Well, I didn't bring up GD. And if it is acceleration of the driver is what you want then just look at it. The physics of the driver motion is pretty straight forward.
Franky I don't know what this discussion is supposed to be about at this point. Everything being discussed is pretty elementary. Seems to be a lot of he said, she said.
The other point I would address is the idea that these is a unique acoustic center. What is typically referred to is the AC based on the on axis response. Find that AC position and measure the same driver at 60 degrees off axis and the AC would be at the same position.
Further to my posts #129 and #139. At 1000 hz, an error in measurement of 500 microseconds or any whole number multiple of it will put the point at which the cone seems to be 180 degrees out of phase with where it actually is. At 5000 hz the window drops to 100 microseconds. To directly measure acceleration, a different device must be used, one which measures reflected laser light over a very short windows around the instant of the applied marker and just afterwards to determing if the dopplar shift is up or down and what change in velocity is occuring to determine the direction of acceleration. That would be the first time derivative of the cone's velocity. I used a similar laser velocimeter about 40 years ago integrating (rather than differentiating) the difference in dopplar upshift and downshift of two laser reflections from the same laser using a beam splitter prism where the beams converged on cold rolled sheet steel moving from 0 to 3500 fpm to determine its length so the technology is hardly new. The instrument was developed by GE in Roanoke for this purpose and was very accurate.
Sooundminded said:
In some cases that might be close enough to figure something out, but understand what is happening here;
Take a crossover or pspice make yourself a perfect minimum phase pretend speaker. Use say a 4th order low pass at 10K to make it clearer and a high pass at say 50Hz..
Examine the phase for this, note the two steps associated with each corner. Look at the GD for this, see that “Time = 0” so far as phase or GD is above the low pass corner at 10KHz.
In other words, the fly in your approach above is any high frequency pip that the speaker can produce, will be produced according to the speakers acoustic phase for those frequencies. That is what will govern where in the lower sine wave the pip happens.
The point I was making before it that in order for an identical response minimum phase loudspeaker to measure like what one predicts for what one has in a hard wired circuit (perfect minimum phase, no delays to speak of),
ALL the fixed delay must be removed AND for a band pass thingy, where Heyser’s Time=0 is, actually corresponds to what one would measure at a frequency well above what the device can produce.
Best,
Tom
getting the vibro - oscillating parts to stop doing so,
has been addressed by at least three manufacturers.
lowther and hartley by ferromagnetic coatings,
and iron foil by the 'exact sound graphic' speakers
```````````````````````````````````````````````````
if you don't mind my asking,
how many fuses were in this piece of equipment?
An externally hosted image should be here but it was not working when we last tested it.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
only if wired in series.
i have a clear recollection of my parents telling me,
'you cant have speakers stacked to the ceiling'
`````````````````````````````````````````````````````````
looks to me like he's selling mid-fi.
Last edited:
Tom, there seems to be some confusion here so I went back to the original posting that brought it up. It's posting #190. The posting referred to a magazine article and then said;
"for F >> resonance:
acceleration is in phase with applied voltage, amplitude is 180 deg out of phase
for F << resonance:
amplitude is in phase with applied voltage, acceleration is 180 deg out of phase"
Reading the article carefully, that is not what it said, at least I didn't read it that way. The measurement of the instantaneous relationship between applied voltage and acceleration and the relationship between applied voltage and amplitude (read that to mean displacement) is not one that can be made in an electronics laboratory but only in an optics laboratory using the kind of sophisticated equipment I described. In fact there is likely no correlation between them and that is unimportant. It is not germaine to the problem of speaker system design. But there seemed to be no end of challenges based on what I can only conclude is a superficial understanding of what was posted and then what was posted in response.
I'm a little curious about when the article was really originally written.
