Hi Group,

This qualifies as a newbie question. Thanks for the insight in
advance.

I am an experienced loudspeaker designer with a background in
mechanics so I am familiar with designing loads, not driving loads. I
am curious if there are some good rule of thumb, generic wisdom type
of tidbits that might help me learn how to best go about matching
various stages in a total design to each other. I'm getting the hang
of determining in/out impedance characteristics, but not really
comfortable with knowing what's acceptable etc.

If anyone know a good reference on this subject this would be
appreciated.

Wessel Dirksen

"Wessel Dirksen" wrote

...am curious if there are some good rule of thumb, generic wisdom
type
of tidbits that might help me learn how to best go about matching
various stages in a total design to each other...

10 to 1

A couple of issues:

Impedance may vary with signal, so the maximum possible output and
minimum possible input impedances should be considered. For example,
a cathode follower may have a very low output impedance at low
signal levels, but if it is pushed near to cut-off its output
impedance soars. As an opposite example, a class A2 power stage may
have a high input impedance until it pushes into grid conduction,
then it plummets.

If the soaring CF meets the plummeting grid, you have a worst case
scenario.

Secondly, the rule of thumb can be modified with due consideration
to how well behaved the impedances are. For example, a badly behaved
speaker load will want rather more than the 10 to 1 ratio. If it
drops to 2 ohms minimum, for example, it could still be said that
you need less than 0.2 ohms to drive it well.

cheers, Ian

Wessel Dirksen wrote:

Hi Group,

This qualifies as a newbie question. Thanks for the insight in
advance.

I am an experienced loudspeaker designer with a background in
mechanics so I am familiar with designing loads, not driving loads. I
am curious if there are some good rule of thumb, generic wisdom type
of tidbits that might help me learn how to best go about matching
various stages in a total design to each other. I'm getting the hang
of determining in/out impedance characteristics, but not really
comfortable with knowing what's acceptable etc.

If anyone know a good reference on this subject this would be
appreciated.

Wessel Dirksen

The general rule for amplifiers is
"high impedance input,
low impedance output"

It means that the Zin for a give amp stage should be high relative to the

impedance of the preceding stage, and generally this is easy to achieve
because
tubes have extremely high input Z unless it is a tube in an output stage
where
grid current is flowing, or you have a grounded grid stage, neither of
which are
used very often in domestic audio gear.


At the output of the amp the Zout should be a small fraction of the
speaker's
minimum impedance, and having Zout = 1/10 of the nominal speaker Z
is a reasonable guide for tube amps.

This means that triodes must be used where little or nor global NFB is
to be used, or using pentode or tetrode output stages with about 20 dB of
NFB.

As you should know, Speakers are designed so that together with their
crossovers
they have a flat acoustic response, and crossovers work with the designed

attenuation slopes in the filters only when the Rout of the amp
is about 1/10 RL.
So it means an amp should have Rout between 0.5 ohms and 1 ohm.

As Rout is raised above 1 ohm, the response becomes unevern ( frequency
distortion )
and the rates of attenuation of the LCR filters become lower and
not as deigned.

Patrick Turner.

Some other Rodents have already answered better than I could do. IMHO I'd
just add one more consideration based on common sense: there should be a
balance between the size and rated power of the various elements in the
chain. For instance: the output impedance of an ECC88 is low to tube
standards, but I wouldn't use a 1.5W tube to drive a 100+W one, say a GK-71,
even if it were "theoretically" OK. This in fact is implied in Mr. Iveson
answer: as soon as the tube is pushed to its limits and some grid current
begins flowing, the "signal" stage shall supply some power too, and it shall
be designed taking into consideration this problem too.
IMO "visual" concepts are easier understood. Think of a 6C4 in front of a
300B: looks bad, isn't it? Now think of an AD1 in front of the same 300B.
Bad, in an opposite way. Finally, think of a 6SN7 in front of the old WE
coke bottle: looks "right", isn't it?

Ciao

Fabio


"Wessel Dirksen" ha scritto nel messaggio
om...
Hi Group,

This qualifies as a newbie question. Thanks for the insight in
advance.

I am an experienced loudspeaker designer with a background in
mechanics so I am familiar with designing loads, not driving loads. I
am curious if there are some good rule of thumb, generic wisdom type
of tidbits that might help me learn how to best go about matching
various stages in a total design to each other. I'm getting the hang
of determining in/out impedance characteristics, but not really
comfortable with knowing what's acceptable etc.

