossier when working
into lower impedance loads.

True. I have typically been assuming 8 Ohms, and reasonable
differences bewteen the transformer peak output voltage and the
amplifier peak output voltage.


The only other dilemma was that the LM12 (and LM3886) do not have
adjustable current limiting, so my amp effectively had no current
limiting.

Actually, sophisiticated SOA limiting is feature that National makes heavy
claims about on their data sheets for these parts. Please see
http://www.national.com/ds/LM/LM3886.pdf .

True again. I should have clarified -- my main concern is about the
SOA of the power supply, since it usually works out that I am using
these IC's at less than their ultimate power rating.

"The performance of the LM3886, utilizing its Self Peak Instantaneous
Temperature (°Ke) (SPiKeT) protection circuitry,
puts it in a class above discrete and hybrid amplifiers by providing an
inherently, dynamically protected Safe Operating Area (SOA). SPiKe
protection means that these parts are completely safeguarded at the output
against overvoltage, undervoltage, overloads, including shorts to the
supplies, thermal runaway, and instantaneous temperature peaks.

True again and again... the SPiKe concept is on my list of all-time
clever circuit ideas.

If you check out the Safe Area curve on page 9, you find the combinations of
voltage and current that the power supply need not exceed.

Thanks for the helpful advice!

On 22 Aug 2004 12:39:26 -0700, (Detector195)
wrote:

Isaac Wingfield wrote in message ...

Get a couple of LM-12's from National Semiconductor. Don't laugh until
you examine the specs.

Isaac

I got a load of LM12's as cast-offs from a canceled project, and used
them to build an amp for bass guitar. They worked great. Due to their
relatively low power supply voltage, I had to build a bridge amp with
two LM12's, but it was a piece of cake.

Nowadays, I would consider National's LM3886 chips, which can easily
give you 100 Watts in bridge mode. Spreading such a modest thermal
load over two of these chips should make heat sinking pretty easy.

My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously?

The rule of thumb is typically to size it to the amp's total power
output rating. Some lab amps have transformers rated to take full
sine-wave output power indefinitely, but no commercial amps do. Some
respected class D amps have a lot of specs taken with random noise
output (pink?) at an average of 1/8 the max sine-wave power, which
doesn't seem unreasonable for music.

My trusty old AT M1000B mosfet amp has a 1000VA transformer, is rated
for 500+500 into 4R+4R (at which point the transformer seems quite
happy on loud music, but would clearly be inadequate for sine-wave.
Worse still, the amp is also allegedly rated at 1200W into 2R+2R, at
which point the transformer needs to supply 60% more current again.
Most commercial amps are rated like this, so I personally seldom use
commercial amps into their lowest rated load impedance, because it's
just too easy to get a combination of signals and complex load
impedance that violate the output device SOA, cap ripple rating,
heatsink capacity, etc, and in any case, performance and long term
reliability are ALWAYS noticeably better into a higher load.

One "trick" I have found useful in this decision - for smaller amps
(up to 250W) I sometimes use a pair of cheap 100VA "bobbin-wound"
transformers oriented and connected so their external magnetic fields
cancel. The bobbin winding create lots more series inductance, so the
RMS current in the transformers and filter caps is a lot lower than
in, say, a toroid, you get DC sag a little like a tube rectifier, the
power amp and heatsink don't get taxed nearly as hard if the amp cops
a sustained overload, and with decent design the noise doesn't need to
be significantly higher than with a toroid.

Tony (remove the "_" to reply by email)

In article ,
(Detector195) wrote:

Isaac Wingfield wrote in message
...

Get a couple of LM-12's from National Semiconductor. Don't laugh until
you examine the specs.

Isaac

I got a load of LM12's as cast-offs from a canceled project, and used
them to build an amp for bass guitar. They worked great. Due to their
relatively low power supply voltage, I had to build a bridge amp with
two LM12's, but it was a piece of cake.

A bridge is a great way to get more power at lower voltages.

Nowadays, I would consider National's LM3886 chips, which can easily
give you 100 Watts in bridge mode. Spreading such a modest thermal
load over two of these chips should make heat sinking pretty easy.

Be sure to do the numbers. Almost every time I calculate heatsink size,
I'm surprised by just how large it has to be. Reducing the maximum chip
temperature is a very straightforward way to improve reliability and
longevity.

My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously?

Clearly, the continuous output power of the supply should equal (or
exceed) the continuous output from the amp *plus losses*. If the voltage
sags under load, you won't get the maximum the amp is capable of.

