or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Amplifiers affecting speaker frequency response
New Posts  All Forums:Forum Nav:

Amplifiers affecting speaker frequency response - Page 3

post #61 of 362
Quote:
Originally Posted by Shaun B View Post

Quote:
Originally Posted by Shaun B View Post


So my question is what kind of practical internal impedances/shenanigans forces the electronic designer to bend over backwards to try cater for such producers? Is this a necessary result of a superior loudspeaker, while other ostensibly equally good producers can leave one with a moderately tolerable impedance vs. frequency? I would like some broadening of my horisons here, please!


I'm hoping someone will clarify the above for me.

The marketing department knows that 8 ohm speakers sell best, so that is the spec they write. Reality strikes and you end up with a few frequency ranges where the speaker goes under 4 ohms.

If you want to tell truth in a way you adopt phrases like "8 ohm compatible"

I've been writing posts for the last 3 days about why this sort of thing can be unlikely to cause real world problems and somehow I don't want to interrupt my movie watching with my wife to repeat is again... ;-)

It's easy to find inexpensive speakers with this same thing going on, so its not peculiar to expensive ones.
post #62 of 362
Quote:
Although this doesn't answer the question... higher end speakers with a lower ohm rating are assumed the owner with have higher end amps as well that will handle them okay.

Speakers that are aimed at the entry level or home theatre market try to be 8 ohm normal rating so to present an easy load for AVRs etc.
I suspect that even most higher end speakers are specified at 8 ohms nominal, in order to be compatible with the widest variety of amps (including AVRs). Look at the B&Ws we've been discussing. And some AVRs can indeed drive 4 ohm speakers. But you're right that 4 ohm speakers are concentrated at the higher end of the market.
post #63 of 362
Thread Starter 

I just find it strange why speaker designers would make life so difficult for end users with some of their speaker designs when they could probably correct for the very low impedance dips.

 

So is it really all just for marketing? The really pricey gear have abnormal impedance dips, the cheap gear are easy to drive...? But who stands to benefit? Surely not the speaker manufactures, but the amplifier manufactures?


Are there any loudspeaker designers posting here? Perhaps you can shed some insight into why the expensive designs appear difficult to drive or at least appear so on paper.

 

Take a speaker like the Infinity Kappa 9.2. It has impedance dips down to 1 ohms. Well known for being very poorly optimised. Another speaker is the Infinity Kingdom.

 

Why do loudspeaker designers do this? Also, would it be possible to compensate for these impedance swings with active bi-amping?

post #64 of 362
Quote:
Originally Posted by arnyk View Post

... I don't want to interrupt my movie watching with my wife to repeat is again... ;-)

....
And, being curious what my friend was watching? wink.gifbiggrin.gif

Maybe we like similar movies too.
post #65 of 362
Quote:
I just find it strange why speaker designers would make life so difficult for end users with some of their speaker designs when they could probably correct for the very low impedance dips.
The problem is that they can't correct for one thing without changing a lot of other things. So they make a design decision that has market implications.
Quote:
So is it really all just for marketing?
Marketing is a big part of everything. Ever buy a car? smile.gif
Quote:
The really pricey gear have abnormal impedance dips, the cheap gear are easy to drive...?
Gross overgeneralization. Pricey speakers can be easy to drive.
Quote:
But who stands to benefit? Surely not the speaker manufactures,
If they don't benefit from their own designs, they are idiots who will soon be out of business. There is a high-end market, and they are catering to it.
Quote:
Perhaps you can shed some insight into why the expensive designs appear difficult to drive or at least appear so on paper.
Again, a gross overgeneralization. But of course, some high-end speaker makers worry less about impedance because they're selling through high-end channels where customers probably do have more capable amps (and salesmen are happy to "help them out" if they don't). I bet there are fewer cup holders in a Ferrari than in my Ford. It's not because Ferrari is trying to discourage drinking while driving! Just different markets with different priorities.
Quote:
Why do loudspeaker designers do this?
Repeat after me: Everything loudspeaker designers do, they do because there's a market for it. Same as every other consumer product.
Quote:
Also, would it be possible to compensate for these impedance swings with active bi-amping?
That's not what active bi-amping is for. The way to handle a speaker with very low impedance is to buy an amp that can drive very low impedance speakers. It's that simple.
post #66 of 362
Quote:
Originally Posted by Shaun B View Post

I just find it strange why speaker designers would make life so difficult for end users with some of their speaker designs when they could probably correct for the very low impedance dips.