"John Watkinson wrote a short article for Electronics World about a dozen years ago"
It looks more like it was written in the 1950s or 1960s, its ultimate point was advocating an active filter to overcome loudspeaker deficiencies. In that era the cost of a 1/3 octave graphic equalizer such as Altec Acoustavoice was $900 per channel, far beyond the means of audiophiles, something you only found in the best recording studios for "calibrating" the FR of monitor loudspeakers used to master recordings. A dozen years ago active equalizers were available in $30Walkmans and had already been written off by the high end hi fi industry because they worked.
"for F >> resonance:
acceleration is in phase with applied voltage, amplitude is 180 deg out of phase
for F << resonance:
amplitude is in phase with applied voltage, acceleration is 180 deg out of phase"
Reading the article carefully, that is not what it said, at least I didn't read it that way. The measurement of the instantaneous relationship between applied voltage and acceleration and the relationship between applied voltage and amplitude (read that to mean displacement) is not one that can be made in an electronics laboratory but only in an optics laboratory using the kind of sophisticated equipment I described. In fact there is likely no correlation between them and that is unimportant. It is not germaine to the problem of speaker system design. But there seemed to be no end of challenges based on what I can only conclude is a superficial understanding of what was posted and then what was posted in response.
I'm a little curious about when the article was really originally written.
"John Watkinson wrote a short article for Electronics World about a dozen years ago"
It looks more like it was written in the 1950s or 1960s, its ultimate point was advocating an active filter to overcome loudspeaker deficiencies. In that era the cost of a 1/3 octave graphic equalizer such as Altec Acoustavoice was $900 per channel, far beyond the means of audiophiles, something you only found in the best recording studios for "calibrating" the FR of monitor loudspeakers used to master recordings. A dozen years ago active equalizers were available in $30Walkmans and had already been written off by the high end hi fi industry because they worked.
I'm not selling anything. At least not in this industry.
BTW, I've concluded that the term mid-fi was invented to deprecate other people's expensive equipment we don't like. I never use it myself. Using live unamplified sound from real musical instruments in real spaces where music is intended to be heard as my reference, I've concluded all that all the consumer audio industry has to offer so far is Lo-fi.
Tomtt asks about fuses.
Geeze, no grey hair back then umm.
Fuses??? Fuses?????? We dooneed no steeenking fuses, they are for girly men.
Don’t get the wrong idea, it looked a lot neater when it flew on the shuttle and had a metal can sealed on it. I spent many hours working on that guy now that I think about it. I designed and built the levitation sound source, the electronics that drove and controlled it, the 100 Amp furnace power modulator and temperature regulator (ran at 1650 C very hot!!) , a high temperature microphone and read out and some other minor instrumentation and minor mechanical design and fabrication.
Oh, and seriously, there are no fuses because nothing is going to fail and if it did , no one could fix it back there, that thing is in a can, in the rearmost white box, on the left corner. We had a 120 amp breaker feeding us, that’s all.
http://en.wikipedia.org/wiki/File:STS-7_Anik_C2_deployment.jpg
I Spent a lot of hours laying inside that payload carrier working on hard to get at stuff haha.
Most crucially, Sally Ride turned me on (well, flicked a switch on the control panel I built, starting the otherwise automated experiment ).
Don’t get the wrong idea like it was fun and games, leading to delivery, we had a pizza party when we had worked 90 days without a day off and then continued to press toward delivery. Crazy days.
Best,
Tom Danley
Geeze, no grey hair back then umm.
Fuses??? Fuses?????? We dooneed no steeenking fuses, they are for girly men.
Don’t get the wrong idea, it looked a lot neater when it flew on the shuttle and had a metal can sealed on it. I spent many hours working on that guy now that I think about it. I designed and built the levitation sound source, the electronics that drove and controlled it, the 100 Amp furnace power modulator and temperature regulator (ran at 1650 C very hot!!) , a high temperature microphone and read out and some other minor instrumentation and minor mechanical design and fabrication.
