You might not like the explanation because it's not consistent with a view that suggests "energy storage is meaningless". Despite any perceived lack of elegance in my description, it does in fact explain why a driver operating close to its fundamental resonance needs an LCR to remove energy provided by the amplifier in the resonance band to keep stored energy from producing a spike in response at the edge of the driver's passband. "Stored energy is meaningless" simply doesn't cut it. And frankly, I expected a good deal more from you.
I'm sure any grade schooler could figure out that a 5 volt sine applied to a driver at resonance where impedance is hovering at 60 ohms is going to produce .4 watts of energy before losses. The same 5 volt sine applied just outside resonance where impedance is down around 4 ohms produces 15 times that amount of energy or 6.25 watts. Putting aside a phase angle that might be somewhere between 30 and 60 degrees - it should be obvious that a severe non linearity will result as we traverse across the two regions of operation. This is what is showing up in the response of the driver without an LCR. Why should any of this be complicated or difficult to understand?
Last edited:
64
Hi Joachim,
Germany will have to bare with you for at least 20 more years if you want to retire when you're 64 years old
atb,
Audiofanatic
Hi Joachim,
Germany will have to bare with you for at least 20 more years if you want to retire when you're 64 years old
atb,
Audiofanatic
I did tempest every thread when talking about Isobaric so I may report my experience .
As Joachim's target was to use medium-low cost drivers , using two lesser power woofers allows some flexibility . First ,if 8 Ohm speakers are used ,impedance halves . So we'll have to cope with an higher distortion coming from the amplifier ...which is then voiced . The air contained in the chamber may still act as a spring i.e. not coupling the two cones perfectly.
A positive factor ,other than the reduction of the 2nd harmonic distortion ,is that the air contained between the cones may work as a low-pass...I mean , I fell that the speaker cannot reach the mids as in a single arrangement .
Two Ciare HW159 in 20 l bass-reflex do-me-good
As Joachim's target was to use medium-low cost drivers , using two lesser power woofers allows some flexibility . First ,if 8 Ohm speakers are used ,impedance halves .
So we'll have to cope with an higher distortion coming from the amplifier ...which is then voiced . The air contained in the chamber may still act as a spring i.e. not coupling the two cones perfectly.A positive factor ,other than the reduction of the 2nd harmonic distortion ,is that the air contained between the cones may work as a low-pass...I mean , I fell that the speaker cannot reach the mids as in a single arrangement .
Two Ciare HW159 in 20 l bass-reflex do-me-good
Hello All,
I agree that energy storage is real and has an effect. I disagree that there is a peak in power output at resonance. For verification plug in your favorite test microphone or simulation software. At resonance there is a corresponding peak in impedance. Plug in the impedance value into the power formula for the assumed perfect voltage source amplifier and there is a corresponding dip in amplifier power output at resonance. Now look at the measured frequency sweep or simulation and typically the output is 3db down at resonance in a sealed box tuned to Q :=0.71. lower at Q :=0.50. It is no wonder that energy storage is ignored by some. I believe that energy storage at resonance is more about accuracy rather than a peak in output.
DT
All just for fun!
Last edited:
Not at all accurate. Why must you attribute everything to nonlinearity? The impedance peak is present in a perfectly linear system. It's like you observe that the impedance varies with frequency and thus conclude that the system it represents must be nonlinear.
I never suggested there was a peak power output at resonance. Energy stored is what I said was peaking at resonance. Actual acoustic output at resonance is a function of a large number of variables. Suffice to say that if the driver were properly loaded at resonance, stored energy would be low and energy dissipated either through acoustic output or other mechanical/electrical losses would be comparatively higher. I think I've been saying all along that energy storage is about control. The power aspect mentioned in my previous post was to point out the available energy (hence force) to exert control over diaphragm motion at the frequencies specified. It doesn't translate into acoustic watts output at those frequencies which also depends as noted on a number of other factors. Very simply, if the watts aren't there to control motion - you can't expect accuracy and low distortion. I've heard less informed folks make statements to the effect of "who cares - if the driver is resonating without the amplifier producing current it's fine because it means the system is more efficient". That kind of superficial perspective is common unfortunately in DIY circles. Others have suggested - "well it's just low bass frequencies and people can't tell the difference between 5% and 20% distortion anyway". To that I say two things:
1) speak for yourself as to what's audible and not
2) the gross non linearity produced by energy storage can have significant negative affects on acoustic output when you attempt to enforce a particular response in the resonance region with signal shaping components
You are living in a grossly oversimplified world. Not all distortion or non linearity is measured in a simple plot of an energy time curve or non linear distortion curve. The fact that the impedance curve is anything but a straight line should give you a clue about the potential for non linearity. It's not an exceedingly simple situation as there are many factors influencing what ultimately comes out in the way of acoustic output for a given input. You simply can't capture all the possibilities of non linear behavior that can occur with a simple plot of distortion when you excite a transducer with a swept sine wave at a fixed drive level across it's operating band. It's just one small part of the entire picture. If you think a CSD, linear distortion plot (ETC) or non linear distortion plot will cover all the bases, you're deluding yourself - especially if you intend to operate a transducer in the region of its fundamental resonance.