So is it really all just for marketing? The really pricey gear have abnormal impedance dips, the cheap gear are easy to drive...? But who stands to benefit? Surely not the speaker manufactures, but the amplifier manufactures?


Are there any loudspeaker designers posting here? Perhaps you can shed some insight into why the expensive designs appear difficult to drive or at least appear so on paper.

Take a speaker like the Infinity Kappa 9.2. It has impedance dips down to 1 ohms. Well known for being very poorly optimised. Another speaker is the Infinity Kingdom.

Why do loudspeaker designers do this? Also, would it be possible to compensate for these impedance swings with active bi-amping?
The impedance dips are not as big a deal as many make it out to be. Of course it depends on the dip-how wide-how many etc etc.

Remember that the amp will only "see" the low impedance if there are notes in the music at those freq. If there aren't-it doesn't matter.

And how long are the notes "staying" at that freq?

It is not always an easy thing to "design without the impedance dips". In many cases you have to drop the impedance a little bit in order to flatten out either the amplitude or phase response-or both.

Of course you could "design" without those dips-but then the response would not be as flat-and that is part of what people "expect" in the upper end speakers.

The cheap speakers generally don't have them-and their response is much more ragged. It costs money to "put the dips there" in terms of crossover parts. This raises the cost of the speaker-but also makes it perform better.

The bigger issue of "hard to drive" is the PHASE of the impedance. This can put a real load on the amp-if it is to capacitive or inductive. Typically past +/-45° is asking for trouble.

Of course you never see the phase response of the impedance-so you never know what it is-but I always monitor it during the design process-to make sure it stays within bounds.
post #67 of 362
+1. Right, if you are worried about stressing an amplifier, the maximum impedance phase angle (how reactive the impedance is) is just as important as the minimum impedance measurement.

A low impedance load will require more current output and will cause additional heat generation from the amplifier output stages - a highly reactive low impedance load will require just as much current but will generate significantly more heat.

Think of it this way. A purely resistive load will convert the power to heat and dissipate the power generated whereas a capacitive or inductive load will store the power and return it to the amplifier. This will not only stress the output stage but will stress the thermal design of the amp.

Some amplifier manufacturers use just this kind of load to "torture test" both the electrical and thermal performance to find and components that will fail prematurely.
post #68 of 362
Quote:
Originally Posted by arnyk View Post

A number of people seem to have figured out that writing and delivering amps that meet that spec is easy, you just make the 8 ohm rating very conservative. IOW the amp actually puts out 150 watts at 8 ohms but you rate it at 100. You rate it at 200 watts at 4 ohms so it appears that the amp is capable of "doubling down". Who is going to fault you for a pessimistic 8 ohm amplifier rating?

They would be losing a lot of potential customers that were in the market for a 150W 8 ohm amp and would turn their nose up at a 100W amp. They all know bigger numbers sell.
post #69 of 362
Quote:
Originally Posted by kiwi2 View Post

Quote:
Originally Posted by arnyk View Post

A number of people seem to have figured out that writing and delivering amps that meet that spec is easy, you just make the 8 ohm rating very conservative. IOW the amp actually puts out 150 watts at 8 ohms but you rate it at 100. You rate it at 200 watts at 4 ohms so it appears that the amp is capable of "doubling down". Who is going to fault you for a pessimistic 8 ohm amplifier rating?