Oh, and seriously, there are no fuses because nothing is going to fail and if it did , no one could fix it back there, that thing is in a can, in the rearmost white box, on the left corner. We had a 120 amp breaker feeding us, that’s all.
http://en.wikipedia.org/wiki/File:STS-7_Anik_C2_deployment.jpg
I Spent a lot of hours laying inside that payload carrier working on hard to get at stuff haha.
Most crucially, Sally Ride turned me on (well, flicked a switch on the control panel I built, starting the otherwise automated experiment ).
Don’t get the wrong idea like it was fun and games, leading to delivery, we had a pizza party when we had worked 90 days without a day off and then continued to press toward delivery. Crazy days.
Best,
Tom Danley
Curious. Tom how did you run a 100 amp furnace and all the rest of that stuff on a 120 amp breaker? Was power 3 phase? Furnace never operated at full power?? Different voltage??? I don't know anything about power distribution on space shuttles.
If the discussions about steady state sinusoidal oscillation there is no need for either electronics or optics. Any SMD (spring, mass, damping) system behaves the same way. well below resonance only the spring has significants. Thus F = kx. If F= sin(wt) then x = const * sin(wt), U = w*const*cos(wT) , a =-w^2 * const * sin(wt), displacement is in phase with F and acceleration is out of phase. At resonance the equation of motion reduces to F = D * U so if F = sin(wt), U = const * sin(wt), x = -(const/w) * cos(wt) and a = w*const*cos(wt) and it is U that s in phase. Well above resonance only mass is significant and F = ma. If F = sin(wt) , a = const* sin(wt), U = -(const/w)* cos(wt) and x = -(const/w^2)* sin(wt). a is in phase x 180 out of phase.
That is the behavior of the driver motion in response to an applied sinusoidal force. Now decide what the polarity of the electrical connection is and F will either have the same phase as V or inverted phase.
Now if you are considering the response of any SMDsystem at rest (a = u = x = 0) to to an applied force that since there is at the instant the force is applied a jumps to a = F/m and a is in the same direction of F.
If you don't like that then attach an accelerometer the the driver an measure acceleration directly.
You do not need optics to measure behavior at frequencies in the audio range.
Sorry I didn't have the exact date handy when I made posting #119...was October 1998 issue, pages 847 & 848.
I posted the article because I thought it had some nice plots to compliment the verbal descriptions Tom Danley was making concerning the motion of a woofer.
I wasn't trying to make any bold claims, was just making some observations from the plotted data.
I thought the behavior of a mass/spring/damper system driven by a steady sinusoidal source was fairly well understood.
I can remember running Runge-Kutta simulations back in college...lots of water under the bridge since then.
Last edited:
When you introduce an energy conversion mechanism, in this case from electrical to mechanical energy and try to compare a dynamic parameter in one form of energy with another dynamic parameter in the other form of energy, that introduces an entirely different problem that is far more complex. I don't think they get it and I'm about to give up trying. It's just more than they can handle. I'm also tired of snide remarks and insults by people whose knowledge is just not sufficient to model anything beyond what they are accostomed to dealing with.
The only insults I've seen are coming from you, in your frustration that your absolute assertions are frequently full of technical holes and we inferior intellects are pointing them out.
The post that started the debate was pointing out the phase relationship between driving voltage and the three output measures of displacement, velocity and accelartion. You initially denied any phase shift or any vector properties: "The statement that amplitude and acceleration are out of phase makes no sense." and then started talking about highly significant phase shift:
"If so the phase relationship between driving voltage and resulting physical acceleration, velocity, and displacement is probably meaningless because it would be in the thousands of degrees at mid frequencies where the crossover between the tweeter and midrange or mid woofer occurs."
We model woofer box systems as high pass filters. The electro mechanical equivalent circuit of a sealed box is in the end a 2nd order high pass. As with any 2nd order high pass we will see 180 degree phase advance in the stop band (half that, 90 degrees at the corner of fs). We could either measure a woofer, measure or simulate an equivalent electrical model or, knowing it is a minimum phase system, compute the Hilbert transform. In all regards we would see the same thing.
