Speaker Impedance and Amp Sound

[Audio Empire]

In an earlier post, I said that my Martin-Logan speakers fell below 2 Ohms
impedance, but that this should be inconsequential. I forgot to add that this
would prove inconsequential for most Solid-State amps (which was what I was
thinking about when I made the comment). Of course, it would likely be very
consequential for most tube amps, as they, generally, do NOT like highly
reactive loads. I still suspect that unless that reactance caused some kind
of non-linearity in the driving amp, that this would likely affect top-end
frequency response rather than distortion.

[Trevor Wilson[_3_]]

Audio Empire wrote:
> In an earlier post, I said that my Martin-Logan speakers fell below 2
> Ohms impedance, but that this should be inconsequential. I forgot to
> add that this would prove inconsequential for most Solid-State amps
> (which was what I was thinking about when I made the comment).

**Not so. In several ways:

1) Few solid state amps can cope with sub-2 Ohm loads without entering into
current limiting. Current limiting can result is very unpleasant sounds.
2) Most solid state amps utilise an output inductor, which causes a rising
output impedance relative to frequency. This can cause frequency response
issues under certain circumstances.

Of
> course, it would likely be very consequential for most tube amps, as
> they, generally, do NOT like highly reactive loads.

**On the contrary. Most valve amps are untroubled by highly reactive loads.
Current limiting is gentle and unobtrusive. Frequency response abberrations,
with highly reactive loads, is, however, a real issue.

I still suspect
> that unless that reactance caused some kind of non-linearity in the
> driving amp, that this would likely affect top-end frequency response
> rather than distortion.

**Of course. That said, sib- 2Ohm loads is likely to trigger current
limiters in many amps. Current limiting can be viewed as a very serious form
of non-linearity.


--
Trevor Wilson
www.rageaudio.com.au

[Audio Empire]

On Mon, 12 Jul 2010 03:31:13 -0700, Trevor Wilson wrote
(in article >):

> Audio Empire wrote:
>> In an earlier post, I said that my Martin-Logan speakers fell below 2
>> Ohms impedance, but that this should be inconsequential. I forgot to
>> add that this would prove inconsequential for most Solid-State amps
>> (which was what I was thinking about when I made the comment).
>
> **Not so. In several ways:
>
> 1) Few solid state amps can cope with sub-2 Ohm loads without entering into
> current limiting. Current limiting can result is very unpleasant sounds.
> 2) Most solid state amps utilise an output inductor, which causes a rising
> output impedance relative to frequency. This can cause frequency response
> issues under certain circumstances.

At 20 Khz and above? What recordings have enough 20 KHz high frequency energy
to trigger current limiting in any amplifier?

And, frankly, I haven't seen an SS amp with an output inductor in series with
the outputs for a long time. Not saying that some don't still use them, but
the amps I have here in house don't seem to have inductors and are, in fact,
direct-coupled to the speaker.

> Of
>> course, it would likely be very consequential for most tube amps, as
>> they, generally, do NOT like highly reactive loads.
>
> **On the contrary. Most valve amps are untroubled by highly reactive loads.
> Current limiting is gentle and unobtrusive. Frequency response abberrations,
> with highly reactive loads, is, however, a real issue.

Agreed, but due to these frequency response abberrations (which is what I was
referring to, rather than any nasty stuff such as oscillation or current
limiting), most electrostatic speakers don't "like" tube amps (unless they
are direct-coupled with NO transformers, like the old Beverage speakers from
the 1980's where the plates of the output tubes are connected directly to the
stators on the ESL element. Then, tubes and ESLs are a match made in heaven
[or some such place]). While my M-Ls sound fine with my VTL 140 monoblocks,
they sound much better with my Krell SS amp.

I once had a pair of Innersound ESLs here and they hated my VTLs - sounded
lousy with them. In fact, Innersound recommended that one NOT try to use them
with valve (Tube) amps because the Innersound ES speakers were very
reactive. Innersound then sent me their "companion" SS amp and it worked
fine as did my two bridged Rockford/Hafler "TransNova" SS amps.

