Ross Matheson wrote:

Patrick Turner wrote:

: I said this :-
:
: " Transistor amps typically have an open loop response curve that looks like
: an arch, with a peak at 500 Hz, and rolling off
: at 6 dB/octave either side. "
:
: Unfortunately, ppl like Phil leapt on me to say how erroneous this is.
: Its because he assumed I said the -3Db point in most amps was close as possible
: to 500 Hz, but of course, a lot of what Phil assumes
: is part of a foolish rush to crucify his target, myself, forever,
: with a pile of foul language.
: His assumption is plain wrong.
: I did not say at what frequency either side of 500 Hz
: was where roll off in response occurs.

I'm sorry, but to me, "a peak at 500 Hz, and rolling off
at 6 dB/octave either side " implies 6dB down at 250 Hz, and at 1000 Hz, &etc.
*quite* specifically. If you had meant a flattened arch, why say a peak at
500Hz? If you meant a broad arch (unspecified slopes) with a final 6db
roll-off at (unspecified freq) ends, why not say so? I find it confusing ...

I think I have spelled out clearly in subsequent postings what describes
the operation of SS amps in better detail, after having got ito trouble because
people assumed too much from what I said, and
rermember, the workings of SS amps cannot be
all pinned down to what one man said in one line.
I was loathe to explain further on a tubes group.
But, after having been engaged in pleasant discussions with
ppl other than Phil Allison, who simply tried to insult his way along,
I think I have now outlined the principles of SS amp operation
with some depth, for the few folks here who are interested.

Patrick Turner.

"Patrick Turner" wrote in message
...


Bob Flint wrote:

On Sat, 29 Nov 2003 19:30:16 +1100, Patrick Turner
wrote:

I said this :-

" Transistor amps typically have an open loop response curve that looks
like
an arch, with a peak at 500 Hz, and rolling off
at 6 dB/octave either side. "

OK - so where are you getting this 500hz from? I take it you are
including the input coupling cap to keep DC out of the
amp, and the output stability cap and natural tendencies of the circuit
to fall off at rising F., but I would have
thought all this contribution would peak at maybe 50Hz and then drop
off...


** I said you would see some writhing from the slimy snake ....

I was being very general about SS amps, in refering to their
OPEN LOOP response, ie, the measured response with
no global feedback connected, but with the input filter present,


** Any " input filter " is external to the open loop response.


and the "compensation" network to control HF gain, which srats to act
somewhere above 500 Hz.
The actual poles where roll off start, ie, the -3 dB points for HF and LF
are usually either side of 500 Hz.


** An RC input filter (where used ) only rolls off frequencies below 20 Hz.

The CHARLATAN is lying - the proof of that is given by the remark
posted a few lines on from the original false claim:

" But at 30 Hz, or 10 kHz, where open loop gain is less, the thd
reduction is much less, ... "



** Squirm, wriggle, writhe, squirm, wriggle, writhe, squirm,
wriggle, writhe .......

He will very soon disappear up his own arse.





............ Phil

"Patrick Turner" wrote in message
...

Ross Matheson wrote:


I'm sorry, but to me, "a peak at 500 Hz, and rolling off
at 6 dB/octave either side " implies 6dB down at 250 Hz, and at 1000 Hz,
&etc.
*quite* specifically. If you had meant a flattened arch, why say a peak
at
500Hz? If you meant a broad arch (unspecified slopes) with a final 6db
roll-off at (unspecified freq) ends, why not say so? I find it confusing
....


** The slimy snake spat :


I think I have spelled out clearly in subsequent postings what describes
the operation of SS amps in better detail, after having got ito trouble
because
people assumed too much from what I said,


** Oh - so now it is ** other people's fault ** for taking what the
CHARLATAN wrote seriously ????


I think I have now outlined the principles of SS amp operation
with some depth, for the few folks here who are interested.


** The CHARLATAN has merely posted more and more pseudo-technical
absurdities - all blown right out his arse hole.

A geriatric bricklayer's comprehension of electronics is just a house made
of cards.

Every time * Turner the Turd * emits one of his smelly farts he blows
it all down as has to rebuild it.





........... Phil

On Sun, 30 Nov 2003 15:07:21 +1100, Patrick Turner wrote:


I was being very general about SS amps, in refering to their
OPEN LOOP response, ie, the measured response with
no global feedback connected, but with the input filter present, and the
"compensation" network to control HF gain, which srats to act somewhere above 500 Hz.
The actual poles where roll off start, ie, the -3 dB points for HF and LF
are usually either side of 500 Hz.

I haven't done any design work on SS amps since the mid 80s, so I would say I'm a bit rusty, but I remember NOT testing
them open loop! I used to have some kind of network all the time. Remember, back then 100 watts was HUGE! I used to
build lots of 5-45 watt amps.




Don't some amps limit the FB loop as well so that DC has minimum gain?

