Slam
 

On Sun, 5 Oct 2003 09:01:08 +0100, Chris Morriss
wrote:

In message , Jim Lesurf
writes
I'd agree that you need plenty of current available. For a 200Wpc I'd go
for well over 30Arms continuous, and bigger short-term available. However
soggy rails don't have to preclude this.

I used to be wary of stabilised rails, as this can lead to problems with
getting the stabilisers to work well. In effect, they end up becoming part
of the amp itself. My own preference was to spend the time and money on the
amp itself, working from a more basic PSU (old fashioned transformer,
bridge, and big caps with low ESR.)


The problem with the use of really big caps (22,000uF or so) is that the
current taken from the mains is in very short, high-amplitude pulses. My
new 200W mains SMPSU I've just designed at work needs a power-factor
correction stage converting the 90 to 260V rms into 400V dc before this
feeds the main islolating stage, to make the PSU look resistive to the
mains.

The law is now that any PSU over 85W needs PF correction. Does this
apply to amplifiers for domestic use?

It's easier to get a reasonable PF by using low value capacitors
(2,200uF or so) and tolerate the dc ripple by regulating down to a
stabilised output.

Or, if you have the money for it, you can use a choke-filtered supply
to maintain the rail voltage without active regulation. Musical
Fidelity use this in their better amps, as do most valve amps. It's
definitely the best electrical solution, since it gives a very quiet
power rail, but those big chokes are *seriously* expensive!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Slam
 

On Sat, 04 Oct 2003 17:56:59 +0100, Jim Lesurf
wrote:

In article , Stewart Pinkerton
wrote:

Note that I'm referring to pretty poor amps here - but then you've
pretty much confined the field to such amps, with your pre-qualification
of extremely saggy power supplies. I take it that we're not dealing with
Class G multi-rail amps here?

No. I'm really meaning the deliberate choice of designing the PSU so the
rails do fall under sustained demand, but hold up for musical transients.

That brings me back to my previous point. Since the reservoir caps are
recharged 100 times per second from the mains supply, you *can't* have
a rail which 'gradually' sags over a few seconds.

As an illustration, published 'tone burst' tests always take place in
less than one cycle of the mains, typically being three or four cycles
of 1kHz tone.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Slam
 

On Sat, 04 Oct 2003 17:56:59 +0100, Jim Lesurf
wrote:

In article , Stewart Pinkerton
wrote:

Note that I'm referring to pretty poor amps here - but then you've
pretty much confined the field to such amps, with your pre-qualification
of extremely saggy power supplies. I take it that we're not dealing with
Class G multi-rail amps here?

No. I'm really meaning the deliberate choice of designing the PSU so the
rails do fall under sustained demand, but hold up for musical transients.

That brings me back to my previous point. Since the reservoir caps are
recharged 100 times per second from the mains supply, you *can't* have
a rail which 'gradually' sags over a few seconds.

As an illustration, published 'tone burst' tests always take place in
less than one cycle of the mains, typically being three or four cycles
of 1kHz tone.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Slam
 

In message , Stewart Pinkerton
writes
On Sun, 5 Oct 2003 09:01:08 +0100, Chris Morriss
wrote:

In message , Jim Lesurf
writes
I'd agree that you need plenty of current available. For a 200Wpc I'd go
for well over 30Arms continuous, and bigger short-term available. However
soggy rails don't have to preclude this.

I used to be wary of stabilised rails, as this can lead to problems with
getting the stabilisers to work well. In effect, they end up becoming part
of the amp itself. My own preference was to spend the time and money on the
amp itself, working from a more basic PSU (old fashioned transformer,
bridge, and big caps with low ESR.)


The problem with the use of really big caps (22,000uF or so) is that the
current taken from the mains is in very short, high-amplitude pulses. My
new 200W mains SMPSU I've just designed at work needs a power-factor
correction stage converting the 90 to 260V rms into 400V dc before this
feeds the main islolating stage, to make the PSU look resistive to the
mains.

The law is now that any PSU over 85W needs PF correction. Does this
apply to amplifiers for domestic use?

It's easier to get a reasonable PF by using low value capacitors
(2,200uF or so) and tolerate the dc ripple by regulating down to a
stabilised output.

Or, if you have the money for it, you can use a choke-filtered supply
to maintain the rail voltage without active regulation. Musical
Fidelity use this in their better amps, as do most valve amps. It's
definitely the best electrical solution, since it gives a very quiet
power rail, but those big chokes are *seriously* expensive!


Funny you should mention that Stewart. I'm building a real 'Luddite'
single-ended Class A amp at this minute. (In fact I am wiring up the
Schottky rectifiers for the output stage PSU in between looking at the
emails).