> I still suspect
>> that unless that reactance caused some kind of non-linearity in the
>> driving amp, that this would likely affect top-end frequency response
>> rather than distortion.
>
> **Of course. That said, sib- 2Ohm loads is likely to trigger current
> limiters in many amps. Current limiting can be viewed as a very serious form
> of non-linearity.

But again, the M-Ls fall below two ohms above 20 KHz. The first harmonic of
20 KHz is 40 KHz, who cares if that harmonic or an odd-order harmonic of
20KHz is distorted? Dogs and mosquitos might hear it, people don't. Those
kinds of frequencies aren't very well represented on recordings anyway and
therefore the 20KHz content of any recording is unlikely to be of sufficient
amplitude to cause any but the poorest designed amplifiers to go into current
limiting mode (unless the reactive load has caused some kind of instability
and the amp is ultrasonically oscillating due to it).

[Trevor Wilson[_3_]]

Audio Empire wrote:
> On Mon, 12 Jul 2010 03:31:13 -0700, Trevor Wilson wrote
> (in article >):
>
>> Audio Empire wrote:
>>> In an earlier post, I said that my Martin-Logan speakers fell below
>>> 2 Ohms impedance, but that this should be inconsequential. I forgot
>>> to add that this would prove inconsequential for most Solid-State
>>> amps (which was what I was thinking about when I made the comment).
>>
>> **Not so. In several ways:
>>
>> 1) Few solid state amps can cope with sub-2 Ohm loads without
>> entering into current limiting. Current limiting can result is very
>> unpleasant sounds. 2) Most solid state amps utilise an output
>> inductor, which causes a rising output impedance relative to
>> frequency. This can cause frequency response issues under certain
>> circumstances.
>
> At 20 Khz and above?

**No. Problems that occur are related to the speaker impedance. At low
impedances (such as those typically found in ESLs), the output impedance of
a regular SS amp is dominated by the reactance of the output inductor. At 2
Ohms and below, that may play a part at frequencies lower than 20kHz. It is
most certainly an issue with Class D amps, where the output inductor/s can
play profound effects with ESLs.

What recordings have enough 20 KHz high
> frequency energy to trigger current limiting in any amplifier?

**Non-sequitur.

>
> And, frankly, I haven't seen an SS amp with an output inductor in
> series with the outputs for a long time.

**I have. I worked on (a late model) one yesterday. They are very very
common in SS amps.

Not saying that some don't
> still use them, but the amps I have here in house don't seem to have
> inductors and are, in fact, direct-coupled to the speaker.

**ALL Class D amps use output inductors. They must, if they are to meet
emission regulations. I acknowledge that some high end linear amps do not
use output inductors.

>
>> Of
>>> course, it would likely be very consequential for most tube amps, as
>>> they, generally, do NOT like highly reactive loads.
>>
>> **On the contrary. Most valve amps are untroubled by highly reactive
>> loads. Current limiting is gentle and unobtrusive. Frequency
>> response abberrations, with highly reactive loads, is, however, a
>> real issue.
>
> Agreed, but due to these frequency response abberrations (which is
> what I was referring to, rather than any nasty stuff such as
> oscillation or current limiting), most electrostatic speakers don't
> "like" tube amps (unless they are direct-coupled with NO
> transformers, like the old Beverage speakers from the 1980's where
> the plates of the output tubes are connected directly to the stators
> on the ESL element. Then, tubes and ESLs are a match made in heaven
> [or some such place]). While my M-Ls sound fine with my VTL 140
> monoblocks, they sound much better with my Krell SS amp.