They would be losing a lot of potential customers that were in the market for a 150W 8 ohm amp and would turn their nose up at a 100W amp. They all know bigger numbers sell.

Yes, but if it's part of a tier, why not make the numbers work for your whole lineup rather than just a particular unit? Might impact their 200W amp to have a 150....and I'll bet the higher wattage amps have comparatively less cost/more profit.
post #70 of 362
Quote:
Originally Posted by kiwi2 View Post

They would be losing a lot of potential customers that were in the market for a 150W 8 ohm amp and would turn their nose up at a 100W amp. They all know bigger numbers sell.
True, the big numbers sell. But I suspect they will be selling the "200 watt" number, and leave the 8 ohm spec in the fine print.
post #71 of 362
Quote:
Originally Posted by kiwi2 View Post

Quote:
Originally Posted by arnyk View Post

A number of people seem to have figured out that writing and delivering amps that meet that spec is easy, you just make the 8 ohm rating very conservative. IOW the amp actually puts out 150 watts at 8 ohms but you rate it at 100. You rate it at 200 watts at 4 ohms so it appears that the amp is capable of "doubling down". Who is going to fault you for a pessimistic 8 ohm amplifier rating?

They would be losing a lot of potential customers that were in the market for a 150W 8 ohm amp and would turn their nose up at a 100W amp. They all know bigger numbers sell.

It's a risk but apparently an acceptable risk. The payoff can be high end panache.

Underrating amplifier power was the essence of HK's "High Current Amp" stategy. Underrating amplifier power is the essence of Krell's double down amp strategy. I believe that Marantz is doing something similar.

Recently in another thread I was challenged to provide a rel world example of a integrated amp that amounted to being an existing AVR design with features removed. I lucked out. I found pictures of an 50 pwc integrated amp that had guts that looked for all the world to be identical with a 100 wpc receiver sold under a value brand for a fraction of the price.
post #72 of 362
The parameters for testing is pretty straight forward but is done with single or double tones. What is music?? Certainly not that simple. Complex notes with sub-harmonics and harmonics galore. That can't be measured or taken into account. Sampling is a waste of time. Current sensors are good but again, impedance plays a big part in current draw. Exceeding the speaker parameters for any amp will push those final stage devices beyond their specs, resulting in damaged speakers and overheated amplifiers.
post #73 of 362
Quote:
Complex notes with sub-harmonics and harmonics galore


What gets amplified is not notes and harmonics, what gets amplified is the voltage (differences) of the signal. Check out an oscilloscope trace of a music signal. The curve is the signal, and that is all. That curve includes all your harmonics etc. There is nothing below that curve that an amp will respond to. A relative simple signal.
http://www.audiofile-engineering.com/spectre/images/screenshots/techspecs/meters/oscilloscope.png


Quote:
The parameters for testing is pretty straight forward but is done with single or double tones

Utter nonsense.

What amplifiers usually are tested with is not a simple sine frequency, what they get hammered with in testing is a square wave.

http://en.wikipedia.org/wiki/Square_wave

http://www.kennethkuhn.com/students/ee351/square_wave_testing.pdf

Now check how any good amp will reply to a really complex signal:



http://www.stereophile.com/content/bryston-b-60r-integrated-amplifier-measurements



http://www.stereophile.com/content/bryston-3b-st-power-amplifier-measurements
Edited by kraut - 11/11/13 at 9:14am
post #74 of 362
Quote:
Originally Posted by arnyk View Post

Underrating amplifier power is the essence of Krell's double down amp strategy..