And when I say Luddite, I mean it. Just to see if the 'zero overall
feedback' guys have any basis for their views, it's a single-ended,
emitter-follower output, choke loaded (125mH at 5A, not bought new of
course!) with a choke-input PSU.)

The SPICE simulation shows remarkably good performance, so it'll be
interesting to see how it measures and sounds in real life.

I'm working in Leicester now, and driving every day from Nottingham to
Leicester, so if it seems to work well, you'll be welcome to see how it
works on any speakers you have. (Being SE Class-A it will only give
about 15W into 4 Ohms though before running out of current.)
--
Chris Morriss

Slam
 

In message , Stewart Pinkerton
writes
On Sun, 5 Oct 2003 09:01:08 +0100, Chris Morriss
wrote:

In message , Jim Lesurf
writes
I'd agree that you need plenty of current available. For a 200Wpc I'd go
for well over 30Arms continuous, and bigger short-term available. However
soggy rails don't have to preclude this.

I used to be wary of stabilised rails, as this can lead to problems with
getting the stabilisers to work well. In effect, they end up becoming part
of the amp itself. My own preference was to spend the time and money on the
amp itself, working from a more basic PSU (old fashioned transformer,
bridge, and big caps with low ESR.)


The problem with the use of really big caps (22,000uF or so) is that the
current taken from the mains is in very short, high-amplitude pulses. My
new 200W mains SMPSU I've just designed at work needs a power-factor
correction stage converting the 90 to 260V rms into 400V dc before this
feeds the main islolating stage, to make the PSU look resistive to the
mains.

The law is now that any PSU over 85W needs PF correction. Does this
apply to amplifiers for domestic use?

It's easier to get a reasonable PF by using low value capacitors
(2,200uF or so) and tolerate the dc ripple by regulating down to a
stabilised output.

Or, if you have the money for it, you can use a choke-filtered supply
to maintain the rail voltage without active regulation. Musical
Fidelity use this in their better amps, as do most valve amps. It's
definitely the best electrical solution, since it gives a very quiet
power rail, but those big chokes are *seriously* expensive!


Funny you should mention that Stewart. I'm building a real 'Luddite'
single-ended Class A amp at this minute. (In fact I am wiring up the
Schottky rectifiers for the output stage PSU in between looking at the
emails).

And when I say Luddite, I mean it. Just to see if the 'zero overall
feedback' guys have any basis for their views, it's a single-ended,
emitter-follower output, choke loaded (125mH at 5A, not bought new of
course!) with a choke-input PSU.)

The SPICE simulation shows remarkably good performance, so it'll be
interesting to see how it measures and sounds in real life.

I'm working in Leicester now, and driving every day from Nottingham to
Leicester, so if it seems to work well, you'll be welcome to see how it
works on any speakers you have. (Being SE Class-A it will only give
about 15W into 4 Ohms though before running out of current.)
--
Chris Morriss

Slam
 

The SPICE simulation shows remarkably good performance, so it'll be
interesting to see how it measures and sounds in real life.


Keep us informed! Sounds interesting.

=== Andy Evans ===
Visit our Website:- http://www.artsandmedia.com
Audio, music and health pages and interesting links.

Slam
 

The SPICE simulation shows remarkably good performance, so it'll be
interesting to see how it measures and sounds in real life.


Keep us informed! Sounds interesting.

=== Andy Evans ===
Visit our Website:- http://www.artsandmedia.com
Audio, music and health pages and interesting links.

Slam
 

In article , Chris Morriss
wrote:
In message , Jim Lesurf
writes
I'd agree that you need plenty of current available. For a 200Wpc I'd
go for well over 30Arms continuous, and bigger short-term available.
However soggy rails don't have to preclude this.

I used to be wary of stabilised rails, as this can lead to problems
with getting the stabilisers to work well. In effect, they end up
becoming part of the amp itself. My own preference was to spend the
time and money on the amp itself, working from a more basic PSU (old
fashioned transformer, bridge, and big caps with low ESR.)


The problem with the use of really big caps (22,000uF or so) is that the
current taken from the mains is in very short, high-amplitude pulses.

The transformer behaviour also comes into this. One of the factors which I
exploited when designing was the saturation and series impedance properties
of the transformer. This can mean that - on heavy load - the transformer
changes from acting like an a.c. voltage source to being more like a pulsed
current source. This has the effect of spreading out the length of the
periods when current is being drawn from the mains, and reducing the peak
current demand. (You can also play similar games with the bridge impedance
and cap ESR to some extent, but the transformer seems to be the main item
that has this effect in my experience.)

My new 200W mains SMPSU I've just designed at work needs a power-factor
correction stage converting the 90 to 260V rms into 400V dc before this
feeds the main islolating stage, to make the PSU look resistive to the
mains.

The law is now that any PSU over 85W needs PF correction. Does this
apply to amplifiers for domestic use?