If anyone know a good reference on this subject this would be
appreciated.

Wessel Dirksen

Thanks Ian and Patrick. This is pretty much what I was looking for.

"Patrick Turner" wrote in message
...


Wessel Dirksen wrote:

Hi Group,

This qualifies as a newbie question. Thanks for the insight in
advance.

I am an experienced loudspeaker designer with a background in
mechanics so I am familiar with designing loads, not driving loads. I
am curious if there are some good rule of thumb, generic wisdom type
of tidbits that might help me learn how to best go about matching
various stages in a total design to each other. I'm getting the hang
of determining in/out impedance characteristics, but not really
comfortable with knowing what's acceptable etc.

If anyone know a good reference on this subject this would be
appreciated.

Wessel Dirksen

The general rule for amplifiers is
"high impedance input,
low impedance output"

It means that the Zin for a give amp stage should be high relative to the

impedance of the preceding stage, and generally this is easy to achieve
because
tubes have extremely high input Z unless it is a tube in an output stage
where
grid current is flowing, or you have a grounded grid stage, neither of
which are
used very often in domestic audio gear.


At the output of the amp the Zout should be a small fraction of the
speaker's
minimum impedance, and having Zout = 1/10 of the nominal speaker Z
is a reasonable guide for tube amps.

This means that triodes must be used where little or nor global NFB is
to be used, or using pentode or tetrode output stages with about 20 dB of
NFB.

As you should know, Speakers are designed so that together with their
crossovers
they have a flat acoustic response, and crossovers work with the designed

attenuation slopes in the filters only when the Rout of the amp
is about 1/10 RL.
So it means an amp should have Rout between 0.5 ohms and 1 ohm.

As Rout is raised above 1 ohm, the response becomes unevern ( frequency
distortion )
and the rates of attenuation of the LCR filters become lower and
not as deigned.

Patrick Turner.

forgive me if this is a stupid question...

i remember being taught that output impedance of the final stage should
match that of the load.

i feel like im missing something here... can u enlighten me?

randy

"xrongor"

forgive me if this is a stupid question...


** Long as you forgive a stupid reply.


i remember being taught that output impedance of the final stage should
match that of the load.

i feel like im missing something here... can u enlighten me?



** Women know that shoes and hats must "match" their frocks.

But they are no so ****ing literal as to put frocks on their heads or
feet.


randy


** More likely "rancid".



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

On Thu, 31 Mar 2005 19:28:45 -0700, "xrongor"
wrote:

i remember being taught that output impedance of the final stage should
match that of the load.

Maximum power into a load occurs when source and load impedances
are equal. This is true for audio purposes, but not a useful
goal.

Others have answered the OP's question better than I could,
but I'd add a skew on the OP's perspective: each stage and its
loading should be considered in the context of the whole
shootin' match.

Meaning, in valve circuits, impedance buffering stages may
often be replaced with beefier preceding stages (not always,
of course). Others have already said this implicitly, so I'd
just add, fewer stages are better stages.

Less really *is* more.

Chris Hornbeck
6x9=42

"Chris Hornbeck"
"xrongor"


i remember being taught that output impedance of the final stage should
match that of the load.

Maximum power into a load occurs when source and load impedances
are equal. This is true for audio purposes, but not a useful
goal.



** The maximum power transfer theorem has two conditions on it - the
source must be linear and the source impedance invariant with load.

This usually excludes active devices or circuits that employ NFB to reduce
output impedance values from complying with the theorem.




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

"Chris Hornbeck" wrote in message
...
On Thu, 31 Mar 2005 19:28:45 -0700, "xrongor"
wrote:

i remember being taught that output impedance of the final stage should
match that of the load.

Maximum power into a load occurs when source and load impedances
are equal. This is true for audio purposes, but not a useful
goal.

Others have answered the OP's question better than I could,
but I'd add a skew on the OP's perspective: each stage and its
loading should be considered in the context of the whole
shootin' match.

Meaning, in valve circuits, impedance buffering stages may
often be replaced with beefier preceding stages (not always,
of course). Others have already said this implicitly, so I'd
just add, fewer stages are better stages.