The only Krell I've seen personally had an overblown fully regulated power supply, more complex than some entire amplifiers. It was stable and doubled down almost to a dead short. This was some time ago, perhaps they've changed their approach in recent years (it sure would make economic sense, and who really needs an amp that can double as an arc welder?), I just don't know, but in that particular case it was not playing with the numbers but over-engineering to the power supply that did it.
post #75 of 362
Quote:
Originally Posted by kraut View Post

Quote:
Complex notes with sub-harmonics and harmonics galore


What gets amplified is not notes and harmonics, what gets amplified is the voltage (differences) of the signal. Check out an oscilloscope trace of a music signal. The curve is the signal, and that is all. That curve includes all your harmonics etc. There is nothing below that curve that an amp will respond to. A relative simple signal.
http://www.audiofile-engineering.com/spectre/images/screenshots/techspecs/meters/oscilloscope.png


Quote:
The parameters for testing is pretty straight forward but is done with single or double tones

Agreed.
Quote:
Utter nonsense.

I'l try to be nice and not return the harsh reply.
Quote:
What amplifiers usually are tested with is not a simple sine frequency, what they get hammered with in testing is a square wave.

The testing to confirm most spec sheets can and is accomplished with single or double tones.

Quick summary:

The classic means for testing amps are based on the following test signals:

Frequency response - chosen spot tones
THD - spot single tones
IM - double tones - SMPTE IM uses a bass tone and a treble tone, while CCIR IM uses two high frequency tones.

In modern times more different kinds of test tones can be used because we have the analysis software to isolate the various components of the complex results that they produce. Swept tones and multiple tones can be used, but they are still composed of pure sine waves.


The problem is one of interpretation. A square wave is just a collection of a the fundamental and a series of odd harmonics of a sine wave (see the equationsin the Wikipedia article for example), with the harmonics weighted in a specific way, namely each odd harmonic present is weighted proportional to the inverse of its order.

In the olden days we lacked the complex analyzers to decode this sort of thing. If you tried to run a THD test with a square wave you had the harmonics of the square wave right where you were hoping to find harmonics generated by the UUT. There were similar problems with IM testing. Frequency response testing had that problem, plus the fact that at high frequencies the harmonics were too close to each other. If you use say a 20 Hz square wave to run a frequency response test, you only test the amp at 20, 60, 100, Hz etc, when you really wanted to test the amp at 20, 25, 30, 40, 50, 80, 100..Hz. At the high frequencies the harmonics are dozens of dB down due to the 1/n weighting, and your harmonics are still odd multiples of 20 Hz which are 40 Hz apart so that around 20 KHz the components are way too close together to analyze.

It just wasn't done. Since waves at the right frequencies were just a crank of a knob away.

If a square wave was used it was use to get an idea of low frequency phase shift, high frequency ringing and slew rate. Note that none of those things show up on a typical amplifier spec sheet. Slew rate was only a serious issue in roughly the 1970s and 1980s and was found to be subsumed by other measurements namely high frequency nonlinear distortion, and more understandable that way. It fell out of fashion and is rarely mentioned these days. Another reason why square waves fell out of favor for high frequency testing is that due to the widespread use of digital media sampled in the 40 KHz range, almost all signals in acttual use are brutally band limited in the 20-25 KHz range.

In modern times we can extract a lot of knowledge using a wide variety of test signals including music - we can do a useful FR test of an amp just by playing music through it, no special test tones required.

I did my first test equipment test of a power amp in or around 1960 and have bench tested several 100 amplifiers over the years. In 1996 I started doing FFT -based equpment tests. I have my own personal AP test set operational in my home. So I literally grew up with this technology and kept on flogging it into my retirement.
post #76 of 362
Quote:
Originally Posted by Wayne Highwood View Post

Quote:
Originally Posted by arnyk View Post

Underrating amplifier power is the essence of Krell's double down amp strategy..

The only Krell I've seen personally had an overblown fully regulated power supply, more complex than some entire amplifiers. It was stable and doubled down almost to a dead short. This was some time ago, perhaps they've changed their approach in recent years (it sure would make economic sense, and who really needs an amp that can double as an arc welder?), I just don't know, but in that particular case it was not playing with the numbers but over-engineering to the power supply that did it.