No idea. No-one required me worry about this when I was designing. 8-]

ahem I just noticed that in some cases, when the amp was turned on, the
room lights flickered and some TV's would roll a frame. ;-

It's easier to get a reasonable PF by using low value capacitors
(2,200uF or so) and tolerate the dc ripple by regulating down to a
stabilised output.

Yes. Although you can get a similar effect by using the approach I outline
above. This has the advantage of avoiding the need for a set of stabilisers
and having to ensure these don't interfere with performance in other ways.

Also, 2,200 uF is perhaps a tad on the small side if you want to support
sustained mean currents of the order of over 30 Amps per channel (i.e. over
60 Amps total for stereo), particularly using a non-SMPS! The resulting
voltage ripple may be too high.

:-)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Slam
 

In article , Chris Morriss
wrote:
In message , Jim Lesurf
writes
I'd agree that you need plenty of current available. For a 200Wpc I'd
go for well over 30Arms continuous, and bigger short-term available.
However soggy rails don't have to preclude this.

I used to be wary of stabilised rails, as this can lead to problems
with getting the stabilisers to work well. In effect, they end up
becoming part of the amp itself. My own preference was to spend the
time and money on the amp itself, working from a more basic PSU (old
fashioned transformer, bridge, and big caps with low ESR.)


The problem with the use of really big caps (22,000uF or so) is that the
current taken from the mains is in very short, high-amplitude pulses.

The transformer behaviour also comes into this. One of the factors which I
exploited when designing was the saturation and series impedance properties
of the transformer. This can mean that - on heavy load - the transformer
changes from acting like an a.c. voltage source to being more like a pulsed
current source. This has the effect of spreading out the length of the
periods when current is being drawn from the mains, and reducing the peak
current demand. (You can also play similar games with the bridge impedance
and cap ESR to some extent, but the transformer seems to be the main item
that has this effect in my experience.)

My new 200W mains SMPSU I've just designed at work needs a power-factor
correction stage converting the 90 to 260V rms into 400V dc before this
feeds the main islolating stage, to make the PSU look resistive to the
mains.

The law is now that any PSU over 85W needs PF correction. Does this
apply to amplifiers for domestic use?

No idea. No-one required me worry about this when I was designing. 8-]

ahem I just noticed that in some cases, when the amp was turned on, the
room lights flickered and some TV's would roll a frame. ;-

It's easier to get a reasonable PF by using low value capacitors
(2,200uF or so) and tolerate the dc ripple by regulating down to a
stabilised output.

Yes. Although you can get a similar effect by using the approach I outline
above. This has the advantage of avoiding the need for a set of stabilisers
and having to ensure these don't interfere with performance in other ways.

Also, 2,200 uF is perhaps a tad on the small side if you want to support
sustained mean currents of the order of over 30 Amps per channel (i.e. over
60 Amps total for stereo), particularly using a non-SMPS! The resulting
voltage ripple may be too high.

:-)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Slam
 

In article , Stewart Pinkerton
wrote:
On Sat, 04 Oct 2003 17:56:59 +0100, Jim Lesurf
wrote:

In article , Stewart Pinkerton
wrote:

Note that I'm referring to pretty poor amps here - but then you've
pretty much confined the field to such amps, with your
pre-qualification of extremely saggy power supplies. I take it that
we're not dealing with Class G multi-rail amps here?

No. I'm really meaning the deliberate choice of designing the PSU so
the rails do fall under sustained demand, but hold up for musical
transients.

That brings me back to my previous point. Since the reservoir caps are
recharged 100 times per second from the mains supply, you *can't* have a
rail which 'gradually' sags over a few seconds.

A few seconds would be quite a long time. However when the transformer is
going into saturation and/or resistive limiting, then the fall can extend
over a number of half-cycles of mains. The reason being that each
half-cycle does not now always recharge the caps up to the same peak
voltage level you'd get on minimal load. Hence you can arrange for a
response time for this which is somewhat longer than a half-cycle of mains.

By using fairly 'large' caps you can control the rate of fall in between
rechargings, and then use the transformer behaviour to control how any fall
under heavy load proceeds from one half-cycle to the next.

As an illustration, published 'tone burst' tests always take place in
less than one cycle of the mains, typically being three or four cycles
of 1kHz tone.

Afraid I can't now recall all the details of the tests I used to do as it
was before the flood. ;- However I think I tended to use 100ms and 500ms
sinewave bursts for mid-term bursts[1], and 'short' burst (i.e. about a
cycle of 20kHz) to test for more 'transient' effects.

Slainte,

Jim

[1] No idea now what frequency, but I suspect it was a few hundred Hz as
opposed to 1kHz as I tend to feel this might be more representative of
music. However i'm now guessing beyond my ability to recall... :-/

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html