There is an additional difficulty, however, because when a transformer
"recharges" the filter caps in a capacitor input supply, current flows
only when the transformer voltage exceeds the cap's voltage. The result
is that the energy is replaced in the caps in "pulses" of very high
current (the cap acts almost like a short whenever the transformer
voltage is great enough to charge it). So the "recharge" power consists
of very high peak current pulses; this (because losses are proportional
to I^2) will cause more than expected heating in the transformer
windings. Watch out for the rectifier peak current ratings, too; the
turn-on transient is tough on diodes. BIG diodes are still cheap.

I don't recall exactly how to derate for this; a good place to start
might be the "Radio Amateur's Handbook".

Isaac

On Mon, 23 Aug 2004 13:40:38 +1000, Tony wrote:

On 22 Aug 2004 12:39:26 -0700, (Detector195)
wrote:

My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously?

The rule of thumb is typically to size it to the amp's total power
output rating. Some lab amps have transformers rated to take full
sine-wave output power indefinitely, but no commercial amps do.

My amplifier is rated at 50 watts per channel, and it has a 1200VA
transformer. Yes, it's a Krell!

To be fair, it's also rated at 400 watts per channel into 1 ohm, which
still gives the 50% overcapacity which is correct for a class AB
amplifier. Bottom line - it's not true that no commercial amps have
transformers rated to take full sinewave power indefinitely.

To answer the OPs question, for full continuous sine wave rating,
which you might need if you're into organ music, the trannie should be
rated at around 1.5 times the full rated output power of the amp.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Mon, 23 Aug 2004 05:58:07 +0000 (UTC), Stewart Pinkerton
wrote:

On Mon, 23 Aug 2004 13:40:38 +1000, Tony wrote:

On 22 Aug 2004 12:39:26 -0700, (Detector195)
wrote:

My only dilemma was: How to size the power transformer appropriately
for the amp output power? Is there a good rule of thumb, or do I
simply have to assume that the power supply can deliver the worst case
average output current continuously?

The rule of thumb is typically to size it to the amp's total power
output rating. Some lab amps have transformers rated to take full
sine-wave output power indefinitely, but no commercial amps do.

My amplifier is rated at 50 watts per channel, and it has a 1200VA
transformer. Yes, it's a Krell!

To be fair, it's also rated at 400 watts per channel into 1 ohm, which
still gives the 50% overcapacity which is correct for a class AB
amplifier. Bottom line - it's not true that no commercial amps have
transformers rated to take full sinewave power indefinitely.

Maybe not, but I've not seen any example where the transformer IS
rated for full continuous sine-wave power - even your Krell falls well
short on that (1200VA is a long way short of the actual RMS current
needed into the storage caps to produce 400+400W sine wave power).

To answer the OPs question, for full continuous sine wave rating,
which you might need if you're into organ music, the trannie should be
rated at around 1.5 times the full rated output power of the amp.

Yes, a 1.5 factor is reasonable for organ music; but at the point of
clipping, even organ music will be putting way less demands on the
system than would long term continuous sinewaves, which really do need
a factor of 3 (typically), mostly because a transformer as
conservative as this will also have conservative storage caps and low
ripple, which, in turn, means the RMS current is much higher than the
average current. Amplifier efficiency isn't the major factor.

Tony (remove the "_" to reply by email)

"Tony" wrote in message


Yes, a 1.5 factor is reasonable for organ music; but at the point of
clipping, even organ music will be putting way less demands on the
system than would long term continuous sinewaves, which really do need
a factor of 3 (typically), mostly because a transformer as
conservative as this will also have conservative storage caps and low
ripple, which, in turn, means the RMS current is much higher than the
average current. Amplifier efficiency isn't the major factor.

Absolutely pure tones are rare events in musical sounds. Pure tones that are
extended and have a constant amplitude just below clipping are even more
rare. So, its safe to say that multiple sine waves of roughly similar
amplitudes better model musical sounds than a pure sine wave.

On a scale that would put pure sine waves at 0 dB, just about any multitone
will have 8-10+ dB LESS energy than a pure sine wave with equal amplitude,
whether you're talking peak or RMS.

I believe this roughly agrees with the following statement, taken from
above:

"...but at the point of clipping, even organ music will be putting way less
demands on the system than would long term continuous sinewaves, which
really do need a factor of 3 (typically)..."

Just for grins, I've built up test amplifiers based on 50-100 watt RMS car
sound amps powered with 14 volt regulated power supplies, driving
inefficient 4 ohm nominal impedance speakers.. I've found that with some
music, power sharply limited to 3 amps suffices, given that there is a
good-sized (4,700-20,000 uF) "stiffening cap" following it. I've
experimented with "stiffening caps" up to about 500,000 uF, and observe that
they don't offer better sound quality or more freedom from clipping with
music than the smaller capacitors.