Here's an example - the actual performance of a Krell amp:

http://www.stereophile.com/content/krell-evolution-202-preamplifier-600-monoblock-power-amplifier-measurements



the specs are: Output power: 600W RMS into 8 ohms (27.8dBW), 1200W into 4 ohms (27.8dBW).

the actual tests shows:
clipping starts at 620 watts @ 8 ohms, 1100 watts @ 4 ohms, and 1800 watts @ 2 ohms
0.1% THD 650 watts @ 8 ohms, 1200 watts @ 4 ohms, and 1950 watts @ 2 ohms

No signs of precise doubling down can be seen. The spec is underrated at 0.1% THD at 8 ohms, about spot on at 4 ohms, and no 2 ohm specification is given.

http://www.stereophile.com/content/krell-evolution-202-preamplifier-600-monoblock-power-amplifier-specifications

There are a ton of threads over at diyaudio about reverse engineering Krell amps including this one: http://www.diyaudio.com/forums/solid-state/222801-krell-ksa-250-mother-clone.html#post3226635
Edited by arnyk - 11/11/13 at 11:01am
post #77 of 362
Arny, you bought your used audio precision analyzer (AP) just a couple of years ago: https://groups.google.com/forum/#!topic/rec.audio.tubes/_B4-nfkFiqM. So that fact doesn't qualify for "growing up" with it. We have not needed complex tools to analyze an amps with square wave in the past or today. Squarewave is revealing by itself in analog domain so you don't need to perform FFT or any other name dropping term. What it reveals is there in time domain display of a scope.

Squarewave analysis of amplifier done decades back and now is one in the same. It is to find hints of amplifier instability and rise time/bandwidth limiting. You mention those aspects but then go off with some random assumption that OP suggested using square waves for frequency response measurement or THD. That is not what he is saying. He is saying that transient response (in the form of square waves) is also important to look at and steady state signals don't show that.
post #78 of 362
Quote:
Originally Posted by amirm View Post

Arny, you bought your used audio precision analyzer (AP) just a couple of years ago: https://groups.google.com/forum/#!topic/rec.audio.tubes/_B4-nfkFiqM. So that fact doesn't qualify for "growing up" with it. We have not needed complex tools to analyze an amps with square wave in the past or today. Squarewave is revealing by itself in analog domain so you don't need to perform FFT or any other name dropping term. What it reveals is there in time domain display of a scope.

False claims, all.

I never ever bought any AP test set ever. The equipment I have was once used by Julian Hirsch at Stereo Review and was given to me, FOB Manhattan, NY, NY.

While the 2011 post says what it says, I had the AP test sets in my possession longer than that. It only says that I was performing a certain project related to them, not when or how I acquired them. The errors are your own invention, to what end I cannot understand.

They are far from the only AP test sets that I have been hands-on with.
Quote:
Squarewave analysis of amplifier done decades back and now is one in the same. It is to find hints of amplifier instability and rise time/bandwidth limiting. You mention those aspects but then go off with some random assumption that OP suggested using square waves for frequency response measurement or THD. That is not what he is saying. He is saying that transient response (in the form of square waves) is also important to look at and steady state signals don't show that.

Only one of the glaring errors above is the claim that square waves are transients. They are in fact steady state collections of sine waves. That's engineering 201 in many good Universities.

The usual and expected odd misinterpretation of the relevant facts is again seen above. In fact the OP said that in most amp tests "The parameters for testing is pretty straight forward but is done with single or double tones" and the post I responded to said, and I quote: "Utter nonsense". It wasn't utter nonsense, it was the way this particular universe works and has worked for decades.
post #79 of 362
Thread Starter 
Quote:
Originally Posted by Ivan Beaver View Post


The impedance dips are not as big a deal as many make it out to be. Of course it depends on the dip-how wide-how many etc etc.

Remember that the amp will only "see" the low impedance if there are notes in the music at those freq. If there aren't-it doesn't matter.

And how long are the notes "staying" at that freq?