**I'm sure they do.

>
> I once had a pair of Innersound ESLs here and they hated my VTLs -
> sounded lousy with them. In fact, Innersound recommended that one NOT
> try to use them with valve (Tube) amps because the Innersound ES
> speakers were very reactive. Innersound then sent me their
> "companion" SS amp and it worked fine as did my two bridged
> Rockford/Hafler "TransNova" SS amps.
>
>> I still suspect
>>> that unless that reactance caused some kind of non-linearity in the
>>> driving amp, that this would likely affect top-end frequency
>>> response rather than distortion.
>>
>> **Of course. That said, sib- 2Ohm loads is likely to trigger current
>> limiters in many amps. Current limiting can be viewed as a very
>> serious form of non-linearity.
>
> But again, the M-Ls fall below two ohms above 20 KHz.

**I don't know which MLs you have. I've measured some ESLs which fall to
less than 0.6 Ohms at around 15kHz. I measured one variant of the CLS which
fell to around 1 Ohm at around 16kHz.

The first
> harmonic of 20 KHz is 40 KHz, who cares if that harmonic or an
> odd-order harmonic of 20KHz is distorted?

**Again. Non-sequitur. The distortion created by current limiting systems is
normally non-harmonically related to the fundamental.

Dogs and mosquitos might
> hear it, people don't. Those kinds of frequencies aren't very well
> represented on recordings anyway and therefore the 20KHz content of
> any recording is unlikely to be of sufficient amplitude to cause any
> but the poorest designed amplifiers to go into current limiting mode
> (unless the reactive load has caused some kind of instability and the
> amp is ultrasonically oscillating due to it).

**And again: Non-sequitur.

--
Trevor Wilson
www.rageaudio.com.au

[Audio Empire]

On Mon, 12 Jul 2010 19:26:00 -0700, Trevor Wilson wrote
(in article >):

> Audio Empire wrote:
>> On Mon, 12 Jul 2010 03:31:13 -0700, Trevor Wilson wrote
>> (in article >):
>>
>>> Audio Empire wrote:
>>>> In an earlier post, I said that my Martin-Logan speakers fell below
>>>> 2 Ohms impedance, but that this should be inconsequential. I forgot
>>>> to add that this would prove inconsequential for most Solid-State
>>>> amps (which was what I was thinking about when I made the comment).
>>>
>>> **Not so. In several ways:
>>>
>>> 1) Few solid state amps can cope with sub-2 Ohm loads without
>>> entering into current limiting. Current limiting can result is very
>>> unpleasant sounds. 2) Most solid state amps utilise an output
>>> inductor, which causes a rising output impedance relative to
>>> frequency. This can cause frequency response issues under certain
>>> circumstances.
>>
>> At 20 Khz and above?
>
> **No. Problems that occur are related to the speaker impedance. At low
> impedances (such as those typically found in ESLs), the output impedance of
> a regular SS amp is dominated by the reactance of the output inductor. At 2
> Ohms and below, that may play a part at frequencies lower than 20kHz. It is
> most certainly an issue with Class D amps, where the output inductor/s can
> play profound effects with ESLs.
>
> What recordings have enough 20 KHz high
>> frequency energy to trigger current limiting in any amplifier?
>
> **Non-sequitur.
>
>>
>> And, frankly, I haven't seen an SS amp with an output inductor in
>> series with the outputs for a long time.
>
> **I have. I worked on (a late model) one yesterday. They are very very
> common in SS amps.
>
> Not saying that some don't
>> still use them, but the amps I have here in house don't seem to have
>> inductors and are, in fact, direct-coupled to the speaker.
>
> **ALL Class D amps use output inductors. They must, if they are to meet
> emission regulations. I acknowledge that some high end linear amps do not
> use output inductors.

I don't care about Class-D amps. Most SS amps are class A-B and analog.