I've said my piece about you position. Comments like "The fact that the impedance curve is anything but a straight line should give you a clue about the potential for non linearity" should actually give a clue about just how misinformed you are. Tell you what,I'll post an impedance curve and you tell me if the system is linear of not. After all, you should have a clue. It is the kind of nonsense that could only influence the unenlightened and uneducated. Why you desire to continue to mislead and misinform I do not know, understand, nor at this point care about. It is just a shame.
Since you clearly think I'm making this stuff up as I go along, read pages 126 and 127 about the definition of LINEARITY in the context of signals and systems.
http://www.ee.sunysb.edu/~ctchen/media/freshlook.pdf
I think I have a pretty good handle on what linear and non linear means not because I'm overflowing with self confidence but because I'm repeating the information other individuals highly respected in the field have been saying. I actually have very little confidence in myself but quite a bit in people like Prof. Chen - someone I studied under.
Hello,
I am not opposed to what you are saying. I am trying to better understand your intent.
This is what I am looking at. At resonance there is increased or at least less controlled stored energy and decreased accuracy. You propose a LCR or tuned notch filter to limit the uncontrolled cone motion at resonance. The other side of the coin is to avoid a notch in the frequency response at the cost of increased distortion near and at the resonate frequency. On target? Apparently you have your choice among the trade offs.
A story from my world; we design air handling systems with glorified 100 plus HP inverters that produce “sin wave” output to control fan speed and air movement. Some of these fan systems develop Tesla like resonance that threaten to shake the building down or at least destroy ceilings and walls. I call it the pipe organ effect. Depending on the number of fan blades beating the air, the resonance of the ducts, building elements and and the RPM of the fan we end up locking out a narrow frequency range on the fan frequency drive to avoid breaking stuff.
DT
All just for fun!
I prefer transmission line loading instead of LCR notches if one is trying to maximize output at resonance. If one doesn't care about output at resonance and one is simply trying to reduce stored energy due to fundamental resonance - as noted earlier, one can't do it effectively with a single capacitor. An LCR trap is needed along with the cap. The cap limits current available and the LCR trap diverts what current is available away from the driver to prevent it from storing enough energy to "ruin" response at the lower edge of the passband.
The air handler is interesting - 100 hp, no less.
Sounds like you need to avoid the fourth harmonic of the fan's fundamental like the plague. This tends to establish the peak standing wave that wants to "shake the house down" 
In practice, the characteristics of the BL curve as the coil goes to the extremes makes a significant difference. Not knowing what kind of drivers you used during your research, it is possible that you can come to a different conclusion.
I have tried that, it is good, but the LCR trap is slightly better. So the decision is how much the improvement is worth.
Not if you think that reference proves any of your points. Write your Prof and ask him. Then post the answer.
If you are not comparing equal HP freq, and sensitivities then comparing box volumes is not valid.
This is partly true. There is still some electrical stored energy. However, the spectrum is more spread out.
As I have shown (I think earlier in this thread) two impedance curves with their original driver impedance curves, the change in impedance is only due to the different enclosure loading. One of the early articles published by Ted Jordan in Wireless World addressed various enclosures, and also shown an electrical equivalent models.
It would be "just the transient decay of the linear system" and leave it at that, if we were describing a closed system. But in a loudspeaker that "system" is coupled to air, where it radiates sound that we (outside "the system") hear. And the "ringing" (that we hear) is sound that was not in the original signal (that we are trying to reproduce). So we've got your "nothing nonlinear about it" black box with an electrical signal going in and an acoustic signal coming out and the acoustic signal has identifiable tones that the electrical signal does not. Further, if you take the ring of the bell, or the tone coming out of the black box, you cannot deconvolute it and re-create the original electrical signal. Somewhere in the box the nature of the signal that set the bell ringing has been lost.
I think that's "distortion" . . .
Last edited:
George, I think at the outset I referenced a high degree of loading via air load interaction as well as the damping (stuffing) of the line itself. These are the two main advantages of transmission lines over other vented and sealed alignments. If the proper length is used, the end result is the broadest and most thorough suppression of impedance associated with resonance of any of the standard alignments.
Yes, there is still electrical and mechanical stored energy but in a proper TL, the double hump of the bass reflex impedance curve essentially goes away when the proper amount of stuffing is added.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.