The phase response of loudspeakers is only slightly complicated in that we observe them from a distance and through a medium of finite propagation speed. This air path delay is constant for the frequencies of interest and results in continuous phase rotation from low to high frequencies. All the modern measurement programs allow us to cancel out the propagation delay or specify the distance to the origin of sound, the acoustic center of the driver. Have no doubt that there is no delay in the commencement of sound propagation other than the phase delay related to the Hilbert Transform predicted phase response, tied to the acoustical response. If we want to measure the acceleration of the voice coil or use a laser probe, we would see that this is so.
The driver has no hidden delay mechanisms.
John K. stated the correct relationship between the parameters of displacement, velocity and acceleration:
X = real(exp(st)) = cos(wt)
U = dx/dt = real(s * exp(st)) = -w sin(wt) = w * cos(wt-Pi/2),
a = dU/dt = real ( s^2 *exp(st)) = -w^2 cos(wt)= w^2 * cos(wt - Pi)
With each being the derivative of the previous, then we pick up the 6dB per Octave tilt and 90 degree phase shift going from displacement to velocity and again from velocity to acceleration . The phase shift is in the -pi/2 and -pi components.
By this factor alone the "statement that amplitude and acceleration are out of phase (does) make sense." They are exactly 180 degrees out of phase. This, of course, has nothing to do with being a woofer/box, or even a high pass filter, but is the relationship between displacement and acceleration for a sinusoidally oscillating body (do I need to explain that to someone who likes to frequently quote Newton?).
I enjoy the discussions on these forums. There are a number of very bright people here, and we kick around some interesting topics and generally make sense of them in the end. When someone comes along with a tone of arrogance and superiority (combined with confusion about some fundamental principals), then he shouldn't be surprised by the push back. With a little more humility I'd guess that even you could learn something here.
David S.
The primary benefit that I’ve seen claimed for current drive is a substantial reduction of (some kinds of) driver induced distortion . . . the figure I commonly see is “10-20dB reduction”. The benefit is generally not had if there are reactive components between the amplifier and driver (current drive of a crossover is pointless), so all crossover and frequency response corrections must be done ”active”, before the amp.
What I have not seen is a substantial body of testing with actual drivers that demonstrates this reduction in distortion, or that with completed loudspeakers we can actually hear it. Driver distortion (of the kinds that current drive can correct, if such exist) is in many cases already below audibility, and it may be that, as with “linear phase” and “transient perfect”, the “distortions” eliminated either cannot be heard, or where they can be heard do not produce a listener preference for or against. It does provide an interesting opportunity for the UE, which by correcting for the changes in frequency response (that come with current drive) would permit both instrument and listening tests for the claimed reduction in driver distortion when current drive is applied.
How would a current drive amplifier handle a multiway loudspeaker system especially if there are no reactive components in a passive crossover network to prevent midrange and tweeter burnout? Wouldn't each driver need its own amplifier in that case?
To be honest I have never heard claims of 20dB reduction in distortion. What I have heard is that it eliminated the effects of VC heating since as Re increases the current is forced to remain constant. I have also heard that it has some affect on reducing distortion due to Le(x) effects. Perhaps that is where the 20dB figure comes from??? But, the thing I see with the CI approach is thermal runaway. With CV as the VC heats up the current is reduced and the power dissapated in the VC (heat generation) decrease. It is sort of a limit cycle type thing. If Re doubles, I decreases by a factor of 2 and the heat generaterd decreases by a factor of 2. VC heats up more slowly.