It is not always an easy thing to "design without the impedance dips". In many cases you have to drop the impedance a little bit in order to flatten out either the amplitude or phase response-or both.

Of course you could "design" without those dips-but then the response would not be as flat-and that is part of what people "expect" in the upper end speakers.

The cheap speakers generally don't have them-and their response is much more ragged. It costs money to "put the dips there" in terms of crossover parts. This raises the cost of the speaker-but also makes it perform better.

The bigger issue of "hard to drive" is the PHASE of the impedance. This can put a real load on the amp-if it is to capacitive or inductive. Typically past +/-45° is asking for trouble.

Of course you never see the phase response of the impedance-so you never know what it is-but I always monitor it during the design process-to make sure it stays within bounds.

 

So a speaker that does not have very low impedance dips, won't have a flat response? And if you correct for it then the response won't be flat? Ditto about the flat response in upper end speakers, although looking through Stereophile it seems many high-end speakers measure poorly. Poor as in very uneven frequency response. Have you seen some of the Wilson Audio designs?

post #80 of 362
Quote:
although looking through Stereophile it seems many high-end speakers measure poorly. Poor as in very uneven frequency response.
Yeah, just because a speaker is expensive doesn't mean it's any good. And you notice how those expensive but lousy speakers still get glowing reviews? That should tell you something about the reviewers.
post #81 of 362

Beefy amplifiers just have more headroom to cope with demanding climatic musical passages when the average volume is already high enough to make you deaf... :D 

post #82 of 362
Quote:
Originally Posted by Shaun B View Post

So a speaker that does not have very low impedance dips, won't have a flat response?

Not necessarily. Speaker impedance curves can get bumpy and low as a byproduct of efforts to smooth frequency response, but that doesn't mean that smooth impedance curves always go with poor frequency response.

For one thing one could hang a 12 ohm resistor across the terminals of many speakers, obtain a magically smoother impedance curve, and sacrifice a little efficiency. More people would try this little deceit, if they thought that enough people cared.
post #83 of 362
Thread Starter 
Quote:
Originally Posted by arnyk View Post


Not necessarily. Speaker impedance curves can get bumpy and low as a byproduct of efforts to smooth frequency response, but that doesn't mean that smooth impedance curves always go with poor frequency response.

For one thing one could hang a 12 ohm resistor across the terminals of many speakers, obtain a magically smoother impedance curve, and sacrifice a little efficiency. More people would try this little deceit, if they thought that enough people cared.

 

If a flat response could be achieved without abnormal impedance dips, then why do loudspeaker designers often take that route? The phase of the impedance is so poor in some high-end designs, but it's not as if the designers are handcuffed in the design process. Surely the idea would be to create a speaker that does not present a super complex load to the amp? I guess what confuses me is that if loudspeaker designers can design a speaker to measure well without all the drama of super low impedance dips and weird phase angles, then why don't they? Otherwise their work could be seen as just very poorly optimised, wouldn't you think?

post #84 of 362
Quote:
If a flat response could be achieved without abnormal impedance dips, then why do loudspeaker designers often take that route?
Well, there are other parameters besides those two. Arny mentioned efficiency, for example. And, especially at the high end, there's a premium on distinctiveness. You don't want to produce the same plain-vanilla speaker everybody else is making. You want something that'll make the boys at Sphile and TAS sit up and say "Wow!" And what makes them sit up isn't a good speaker; it's a distinctive speaker.
Quote:
The phase of the impedance is so poor in some high-end designs, but it's not as if the designers are handcuffed in the design process. Surely the idea would be to create a speaker that does not present a super complex load to the amp? I guess what confuses me is that if loudspeaker designers can design a speaker to measure well without all the drama of super low impedance dips and weird phase angles, then why don't they? Otherwise their work could be seen as just very poorly optimised, wouldn't you think?
I would agree with you that a lot of high-end speakers are "poorly optimized." But I also think you're putting way too much emphasis on this impedance problem. Lots of speakers which the Internet audiophile echo chamber has deemed "difficult to drive" aren't necessarily so. And the people who can afford those distinctive high-end designs with the wavy impedance and phase angle curves can certainly afford the amps necessary to drive them.