>>> Of
>>>> course, it would likely be very consequential for most tube amps, as
>>>> they, generally, do NOT like highly reactive loads.
>>>
>>> **On the contrary. Most valve amps are untroubled by highly reactive
>>> loads. Current limiting is gentle and unobtrusive. Frequency
>>> response abberrations, with highly reactive loads, is, however, a
>>> real issue.
>>
>> Agreed, but due to these frequency response abberrations (which is
>> what I was referring to, rather than any nasty stuff such as
>> oscillation or current limiting), most electrostatic speakers don't
>> "like" tube amps (unless they are direct-coupled with NO
>> transformers, like the old Beverage speakers from the 1980's where
>> the plates of the output tubes are connected directly to the stators
>> on the ESL element. Then, tubes and ESLs are a match made in heaven
>> [or some such place]). While my M-Ls sound fine with my VTL 140
>> monoblocks, they sound much better with my Krell SS amp.
>
> **I'm sure they do.
>
>>
>> I once had a pair of Innersound ESLs here and they hated my VTLs -
>> sounded lousy with them. In fact, Innersound recommended that one NOT
>> try to use them with valve (Tube) amps because the Innersound ES
>> speakers were very reactive. Innersound then sent me their
>> "companion" SS amp and it worked fine as did my two bridged
>> Rockford/Hafler "TransNova" SS amps.
>>
>>> I still suspect
>>>> that unless that reactance caused some kind of non-linearity in the
>>>> driving amp, that this would likely affect top-end frequency
>>>> response rather than distortion.
>>>
>>> **Of course. That said, sib- 2Ohm loads is likely to trigger current
>>> limiters in many amps. Current limiting can be viewed as a very
>>> serious form of non-linearity.
>>
>> But again, the M-Ls fall below two ohms above 20 KHz.
>
> **I don't know which MLs you have. I've measured some ESLs which fall to
> less than 0.6 Ohms at around 15kHz. I measured one variant of the CLS which
> fell to around 1 Ohm at around 16kHz.
>
> The first
>> harmonic of 20 KHz is 40 KHz, who cares if that harmonic or an
>> odd-order harmonic of 20KHz is distorted?
>
> **Again. Non-sequitur. The distortion created by current limiting systems is
> normally non-harmonically related to the fundamental.

Believe me, it doesn't matter. While amps COULD current limit at the
frequencies where modern Martin-Logan ELS fall below 2 ohms, the amplitude of
the signals at those frequencies (20 KHz+) will most likely be so small as to
not cause the amp to go anywhere near current limiting. There is, simply,
very little energy above about 12-15 KHz on most recordings.

> Dogs and mosquitos might
>> hear it, people don't. Those kinds of frequencies aren't very well
>> represented on recordings anyway and therefore the 20KHz content of
>> any recording is unlikely to be of sufficient amplitude to cause any
>> but the poorest designed amplifiers to go into current limiting mode
>> (unless the reactive load has caused some kind of instability and the
>> amp is ultrasonically oscillating due to it).
>
> **And again: Non-sequitur.

Not at all a non-sequitur. You are talking about theoretical "problems" with
analog, linear amplifiers that just don't really exist in the real world. An
analog SS amplifier capable of delivering adequate current in the audio
passband is not going to choke, even with low impedance loads, on the small
amplitudes of ultra-high frequency program material present in any musical
recording made with microphones. direct-fed synthesized performances could,
possibly, have significant 20KHz+ material in them, but I doubt even that
would make it to a commercial recording.

[Arny Krueger]

"Audio Empire" > wrote in message


> On Mon, 12 Jul 2010 03:31:13 -0700, Trevor Wilson wrote
> (in article >):

>> Audio Empire wrote:
>>> In an earlier post, I said that my Martin-Logan
>>> speakers fell below 2 Ohms impedance, but that this
>>> should be inconsequential. I forgot to add that this
>>> would prove inconsequential for most Solid-State amps
>>> (which was what I was thinking about when I made the
>>> comment).

>> **Not so. In several ways:

>> 1) Few solid state amps can cope with sub-2 Ohm loads
>> without entering into current limiting.

