I expect that Pass article is what resurrected this thread--In fact, I just read it this morning.
I agree that less amplifier damping is not always to be feared, and often to be embraced. This high-Ro territory has already been well-trodden by our ancestors, and they addressed the "problem" wisely and well with better diaphragm/suspension damping, large boxes, aperiodic loading, and open baffles, which turn the apparent liability of high-Q into an asset that paid dividends in low-frequency pattern control, etc.
There are many who avow that open-baffle is the least compromised alignment for dynamic transducers. I don't argue--that's what I listen to. My 21" OB bass units have a Qts of about 1.3, and so require very little EQ down to 30Hz--I love 'em.
I believe LF pattern control is golden. Besides dipole, I've been exploring other means of LF pattern control, like cardioid or even 2nd-order gradient alignments. John Kreskovsky's NaO cardioid bass section is interesting. I suspect it's a fine place for a high-Q woofer, which would cut down on EQ and allow more stuffing, better taming the pipe resonances of the open-back box. (I've also been thinking about adding an exponentially tapered slot, ala Carlson, to the top wall to spread the resonance, but that's another story...)
So I often find myself looking at one or another of my drivers (like my AuraSound 1808s) and thinking, "Dang, I wish it was higher Q!) This is where I think high-Ro amps (or adjustable Ro, please!) might be just the thing--making fine drivers with unusably low-Q usable in alignments that utilize cancellation for LF pattern control.
Hi Steve,
Just guessing, but I think KYW's "leaky box" may refer to aperiodic loading.
You probably remember Goodmans' ARU (acoustical resistance units) and Dynaco's A25 & A10.
I believe resistive damping of air movement behind the cone can significantly lower system Q (though I'm sure it has its limits). It can also be used to good effect in Cardioid alignments, which interest me.
Konnichiwa,
With notable exceptions, like Hartleys and Eckmillers plus a number of Pro-Drivers destined for horn use.
Correctly implemented it will take energy from the resonant system (to which it is coupled by the air behind the cone) and convert it into heat.
Maybe "leaky box" was too cavalier a term. Shoudl we term it a box with opening that is at least the same surface was the driver or larger with a flow resistance fitted? The same applies in fact to a "acoustic swamp" (literal translation from german), aka heavily damped labyrinth.
Correctly implemented you end up with a system Q that is a small fraction of the drivers Qm. Of course, a driver with a reasonable and not excessively high Qm would not be a bad idea to start with.
NOT NECCESARILY. The Eckmiller driver uses a flow resistance build into the driver basket to accomplish that job despite having a low loss suspension. Usung a box with a very small chamber for the driver with an exit out of this box in multiple small louvres amounting to more than the drivers cone surface and filled with a suitable absorbent material can be used in a similar fashion.
All that is needed is a modicum of intelligence and you can a number of possible solution, assuming you have deconditioned yourself from teh prevailing orthodox view....
Sayonara
Steve Eddy said:
With notable exceptions, like Hartleys and Eckmillers plus a number of Pro-Drivers destined for horn use.
Steve Eddy said:
Correctly implemented it will take energy from the resonant system (to which it is coupled by the air behind the cone) and convert it into heat.
Steve Eddy said:
Maybe "leaky box" was too cavalier a term. Shoudl we term it a box with opening that is at least the same surface was the driver or larger with a flow resistance fitted? The same applies in fact to a "acoustic swamp" (literal translation from german), aka heavily damped labyrinth.
Steve Eddy said:
Correctly implemented you end up with a system Q that is a small fraction of the drivers Qm. Of course, a driver with a reasonable and not excessively high Qm would not be a bad idea to start with.
Steve Eddy said:
NOT NECCESARILY. The Eckmiller driver uses a flow resistance build into the driver basket to accomplish that job despite having a low loss suspension. Usung a box with a very small chamber for the driver with an exit out of this box in multiple small louvres amounting to more than the drivers cone surface and filled with a suitable absorbent material can be used in a similar fashion.
All that is needed is a modicum of intelligence and you can a number of possible solution, assuming you have deconditioned yourself from teh prevailing orthodox view....
Sayonara
Bill F. said:
Could be. All I know is he said "leaky box."
No, but I remember the "Variovent" which I beleive was claimed to provide some aperiodic loading.
Hmmm. Well lowering system Q doesn't necessarily mean that the system Q is lowered below that of the driver Q. Are you sure that the devices you're talking about can give a system Q lower than that of the driver's Q?
se
Kuei Yang Wang said:
Ok. Though if I understand what you're saying below correctly, it wouldn't take energy from the resonant system, but rather limits the amount of energy that can be stored in that resonant system.
Gotcha. Makes more sense now than just "leaky box." You're basically damping the driver's mechanical resonance by fitting it with a gas mask.
Yeah, I can see where implementation could be rather tricky. It would need to provide sufficient damping at resonance to bring the mechanical Q down, but you'd also want it to unload just right above resonance to prevent it from causing losses where you don't want them.
Ok. But the Eckmiller's basically bolting an enclosure onto the driver. I was thinking of how to address it in the driver itself.
Sure.
se
Eric wrote:
I think it is quite easy to see how it deals with that: It simply doesn't !!
You can restrict the maximum "loss of damping" by restricting the basic gain of the amp by introducing another feedback resistor from the putput to the inverting input.
You can go a little further and achieve full damping around resonance with a little more effort. If anyone is interested I can post a schematic (risking that someone is going to patent it).
Regards
Charles
I think it is quite easy to see how it deals with that: It simply doesn't !!
You can restrict the maximum "loss of damping" by restricting the basic gain of the amp by introducing another feedback resistor from the putput to the inverting input.