I'm also not sure why this is so important to you. If I wanted to assemble, say, a $10K system, I'd go out and choose the best-sounding speakers I could find. Then I'd buy whatever kind of amp was appropriate to drive them. It'd be great if a $100 receiver could do the trick, but I wouldn't feel the speaker maker had done me wrong if I needed something better.
post #85 of 362
@arny

I responded to the claim how amps are tested:
Quote:
The parameters for testing is pretty straight forward but is done with single or double tones
A statement i know not to be true at least for the amps I used to own.


http://www.roger-russell.com/truth/truth.htm
Quote:
How Fast Is Your Audio System?

Many years ago, Harman-Kardon began emphasizing the importance of square wave response in audio electronics. Today, the term "square wave" is no longer being used but a more obscure term called "fast" is used. Capacitors, speakers and amplifiers, etc. are referred to as being fast or perhaps slow. The implication that something might be slow in today's world of faster computers would be undesirable. Fast computer processing time is, in fact, desirable. Audio system response and hearing requirements can be something entirely different.

A square waveform rises infinitely fast, holds at constant amplitude and then falls infinitely fast to an equal negative value, holds at constant amplitude and then rises infinitely fast to repeat over again. A square wave signal can be generated electronically and can cover a frequency range from of 20Hz to 20kHz or even wider. It can be applied to a preamplifier, power amplifier or other audio equipment and the resulting output observed on an oscilloscope.

Harmonic analysis of a square wave shows that it is made of an infinite array of odd order harmonics that decrease in amplitude. It consists of the fundamental frequency that has an amplitude of one. The 3rd harmonic has an amplitude of 1/3 of the fundamental. The 5th harmonic has an amplitude of 1/5th of the fundamental, etc. The fundamental and each of these harmonics can be thought of as a pure tone or what is called a sine wave. The lower harmonics are the most significant in contributing to the shape or "speed" of a square wave and the shape of the leading and falling edges. As the harmonics get even higher, they contribute less and less to the leading and trailing edges until they become negligible.

As a rule, the bandwidth of an amplifier that will reproduce a good looking square wave must extend from 1/10th of the fundamental frequency of a test square wave to 10 times its frequency. For example, if an amplifier passes a 2000 Hz square wave without distortion, the amplifier should have flat response and be free from phase shift or oscillation from at least 200 to 20,000 Hz. But what if a 20kHz square wave is applied? Should the amplifier output also be a square wave? The fundamental is 20kHz, third harmonic is 60kHz and the 5th harmonic is 100kHz.. Tests indicate that we don't hear pure tones beyond 20kHz.

I remember square wave tests of audio equipment in German audio magazines done in the early seventies.

Even tube amps

http://www.r-type.org/articles/art-125.htm
Quote:
Square Wave Testing

Sine waves can be used, as has been shown, to check for distortion in audio amplifiers, but a more definite test is afforded by square waves. The effects of square waves can be seen in the diagrams below..
post #86 of 362
Quote:
Originally Posted by Shaun B View Post

So a speaker that does not have very low impedance dips, won't have a flat response? And if you correct for it then the response won't be flat? Ditto about the flat response in upper end speakers, although looking through Stereophile it seems many high-end speakers measure poorly. Poor as in very uneven frequency response. Have you seen some of the Wilson Audio designs?
I was talking in "general terms". I did not mean to imply that you have to have a low impedance dip to have a flat response. Just that it happens sometimes

I could not find any measurements (freq response and impedance) on the Wilson Audio site-could you provide a link?
post #87 of 362
Quote:
Originally Posted by Shaun B View Post

Quote:
Originally Posted by arnyk View Post

Not necessarily. Speaker impedance curves can get bumpy and low as a byproduct of efforts to smooth frequency response, but that doesn't mean that smooth impedance curves always go with poor frequency response.