Less really *is* more.


so why cant you have both? a balanced load where Zout = Rload, and flat
frequency response? is there some fundamental conflict?

randy

"Chris Hornbeck" wrote

i remember being taught that output impedance of the final stage
should
match that of the load.

Maximum power into a load occurs when source and load impedances
are equal. This is true for audio purposes, but not a useful
goal....

This needs qualification to avoid two very common misunderstandings.

If you have a given voltage source of a given resistance, where
neither is dependent on current output, then maximum power will be
driven into a load of equal resistance. This is universally true at
any given instant in time.

If you have a load of given resistance, and a given voltage source,
then maximum power is driven into the load if the source resistance
is zero.

The first condition is broken by a voltage source whose resistance
depends on current, like a valve.

The second partly explains the 10-1 rule, which is really a guide to
the minimum acceptable. The power dissipated in a non-linear load
should be dependent only on the input voltage, not on the source
resistance. Otherwise the distortion in current will appear on the
voltage.

Finally, "matching" doesn't mean "making the same". As Phil, who I
have been missing, points out.

cheers, Ian

xrongor wrote:

"Patrick Turner" wrote in message
...


Wessel Dirksen wrote:

Hi Group,

This qualifies as a newbie question. Thanks for the insight in
advance.

I am an experienced loudspeaker designer with a background in
mechanics so I am familiar with designing loads, not driving loads. I
am curious if there are some good rule of thumb, generic wisdom type
of tidbits that might help me learn how to best go about matching
various stages in a total design to each other. I'm getting the hang
of determining in/out impedance characteristics, but not really
comfortable with knowing what's acceptable etc.

If anyone know a good reference on this subject this would be
appreciated.

Wessel Dirksen

The general rule for amplifiers is
"high impedance input,
low impedance output"

It means that the Zin for a give amp stage should be high relative to the

impedance of the preceding stage, and generally this is easy to achieve
because
tubes have extremely high input Z unless it is a tube in an output stage
where
grid current is flowing, or you have a grounded grid stage, neither of
which are
used very often in domestic audio gear.


At the output of the amp the Zout should be a small fraction of the
speaker's
minimum impedance, and having Zout = 1/10 of the nominal speaker Z
is a reasonable guide for tube amps.

This means that triodes must be used where little or nor global NFB is
to be used, or using pentode or tetrode output stages with about 20 dB of
NFB.

As you should know, Speakers are designed so that together with their
crossovers
they have a flat acoustic response, and crossovers work with the designed

attenuation slopes in the filters only when the Rout of the amp
is about 1/10 RL.
So it means an amp should have Rout between 0.5 ohms and 1 ohm.

As Rout is raised above 1 ohm, the response becomes unevern ( frequency
distortion )
and the rates of attenuation of the LCR filters become lower and
not as deigned.

Patrick Turner.

forgive me if this is a stupid question...

i remember being taught that output impedance of the final stage should
match that of the load.

i feel like im missing something here... can u enlighten me?

Consider the output stage without any global loop of negative FB applied.

There is no need for output impedance of the output stage to equal the load
impedance.

Pentode amps have a very high Ro, maybe 60 ohms,
UL stages give around 8 ohms, and are about the same
as the load,
and triodes can give a fraction of the RL, say 3 ohms.

Applying NFB around all these output stages will reduce the output impedance,
(or Ro.)

Nearly all amplifiers made use output stages
with higher source impedance without NFB than the load value.
Most are either pentode or UL.

NFB is what determines Rout more than any other factor, including
triode amps where the NFB is from the anode, and applied electrostatically
to the electrom stream.

Patrick Turner.




randy

On Fri, 1 Apr 2005 00:32:11 -0700, "xrongor"
wrote:

so why cant you have both? a balanced load where Zout = Rload, and flat
frequency response? is there some fundamental conflict?

Ah, OK, you're coming from some RF perspective. For audio purposes,
sources are *always* current limited, so power matching just doesn't
matter, (except very special cases like ribbon mic's), even back in
the old days of terminated lines.

In modern times frequency extension is acheived by brute force, low
source impedance, usually a result of negative feedback.

This works fine for most folks, but if you're needing to run lines
a kilometer or more, you'll need to think about terminating the ends.

For interstage coupling, impedance matching usually compromises
linearity, and usually to no other benefit.

Chris Hornbeck
6x9=42