You can go a little further and achieve full damping around resonance with a little more effort. If anyone is interested I can post a schematic (risking that someone is going to patent it).
Regards
Charles
Charles,
please go ahead and post. My imagination is not up to speed this morning
I suppose you want to achieve voltage drive around resonance by paralleling the driver with some resonant electric circuit that has finite impedance around driver fs and hence does the current to voltage conversion?
This way, you would still have the benefit of reduced midband distortion from eddy currents etc.. However, you would loose the benefit of a more linear dependence on Bxl exactly where it matters, at high excursion.
Regards,
Eric
please go ahead and post. My imagination is not up to speed this morning
I suppose you want to achieve voltage drive around resonance by paralleling the driver with some resonant electric circuit that has finite impedance around driver fs and hence does the current to voltage conversion?
This way, you would still have the benefit of reduced midband distortion from eddy currents etc.. However, you would loose the benefit of a more linear dependence on Bxl exactly where it matters, at high excursion.
Regards,
Eric
Konnichiwa,
How is your German? The Eckmiller was designed in the late 1930's in Germany and used as Monitor for the first stereophonic music recordings to tape. It is a very interresting driver. I have some details in a File areea about German Vintage Gear in one of my yahoo groups. This stuff used to be open to all, nowadays you must be a subscribed member, sorry.
http://groups.yahoo.com/group/Thunderstone_technical/files/German%20Vintage/Eckmiller%20Koaxial/
Sayonara
enochRoot said:
How is your German? The Eckmiller was designed in the late 1930's in Germany and used as Monitor for the first stereophonic music recordings to tape. It is a very interresting driver. I have some details in a File areea about German Vintage Gear in one of my yahoo groups. This stuff used to be open to all, nowadays you must be a subscribed member, sorry.
http://groups.yahoo.com/group/Thunderstone_technical/files/German%20Vintage/Eckmiller%20Koaxial/
Sayonara
Since resonant systems are fairly predictable and easily modeled (and corrected, for the most part), I don't know why one would need anything as complex as motion feedback to control damping around resonance. As the system is a combination of the electrical and mechanical, Q adjustments in the electrical realm are nearly as effective as Q adjustments in the mechanical realm (with some obvious inability to stop single cycle input ringing, which may not be that critical for real music content).
Appropriate filtering of the signal before amplification should be able to transform the system Q into something more desireable without resorting to high loss mechanical damping and the resultant decrease in sensitivity. A biquad filter seems perfectly suited to this role.
Considering all the pros/cons, I think an "ultimate" solution would probably lean towards pure mechanical damping with high power to compensate (is this easily doable with transconductance amplifiers?). Sure, it would be terribly inefficient, but who said the best absolute quality should (or more importantly could) maximize all possible performance characteristics? Typically, you have to sacrifice something to gain something, and if increased heat generation and amplifier power requirements are the penalties for lowered distortion, I think that is an equitable trade for some applications.
Appropriate filtering of the signal before amplification should be able to transform the system Q into something more desireable without resorting to high loss mechanical damping and the resultant decrease in sensitivity. A biquad filter seems perfectly suited to this role.
Considering all the pros/cons, I think an "ultimate" solution would probably lean towards pure mechanical damping with high power to compensate (is this easily doable with transconductance amplifiers?). Sure, it would be terribly inefficient, but who said the best absolute quality should (or more importantly could) maximize all possible performance characteristics? Typically, you have to sacrifice something to gain something, and if increased heat generation and amplifier power requirements are the penalties for lowered distortion, I think that is an equitable trade for some applications.
RHosch said:
I don't see why you couldn't ultimately ahceive the damping you need to bring down the Q at resonance without losing efficiency elsewhere if the gas mask approach (i.e. the Eckmiller driver Thorsten mentions) is done right.
Seems to me that if you have the right combination of loss (to provide the appropriate amount of damping) and enclosure volume (so that the loss mechanism effectively unloads above resonance) you could achieve the more techincally ideal solution.
se
Konnichiwa,
Absolutely right. Funny that this is how they made high sensitivity speakers in the 1930's, 40's and 50's, which they buiild with suitable damping in the driver and the enclosure and then drove with SE Class A Pentodes (current sources). If you need good examples look into many an old radio (untill they started throwing tons of feedack in and separating the speakers from the amplifiers).
Interesting that Nelson Pass's approach is basically the same thing again 50 or 60 years later. I guess what goes around comes around....
Sayonara
Steve Eddy said:
Absolutely right. Funny that this is how they made high sensitivity speakers in the 1930's, 40's and 50's, which they buiild with suitable damping in the driver and the enclosure and then drove with SE Class A Pentodes (current sources). If you need good examples look into many an old radio (untill they started throwing tons of feedack in and separating the speakers from the amplifiers).
Interesting that Nelson Pass's approach is basically the same thing again 50 or 60 years later. I guess what goes around comes around....
Sayonara
Nelson Pass said:I recall stating the same, but it seems to need refreshing in
today's market mentality and given a new crop of drivers to
play with.
But all of those drivers have rather high mechanical Qs. And from what I read in the article, this wasn't addressed at all. So it's not quite the same as what Thorsten's talking about, which is reducing mechanical Q to avoid the decidedly underdamped bass response that's evident in pretty much all of the plots in the article.
se
Steve Eddy said:
This is addressed in the text, where it is made clear that you
can set the damping as you like. In listening tests, most people
preferred a Q around 1 or so. If you want a lower Q, that's
very easy to achieve, just adjust the source impedance. Since
most of the case shown were around 47 ohms, you have a lot
of room to play.
And let's look at the reverse case - driven by a pure voltage
source, these drivers suffered from extremely low Q in other
than rear loaded enclosures, with loss of bottom end as much
as 15 dB. Take your pick.