For one thing one could hang a 12 ohm resistor across the terminals of many speakers, obtain a magically smoother impedance curve, and sacrifice a little efficiency. More people would try this little deceit, if they thought that enough people cared.

If a flat response could be achieved without abnormal impedance dips, then why do loudspeaker designers often take that route?

I was talking about a cosmetic change, not what you are talking about.

First off, characterizing many dips as "abnormal impedance dips" is too severe. A dip down to 2 ohms is abnormal, but 3 or 4 ohms is OK. It is the very rare speaker that dips down below 3 ohms.
Quote:
The phase of the impedance is so poor in some high-end designs, but it's not as if the designers are handcuffed in the design process.

Again what's a poor impedance phase? If the impedance is high, then large phase angles are permissable. It;s the combination of low impedance high phase angle and presence in a wide region where music is very energetic that can cause problems for some amplifiers. I'm not sure that everybody understands what modern amplifiers find daunting.
Quote:
Surely the idea would be to create a speaker that does not present a super complex load to the amp?

It's not the complexity but ratther its actual impact.
Quote:
I guess what confuses me is that if loudspeaker designers can design a speaker to measure well without all the drama of super low impedance dips and weird phase angles, then why don't they?

Most designers don't want to make a speaker that will be hard to drive. They only do so as a last resort. There is trading off of a weird looking impedance curve for a speaker that has good frequency response. However, just because the curve looks weird to a non-technical reader doesn't mean that there is necessarily a problem.
Quote:
Otherwise their work could be seen as just very poorly optimised, wouldn't you think?

I think the biggest fear is that the speaker will sound bad with a lot of amps that it is likely to be used with. Returns are very expensive.
post #88 of 362
Square wave testing has largely fallen out of favor and interest, as a practical matter.

For example this is a recent amplifier test from Stereophile.

http://www.stereophile.com/content/constellation-audio-performance-centaur-mono-monoblock-power-amplifier-measurements

1 square wave test is almost lost among dozens of other kinds of test. There are 10 tolal figures of results with each showing the results of several (usually abut 3) different actual tests. The square wave test is therefore only about 3% of the test results - nearly lost.

And here is a recent test from Sound and Vision

http://www.soundandvision.com/content/nad-d-3020-hybrid-digital-integrated-amplifier-sv-labs-measures

No results of square wave testing is shown at all

Here is a typical power amp test test by the NRC for Soundstage! webizine

http://www.soundstagemagazine.com/measurements/threshold_s5000e/

No square wave tests of any kind.

Finally a typical Audioholics amplfiier test]

http://www.audioholics.com/amplifier-reviews/classe-ct-2300/classe-ct-2300-measurements

No square wave tests at all.
post #89 of 362
Moreover, S'phile's square wave tests are, IIRC, uniformly "small signal" not anywhere near the amp's clipping point and so not about assessing harmonic distortion. But they're fun to look at . . .
post #90 of 362
Quote:
Originally Posted by arnyk View Post

Most designers don't want to make a speaker that will be hard to drive. They only do so as a last resort. There is trading off of a weird looking impedance curve for a speaker that has good frequency response. However, just because the curve looks weird to a non-technical reader doesn't mean that there is necessarily a problem.

Well said. As with other engineering, speaker design is full of trade offs - there is size, cost, performance, compatibility, reliability, etc - usually each attribute comes at the expense of another. If you design to a particular size and cost, you will sacrifice performance. But that same size with more expensive drivers and crossover components may have much better performance.

If you want the flattest possible frequency response (cost no object design) you might trade size (larger / more drivers), cost, and a less than optimal impedance curve (compatibility). But as others have said, it is unlikely that someone that purchases such an expensive speaker would try to drive them with an amp that couldn't handle a "difficult" load or would be too upset if they needed a more robust amp.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Amplifiers affecting speaker frequency response