These days laptop cord wart type power supplies in the 16-18 volt, 3-5 amp
range come up surplus, or when the laptop it was used with becomes totally
obsolete or unrepairable. Hook them up to an appropirate car sound amp, and
you may find the basis of a good punchy small, but high quality system,
suitable for use with computers, etc.

Yes, you can tap off the 12 VDC supply in a computer, but you risk forcing
an outage or a reboot if your amp is too big and/or your tastes in music are
too loud. You can manage the initial turn-on surge by runing a 1 K resistor
in series with, and a 100 uF electrolytic to ground, attached to the remote
power control terminal on the amp.

Tony wrote in message . ..

factor is reasonable for organ music; but at the point of
clipping, even organ music will be putting way less demands on the
system than would long term continuous sinewaves, which really do need
a factor of 3 (typically), mostly because a transformer as
conservative as this will also have conservative storage caps and low
ripple, which, in turn, means the RMS current is much higher than the
average current. Amplifier efficiency isn't the major factor.

Actually, I have been assuming even worse conditions -- amplifier
driven to clipping by a square wave input would be not be the most
taxing load on the output transistors, but it would be the worst-case
load on the power supply transformer. Then the transformer has to
sustain the product of its peak output voltage times the peak output
current rating of the amplifier. Accounting for the dropout voltages
of the electronics, plus filter ripple, is where my supposed factor of
3 comes from.

Granted, this is not a musically meaningful situation unless your
musical tastes are different than mine.

Now that my curiosity is piqued, it should be easy for me to
simultaneously measure both the peak voltage output and average
current output of an amplifier when playing music. In my case "music"
consists of the output of a bass guitar or upright bass, but it should
be possible to make this measurement with any recorded music, just by
digitizing it. Or even by extracting the digital data from a CD.

On Mon, 23 Aug 2004 10:56:33 -0400, "Arny Krueger"
wrote:

"Tony" wrote in message


Yes, a 1.5 factor is reasonable for organ music; but at the point of
clipping, even organ music will be putting way less demands on the
system than would long term continuous sinewaves, which really do need
a factor of 3 (typically), mostly because a transformer as
conservative as this will also have conservative storage caps and low
ripple, which, in turn, means the RMS current is much higher than the
average current. Amplifier efficiency isn't the major factor.

Absolutely pure tones are rare events in musical sounds. Pure tones that are
extended and have a constant amplitude just below clipping are even more
rare. So, its safe to say that multiple sine waves of roughly similar
amplitudes better model musical sounds than a pure sine wave.

On a scale that would put pure sine waves at 0 dB, just about any multitone
will have 8-10+ dB LESS energy than a pure sine wave with equal amplitude,
whether you're talking peak or RMS.

I believe this roughly agrees with the following statement, taken from
above:

"...but at the point of clipping, even organ music will be putting way less
demands on the system than would long term continuous sinewaves, which
really do need a factor of 3 (typically)..."

Just for grins, I've built up test amplifiers based on 50-100 watt RMS car
sound amps powered with 14 volt regulated power supplies, driving
inefficient 4 ohm nominal impedance speakers.. I've found that with some
music, power sharply limited to 3 amps suffices, given that there is a
good-sized (4,700-20,000 uF) "stiffening cap" following it. I've
experimented with "stiffening caps" up to about 500,000 uF, and observe that
they don't offer better sound quality or more freedom from clipping with
music than the smaller capacitors.

These days laptop cord wart type power supplies in the 16-18 volt, 3-5 amp
range come up surplus, or when the laptop it was used with becomes totally
obsolete or unrepairable. Hook them up to an appropirate car sound amp, and
you may find the basis of a good punchy small, but high quality system,
suitable for use with computers, etc.

Good idea - still, presumably, with a decent stiffening cap (I don't
imagine the SMPS transient response is good enough to follow the
demands of an audio power amp with only the paltry ecaps they have
inside them).

Tony (remove the "_" to reply by email)

On 23 Aug 2004 10:30:47 -0700, (Detector195)
wrote:

Tony wrote in message . ..

factor is reasonable for organ music; but at the point of
clipping, even organ music will be putting way less demands on the
system than would long term continuous sinewaves, which really do need
a factor of 3 (typically), mostly because a transformer as
conservative as this will also have conservative storage caps and low
ripple, which, in turn, means the RMS current is much higher than the
average current. Amplifier efficiency isn't the major factor.

Actually, I have been assuming even worse conditions -- amplifier
driven to clipping by a square wave input would be not be the most
taxing load on the output transistors, but it would be the worst-case
load on the power supply transformer. Then the transformer has to
sustain the product of its peak output voltage times the peak output
current rating of the amplifier.

I know what you mean. But - and this is important - the transformer
really has to sustain a lot more than that - it's internal heating is
related only to RMS secondary current.