Not many.
Most SS amps have the maximum amount of FB applied AT 0.0Hz, ( DC )
There are two inputs to most SS amps, the one for the signal, from the input
filter which bandwidth limits the input,
and one for the feedback.
The feedback is derived with a reistance divider from the output, so
say 28v output is reduced to 1 v at the FB terminal.
The bottom resistor of the divider is often grounded
to 0V via a cap of around 100uF.
So in effect, at 0 Hz, the output signal is ALL fed back to the input,
and the amount of applied FB at 0 Hz is the maximum amount which can be applied.
Such an arrangement is possible in an otherwise totally direct coupled amplifier,
and the DV measured at the output has splendid stability.

Isn't that what I just said? The FB assures the DC gain is minimum? I was referring to that very cap!!



The LF open loop BW even without the FB connected is
also somewhat marred by the PS caps, and if they were say 4,700 uF,
which isn't unusual, their impedance at 5 Hz would be 6.8 ohms,
and rising at 6 dB/octave towards DC.
The other rail filter caps also have rising impedance,
and basically, the SS amp has a useless open loop power ability
as F falls low.

Wouldn't problems due to power transfer from the supply caps show up with or without FB?? Normally it's just increased
distortion, why would the power out change due to lack of FB?

The power supply caps become open circuit at 0.0 Hz.
So as F falls, the amp output signal finds itself being intermodulated
by the 100 Hz ripple signal from the PS.
With FB, the effects of this are fed back, and the distortions caused are
amplified in opposite phase to cancel the N&D as it occurs, and the
PS IMD isn't a high figure.
But there is a power bandwidth, ie,
a power level where N&D rises above 0.1%,
and that rarely means full power at 1 Hz is available from an SS amp.

In my 2 x 300 watt mosfet class AB amp, I have 100,000 uF caps
for the +/- 70 volt rails, and the amp does
manage some very low F, still at low N&D, with or without any FB.

Is this an amp you designed? I'd love to see the schemo!


The other thing to remember is that most SS amps have their output devices set up in
emitter follower, or source follower, which is like cathode follower.
This is a form of local series voltage FB. It is usually about
20 to 40 dB in magnitude, and this reduces the N&D
in output stages to below maybe 2% at quite high power, without the
additional 50 dB of global FB.
So its not uncommon for an SS amp to have a total of 80 dB of FB.

I used to build amps with complementary symmetry compound common emitter output stages, having about 50% local feedback
in the pair, driven by a diff amp. Sometimes I drove the output with an op amp. You remember the amps designed in the
70s by Lancaster and others? That's about the last SS stuff I ever read up on. I used to build 'junk box' amps out of
scrap at a 2-way radio shop I worked in.

I use SS for my stereo, tube for my guitars...

Ah, so for guitars, the amp is part of the instrument,
and tube distortions are deliberate, yet musical.

But Wagner sounds awful through a guitar amp......

But through a good hi-fi tube amp, well,
don't be surprised if the tube amp sounds better than your SS.

Actually, I build quite nice tube amps, usually for guitar (my main interest) but I also built one for music
reproduction ... and I was surprised by how nice it sounded! But I'm not sure if I would compare it to my 500 watt
stereo...

It's only problem is a lack of highs... I am losing highs like crazy in there - I blame the transformer... I selected it
for it's high inductance and good bass, but.... I installed a parallel network to give me more above about 2k but it
still isn't good enough. Of course, it is a killer transformer for guitar! Oh, I don't use any feedback in tube amps,
I think it kills the 'liveness' of the sound.


Patrick Turner.

Bob Flint wrote:

On Sun, 30 Nov 2003 15:07:21 +1100, Patrick Turner wrote:


I was being very general about SS amps, in refering to their
OPEN LOOP response, ie, the measured response with
no global feedback connected, but with the input filter present, and the
"compensation" network to control HF gain, which srats to act somewhere above 500 Hz.
The actual poles where roll off start, ie, the -3 dB points for HF and LF
are usually either side of 500 Hz.

I haven't done any design work on SS amps since the mid 80s, so I would say I'm a bit rusty, but I remember NOT testing
them open loop! I used to have some kind of network all the time. Remember, back then 100 watts was HUGE! I used to
build lots of 5-45 watt amps.

Few folks do test them open loop, because what you get is real awful,
lots of hum, noise, distortion, and low bandwidth.
But with the better built types, they can be tested open loop, and its not uncommon to need
only 2mV to drive them to full power at around 500 Hz.
One has to usually have the FB network only partially disconnected,
ie, have the AV filtered out of the FB signal, yet still allow
the DV from the output to control the DV offset at the output, lest it drift
badly during tests.




Don't some amps limit the FB loop as well so that DC has minimum gain?

Not many.
Most SS amps have the maximum amount of FB applied AT 0.0Hz, ( DC )
There are two inputs to most SS amps, the one for the signal, from the input
filter which bandwidth limits the input,
and one for the feedback.
The feedback is derived with a reistance divider from the output, so
say 28v output is reduced to 1 v at the FB terminal.
The bottom resistor of the divider is often grounded
to 0V via a cap of around 100uF.
So in effect, at 0 Hz, the output signal is ALL fed back to the input,
and the amount of applied FB at 0 Hz is the maximum amount which can be applied.
Such an arrangement is possible in an otherwise totally direct coupled amplifier,
and the DV measured at the output has splendid stability.