The assertion is patently false given that there are so many amps that are
rated for 2 ohm operation, 20-20 KHz, that are readily available for
reasonable prices.


2) Most solid
>> state amps utilise an output inductor, which causes a
>> rising output impedance relative to frequency. This can
>> cause frequency response issues under certain
>> circumstances.

This is real. The usual design rule is to make this inductor as large as
posible consistent with minimal (usually 0.5 dB) loss at 20 KHz and minimum
rated resistive load. Since most speaker loads even planar and
electrostatic end up with rising impedance at the highest frequencies, they
generally perform better than this in the real world.

> At 20 Khz and above? What recordings have enough 20 KHz
> high frequency energy to trigger current limiting in any
> amplifier?

An excellent point. Spectral analysis of most recordings proves this
assertion to be correct.

> And, frankly, I haven't seen an SS amp with an output
> inductor in series with the outputs for a long time.

The purpose of the inductor relates to stability and longetivity. As output
device technology and other forms of protection improve, there's less need
for it.

However, I see a lot new product with the inductor and also many without it.
A blanket statement one way or the other seems unwarranted.

>> Of
>>> course, it would likely be very consequential for most
>>> tube amps, as they, generally, do NOT like highly
>>> reactive loads.

>> **On the contrary. Most valve amps are untroubled by
>> highly reactive loads. Current limiting is gentle and
>> unobtrusive. Frequency response abberrations, with
>> highly reactive loads, is, however, a real issue.

Most tubed amps will have greater frequency response aberrations due to
reactive loads simply because they present a much higher source impedance.
If you call that "untroubled" that would probably speak to some other
preferences.

> Agreed, but due to these frequency response abberrations
> (which is what I was referring to, rather than any nasty
> stuff such as oscillation or current limiting), most
> electrostatic speakers don't "like" tube amps (unless
> they are direct-coupled with NO transformers, like the
> old Beverage speakers from the 1980's where the plates of
> the output tubes are connected directly to the stators on
> the ESL element. Then, tubes and ESLs are a match made in
> heaven [or some such place]). While my M-Ls sound fine
> with my VTL 140 monoblocks, they sound much better with
> my Krell SS amp.

This all makes sense, again due to the typically lower source impedance
provided by good, modern SS amps.

> I once had a pair of Innersound ESLs here and they hated
> my VTLs - sounded lousy with them. In fact, Innersound
> recommended that one NOT try to use them with valve
> (Tube) amps because the Innersound ES speakers were very
> reactive. Innersound then sent me their "companion" SS
> amp and it worked fine as did my two bridged
> Rockford/Hafler "TransNova" SS amps.

FET-based amps were among the first SS amps to do well with tough loads.
However modern BJT output devices are more than up to any reasonble usage
sitaution.

>> I still suspect
>>> that unless that reactance caused some kind of
>>> non-linearity in the driving amp, that this would
>>> likely affect top-end frequency response rather than
>>> distortion.
>>
>> **Of course. That said, sib- 2Ohm loads is likely to
>> trigger current limiters in many amps. Current limiting
>> can be viewed as a very serious form of non-linearity.

Again, there are a ton of modern SS amps that are rated for and work well
with 2 ohm loads.

[Audio Empire]

On Tue, 13 Jul 2010 06:48:58 -0700, Arny Krueger wrote
(in article >):