With CI, as the VC heats up and Re increases, the power dissapated in the VC increase linearly with Re. Double Re, I remains constant, heat generated doubles, VC now heats up faster. That just seem like a bad thing.
speaker dave posted some data showing significant reduction in 3rd harmonic distortion with increased source resistance.
http://www.diyaudio.com/forums/mult...er-qts-you-cant-tuna-fish-10.html#post2755619
Overlays of the data were posted here:
http://www.diyaudio.com/forums/mult...er-qts-you-cant-tuna-fish-10.html#post2755882
The only factor that I found that was significantly improved by high source impedance was distortion due to ferrite magnet hysteresis effects. (as would typically be treated with aluminum shorting rings).
David
Your tests were single swept tone harmonic distortion measurements only though weren't they ?
I'd expect the effects of Le(x) to show up very little if at all in single tone harmonic distortion testing.
The effect of Le(x) would be most prominent in two tone IMD distortion measurements where one tone is low enough in frequency to generate significant excursion (<200Hz) and the other tone was high enough in frequency that voice coil impedance was well up the curve caused by inductance. (>2Khz or more depending on driver)
Now your high tone is going to be amplitude modulated by the low tone even without the voice coil needing to leave its linear x-max region. If the inductance primarily increases on the in-stroke and decreases on the out stroke (this would be typical ?) it's going to be primarily 2nd order IMD products.
A single bass tone would have the excursion required but the driver would have no significant contribution to impedance from voice coil inductance at those low frequencies, so modulation of Le wouldn't distort the bass waveform, hence no increase in harmonic distortion over and above other distortion mechanisms.
Meanwhile a single upper midrange tone would be at a frequency where impedance varied significantly with excursion due to Le(x) but would not by itself generate the necessary excursion to cause any modulation of Le, hence no increase in harmonic distortion beyond that caused by other distortion mechanisms.
The simultaneous presence of fairly low and fairly high frequencies is required, with a reasonable amplitude of the low frequency.
From that it seems likely that a 3+ way system would be fairly immune to the distortion effects of Le(x) as you're unlikely to have a single driver producing both large excursions and high frequencies where Le is a significant contributor to impedance.
However it's a potential significant problem for both 2 way and full range designs.
A full range driver is obviously trying to produce bass, midrange and treble from the same driver, so the driver is subject to significant excursions, yet at the same time it's being operated a long way up its impedance curve, which could be 20-30 ohms or more by the time you get to high treble due to voice coil inductance.
The higher the reactive contribution to the impedance the greater the modulation effect of high frequencies from bass would be. Current drive could make a big improvement here, although an equally satisfactory and far simpler solution could be a copper shorting ring to dramatically lower and stabilise the inductance. Many full range drivers now have this, and this is probably why. (A side effect is a boost of high frequencies which may have otherwise been lacking, although some copper shorting ring full range drivers actually have an upwards sloping treble that then needs compensating, although doing so won't dilute the other benefits of lowering Le and stabilising Z)
The other case is a 2 way system - you're not trying to use it as high in frequency as a full range driver so impedance modulation is less within the operating range, however you're typically driving it with a passive crossover which is very sensitive to changes in load impedance near the crossover point.
Le(x) modulation may be enough even at crossover frequencies like 2-3Khz to modulate the driver impedance enough to cause the crossover to vary from underdamped to overdamped on every half cycle of bass. This is unlikely to sound good, and if its happening in the presence region it could be even more noticeable, perhaps as harshness.
A zobel network would swamp some of the impedance variation of the driver and thus improve it slightly over having a crossover optimised for the raw impedance of the driver, but it will still allow it to vary somewhat. Again, current drive would eliminate this effect, but now you're unable to use a passive network, once again perhaps a better solution is simply a copper shorting ring to lower and stabilise the impedance against variations with excursion.
Has anyone ever attempted to take measurements on a passive 2 way system to see if Le(x) variation from bass is modulating the Q of the low pass filter at the crossover frequency ? To find out it might be necessary to measure the drive current of the woofer, by sampling it with a small sensing resistor in the earth return leg, and looking at what happens around the crossover frequency with and without a bass signal present.
Last edited:
- Status
- Not open for further replies.