Isn't that what I just said? The FB assures the DC gain is minimum? I was referring to that very cap!!

Well yes, the FB is at its maximum level at DV, or 0.0Hz.
In fact at DV, many SS amps act as followers, with DV gain
very very close to 1.0000, and DV gain of 0.996 isn't unusual.
Some have their R divider taken to 0V, without a cap, and
the DV gain is similar to that at higher F.




The LF open loop BW even without the FB connected is
also somewhat marred by the PS caps, and if they were say 4,700 uF,
which isn't unusual, their impedance at 5 Hz would be 6.8 ohms,
and rising at 6 dB/octave towards DC.
The other rail filter caps also have rising impedance,
and basically, the SS amp has a useless open loop power ability
as F falls low.

Wouldn't problems due to power transfer from the supply caps show up with or without FB?? Normally it's just increased
distortion, why would the power out change due to lack of FB?

The power supply caps become open circuit at 0.0 Hz.
So as F falls, the amp output signal finds itself being intermodulated
by the 100 Hz ripple signal from the PS.
With FB, the effects of this are fed back, and the distortions caused are
amplified in opposite phase to cancel the N&D as it occurs, and the
PS IMD isn't a high figure.
But there is a power bandwidth, ie,
a power level where N&D rises above 0.1%,
and that rarely means full power at 1 Hz is available from an SS amp.

In my 2 x 300 watt mosfet class AB amp, I have 100,000 uF caps
for the +/- 70 volt rails, and the amp does
manage some very low F, still at low N&D, with or without any FB.

Is this an amp you designed? I'd love to see the schemo!

I have not posted the schemo at my website, but here is a pic of what I made in 1996,
http://www.turneraudio.com.au/webpic...et400w302h.jpg
Its about 350mm wide, 300mm deep, 200mm high.
I should have made it bigger, its a bit crammed inside.
Weight is 30 Kg.

But through a good hi-fi tube amp, well,
don't be surprised if the tube amp sounds better than your SS.

Actually, I build quite nice tube amps, usually for guitar (my main interest) but I also built one for music
reproduction ... and I was surprised by how nice it sounded! But I'm not sure if I would compare it to my 500 watt
stereo...

These amps might manage....
http://www.turneraudio.com.au/htmlwe...00monobloc.htm

But then power isn't everything if all you need is 2 watts max,
on average sensitivity speakers.
I find no difference to the sound with 50 watt amps, compared to
300 watt amps, if all I need is 2 watts.
In fact visitors often prefer 25 watt SE amps.
There is a sense of utter effortlessness about a 300 watt amp.




It's only problem is a lack of highs... I am losing highs like crazy in there - I blame the transformer... I selected it
for it's high inductance and good bass, but.... I installed a parallel network to give me more above about 2k but it
still isn't good enough. Of course, it is a killer transformer for guitar! Oh, I don't use any feedback in tube amps,
I think it kills the 'liveness' of the sound.

Most good OPTs have a bandwidth that exceeds the AF band
if driven by low impedance sources.
Most guitar amps are configured in pentode, or beam tetrode, without FB,
and as this mode is a high impedance source to the load,
then the bandwidth isn't as flat as when FB is used.
The high inductance is thus required to get good bass, and the OPT
should have low shunt capacitance and low leakage inductance.
Some pentode amps have poles at 50 Hz and 7 kHz,
if no FB is used. I have seen worse.
Many guitar amps do use some FB, often 6 to 12 dB,
just to extend the F range, but not enough to ruin the "tube sound".

Patrick Turner.

On Mon, 01 Dec 2003 13:41:00 +1100, Patrick Turner wrote:


These amps might manage....
http://www.turneraudio.com.au/htmlwe...00monobloc.htm

Now that's one big tube amp!! The power supplies are separate? How do you run the high volts to the amp? Ignition
wire?!?! lol!!

Do you make the transformers yourself? I'd like to get some info on that, also why separate windings... I guess
capacitance?


But then power isn't everything if all you need is 2 watts max,
on average sensitivity speakers.
I find no difference to the sound with 50 watt amps, compared to
300 watt amps, if all I need is 2 watts.

The speakers I use in my stereo are quite efficient, so 100 watts per channel can blow people out of the room. But I was
planning on building a much smaller system, but the type of speaker I've been looking at is only about 85db compared to
the 100db I have now. That will require an amp upgrade...

In fact visitors often prefer 25 watt SE amps.
There is a sense of utter effortlessness about a 300 watt amp.

what do you think of all these high power FET amps showing up these days?


Many guitar amps do use some FB, often 6 to 12 dB,
just to extend the F range, but not enough to ruin the "tube sound".

I had originally put a feedback control (presence) but liked the sound better when it was off, so I just removed it. I
like the open sound it gives to the guitar.