Accounting for the dropout voltages
of the electronics, plus filter ripple, is where my supposed factor of
3 comes from.

Granted, this is not a musically meaningful situation unless your
musical tastes are different than mine.

But it's fair enough, if it's appropriate for the application. There
was never any suggestion that continuous sinewave was worst case. A
square wave output will increase current by about 1.4, and using the
frequency at which the load is minimum impedance will also typically
increase current by a further 1.2. With all of these factors applied
simultaneously with a stiff transformer and caps, you might well need
a factor of 5 or more.

Now that my curiosity is piqued, it should be easy for me to
simultaneously measure both the peak voltage output and average
current output of an amplifier when playing music. In my case "music"
consists of the output of a bass guitar or upright bass, but it should
be possible to make this measurement with any recorded music, just by
digitizing it. Or even by extracting the digital data from a CD.

You're not really concerned with peak voltage - just peak current. But
even that's not very meaningful for the sizing of the transformer, for
which you need the RMS secondary current.

Tony (remove the "_" to reply by email)

"Tony" wrote in message

On Mon, 23 Aug 2004 10:56:33 -0400, "Arny Krueger"
wrote:

These days laptop cord wart type power supplies in the 16-18 volt,
3-5 amp range come up surplus, or when the laptop it was used with
becomes totally obsolete or unrepairable. Hook them up to an
appropirate car sound amp, and you may find the basis of a good
punchy small, but high quality system, suitable for use with
computers, etc.

Good idea - still, presumably, with a decent stiffening cap (I don't
imagine the SMPS transient response is good enough to follow the
demands of an audio power amp with only the paltry ecaps they have
inside them).

You need the stiffening cap if the amp is capable of *serious* power levels
like say 35-100 wpc. 16 volts at 3 amps is only 48 watts.

Tony wrote:

Good idea - still, presumably, with a decent stiffening cap (I don't
imagine the SMPS transient response is good enough to follow the
demands of an audio power amp with only the paltry ecaps they have
inside them).

Errrr....

The reason SMPS supplies don't need large 'stiffening - lol ' caps is because the
storage caps are recharged at around 1000 times the rate of a 'conventional'
supply.

In fact, the ripple voltage ( which is an accurate inverse measure of the
supply's ability to deliver load current - low ripple = high current ability ) is
likely to be *less* than a 50/60 Hz design.

You might be impressed by 10,000 uF caps in a 50/60 Hz psu. For equivalent
performance, a modern high frequency SMPS needs only 10 uF to equal the
performance ! Usually the caps are *WAY* bigger than this anyway.

Graham

"Pooh Bear" wrote in message

Tony wrote:

Good idea - still, presumably, with a decent stiffening cap (I don't
imagine the SMPS transient response is good enough to follow the
demands of an audio power amp with only the paltry caps they have
inside them).

Errrr....

The reason SMPS supplies don't need large 'stiffening - lol ' caps is
because the storage caps are recharged at around 1000 times the rate
of a 'conventional' supply.

This is a real world advantage of SMPS, which allows a given SMPS have far
smaller filter caps than an equivalent 60 Hz design.

However, in the specific application, that the power amp can at times draw,
several times more current than the SMPS is rated to deliver. The SMPS
mentioned have a highly effective current limiter that acts almost
instantly. Without the stiffening cap, a maximum of the rated 3-4 amps can
be delivered the amp at any one time. With the stiffening cap, the amp can
draw several times more current, but only briefly. Due to the high
peak-to-average ratio of music, the brief needs for higher currents can be
satisfied from the stiffening cap.

In fact, the ripple voltage ( which is an accurate inverse measure of
the supply's ability to deliver load current - low ripple = high
current ability ) is likely to be *less* than a 50/60 Hz design.

Agreed.

You might be impressed by 10,000 uF caps in a 50/60 Hz PSU. For
equivalent performance, a modern high frequency SMPS needs only 10 uF
to equal the performance !

That's true in general, but not in the specific application which exploits
the high peak-to-average ratio of music.

Usually the caps are *WAY* bigger than this anyway.

I think that many would find observing a 50 WPC car sound amp running from a
3-4 amp laptop power supply to be interesting. 50 WPC car sound amps usually
come with 15 amp fuses, which is some kind of indication of the amount of
current they can draw under peak load conditions. Note that a 50 WPC car
sound amp might easily rated at 200 WPC or more, but these are advertising
watts.

BTW, Car sound amps with true RMS ratings of 50 WPC or more are generally
based on built-in SMPS with yet another layer of energy storage. IME the
filter caps in these amps are too small to provide long-term energy storage.
Because of the higher operating voltages at this point in the circuit, the
energy storage caps inside the amps are the ones that are most efficient to
enlarge, if operation from power supplies with reduced continuous power
drain is desired.