nigelwright7557 said:
And then people do even sillier things like putting a varying current through the coil and not having consistent suspensions ('n' stuff).
In a world where many people have trouble identifying a 3dB change (that is, being able to point out what it is) and also have trouble distinguishing a 1dB change (that is, recognising it is there) in an unfamiliar system, these sometimes aren't significant.
Hey, I found it on wiki (here), and it's even more complicated than I thought!
And even the model in the Wiki link is ambiguous, ie in my speaker design textbook the mechanical part (the middle part of the drawing) is a series circuit with L=Mms and C=Cms, which is perfectly equivalent, though.
I think that Nigel got confused by the fact that the Elliott diagram shows the equivalent electrical network that produces the same impedance (as seen by the amp) as a real driver, while the Wiki drawing uses electromechanical analogies to model the complete speaker's behaviour.
- Klaus
I think that Nigel got confused by the fact that the Elliott diagram shows the equivalent electrical network that produces the same impedance (as seen by the amp) as a real driver, while the Wiki drawing uses electromechanical analogies to model the complete speaker's behaviour.
- Klaus
and that Wiki shows a single speaker.
Now put two or three of them joined up with a complex crossover, maybe even a Zobel or two to flatten the sinewave impedance graph.
Feed in a fast starting or fast stopping signal and measure the peak current.
For most speakers, I'm told the Ipk>Vpk/Re. I have seen a few publish figures showing peak currents approaching 3*Vpk/Rnom.
The crux of that Ipk is the fast starting/stopping music signal.
I wonder if a square wave test signal could be used to examine the differences that various speakers might exhibit even though ALL the test drivers are a nominal 8ohm in the 8ohm speaker. The test square wave must of necessity pass through the RF filter built into the power amp input, thus automatically limiting the rise time of the test signal.
Now put two or three of them joined up with a complex crossover, maybe even a Zobel or two to flatten the sinewave impedance graph.
Feed in a fast starting or fast stopping signal and measure the peak current.
For most speakers, I'm told the Ipk>Vpk/Re. I have seen a few publish figures showing peak currents approaching 3*Vpk/Rnom.
The crux of that Ipk is the fast starting/stopping music signal.
I wonder if a square wave test signal could be used to examine the differences that various speakers might exhibit even though ALL the test drivers are a nominal 8ohm in the 8ohm speaker. The test square wave must of necessity pass through the RF filter built into the power amp input, thus automatically limiting the rise time of the test signal.
graaf said:
Hi,
Your welcome to take a graph produced for marketing purposes seriously
and ignore the concensus opinion on the meaningfulness of that graph.
Sign up for your psuedotechnobabble course whilst your at it ...
😉sreten.
sreten said:
and what is not produced for marketing purposes nowadays?
sreten said:
consesus? what consensus?
sreten said:
my pseudotechnobabble? I am just a music lover, technical layman asking, trying to understand
what is wrong with me?
and who are you? what is wrong with You?
rethorical questions
🙂 /graaf
To the original question, if you take the text at face value, they did measure impedance using a sine wave. Then they measured something else using a square wave. It was not impedance because the definition of that is very specific and the correct answer is the one they got with the sine wave. I don't see where the lower plot tells you anything useful, and it's quite misleading to plot it alongside a true impedance curve. As I said above in different words, you really need to understand the reactance of the load vs frequency to be able to predict anything useful about the speakers effect on the amplifier. As a stupid example, let's say a speaker impedance was 8 ohms at 1kHz. You drive it with 28.28 volts RMS at 1kHz. How much power does the speaker dissipate? You'd be completely wrong if you said 99.97 Watts. Why? Because the 8 ohms was an impedance, not a resistance. I neglected to tell you that the impedance was purely capacitive reactance, about 19.9uF, meaning that 100% of the power would be dissipated in the power amp. A reactance can't dissipate real power, but it will stress the amplifier outputs like crazy. Thus, you need to know more than just the impedance number- you need to know the phase angle or the reactance. BTW, you can derive all the info you need from a single step (or a square wave) via FFT analysis- that's one way to measure speakers and beat the energy reflected from the room. Or measure the electrical response. But there's no evidence they did that.
Conrad Hoffman said:
thank You for Your interest and kind response 🙂
so the claim of the Audiograph people that:
is simply not true because the type of the signal (sine wave, square wave, sine wave + pink noise, music) is irrelevant ie. the load remains the same (and depending rather on sine wave impedance response plus its "phase angle or the reactance")?
best,
graaf
Najlepiej opisa³ to moim zdaniem Cordell:
http://www.cordellaudio.com/papers/interface_intermodulation_distortion.pdf
Spróbuj siê przez to przegry¼æ, najbardziej wymowne jest "Figure13", opisuje w zasadzie to samo co w tym artykuliku ale bardziej sci¶le.
Pozdrawiam
http://www.cordellaudio.com/papers/interface_intermodulation_distortion.pdf
Spróbuj siê przez to przegry¼æ, najbardziej wymowne jest "Figure13", opisuje w zasadzie to samo co w tym artykuliku ale bardziej sci¶le.
Pozdrawiam
graaf said:
But is the load always the same ?
Sure the reactance goes up in the speaker as frequency rises ?
Unless the Zobel network perfectly cancels the inductance in the speaker then the impedance is not going to be flat.
graaf said:
Hi,
A carefully constructed test signal with transitions timed to relate
to the complex impedance time constants and stored energy of a
loudspeaker will indicate peak currents are required not indicated
by the impedance curve.
However music signals do not resemble such test signals in any
way is the concensus. The given "square wave" impedance graph
is stated without further explanation and as many have said in this
thread it is essentially meaningless (not a standard measurement)
without full details of the derivation of the curve.
One could add a badly designed 4 ohm speaker might, but
the implication of the graph is vented speakers have very
low impedance in the bass and this is unaffected by crossover
design. As far as I know this is simply untrue, unless you are
taking severe liberties with what is "actual impedance".
🙂/sreten.
- Status
- Not open for further replies.