> "Audio Empire" > wrote in message
>
>
>> On Mon, 12 Jul 2010 03:31:13 -0700, Trevor Wilson wrote
>> (in article >):
>
>>> Audio Empire wrote:
>>>> In an earlier post, I said that my Martin-Logan
>>>> speakers fell below 2 Ohms impedance, but that this
>>>> should be inconsequential. I forgot to add that this
>>>> would prove inconsequential for most Solid-State amps
>>>> (which was what I was thinking about when I made the
>>>> comment).
>
>>> **Not so. In several ways:
>
>>> 1) Few solid state amps can cope with sub-2 Ohm loads
>>> without entering into current limiting.
>
> The assertion is patently false given that there are so many amps that are
> rated for 2 ohm operation, 20-20 KHz, that are readily available for
> reasonable prices.
>
>
> 2) Most solid
>>> state amps utilise an output inductor, which causes a
>>> rising output impedance relative to frequency. This can
>>> cause frequency response issues under certain
>>> circumstances.
>
> This is real. The usual design rule is to make this inductor as large as
> posible consistent with minimal (usually 0.5 dB) loss at 20 KHz and minimum
> rated resistive load. Since most speaker loads even planar and
> electrostatic end up with rising impedance at the highest frequencies, they
> generally perform better than this in the real world.
>
>> At 20 Khz and above? What recordings have enough 20 KHz
>> high frequency energy to trigger current limiting in any
>> amplifier?
>
> An excellent point. Spectral analysis of most recordings proves this
> assertion to be correct.
>
>> And, frankly, I haven't seen an SS amp with an output
>> inductor in series with the outputs for a long time.
>
> The purpose of the inductor relates to stability and longetivity. As output
> device technology and other forms of protection improve, there's less need
> for it.
>
> However, I see a lot new product with the inductor and also many without it.
> A blanket statement one way or the other seems unwarranted.

I agree. None of the SS amps I have on-hand seem to have an inductor in
series with the output load, but I'm equally sure that many do.
>
>>> Of
>>>> course, it would likely be very consequential for most
>>>> tube amps, as they, generally, do NOT like highly
>>>> reactive loads.
>
>>> **On the contrary. Most valve amps are untroubled by
>>> highly reactive loads. Current limiting is gentle and
>>> unobtrusive. Frequency response abberrations, with
>>> highly reactive loads, is, however, a real issue.
>
> Most tubed amps will have greater frequency response aberrations due to
> reactive loads simply because they present a much higher source impedance.
> If you call that "untroubled" that would probably speak to some other
> preferences.
>
>> Agreed, but due to these frequency response abberrations
>> (which is what I was referring to, rather than any nasty
>> stuff such as oscillation or current limiting), most
>> electrostatic speakers don't "like" tube amps (unless
>> they are direct-coupled with NO transformers, like the
>> old Beverage speakers from the 1980's where the plates of
>> the output tubes are connected directly to the stators on
>> the ESL element. Then, tubes and ESLs are a match made in
>> heaven [or some such place]). While my M-Ls sound fine
>> with my VTL 140 monoblocks, they sound much better with
>> my Krell SS amp.
>
> This all makes sense, again due to the typically lower source impedance
> provided by good, modern SS amps.
>
>> I once had a pair of Innersound ESLs here and they hated
>> my VTLs - sounded lousy with them. In fact, Innersound
>> recommended that one NOT try to use them with valve
>> (Tube) amps because the Innersound ES speakers were very
>> reactive. Innersound then sent me their "companion" SS
>> amp and it worked fine as did my two bridged
>> Rockford/Hafler "TransNova" SS amps.
>
> FET-based amps were among the first SS amps to do well with tough loads.
> However modern BJT output devices are more than up to any reasonble usage
> sitaution.


That's more or less what I suspected.

>>> I still suspect
>>>> that unless that reactance caused some kind of
>>>> non-linearity in the driving amp, that this would
>>>> likely affect top-end frequency response rather than
>>>> distortion.
>>>
>>> **Of course. That said, sib- 2Ohm loads is likely to
>>> trigger current limiters in many amps. Current limiting
>>> can be viewed as a very serious form of non-linearity.
>
> Again, there are a ton of modern SS amps that are rated for and work well
> with 2 ohm loads.

....and even lower. Many SS amps use MOSFETS or, as you stated above, modern
bi-polar transistors such as ring-mesa types which tend to have astonishingly
low "on" impedances of less than an ohm. They have no problems with very low
speaker impedances.