Current drive and bass boost.

[GK]

I was flicking through a 1959 copy of Radio News' Hi-Hi Annual the other day and came across an article on loudspeaker damping. The article detailed an experiment in de-damping an amplifier that, utilising global negative feedback, has a high damping factor, by connecting a fixed resistance equal to the nominal speaker impedance in series with the output.
Comparative frequency response plots are provided and it is shown that while the overall amplitude is reduced in the de-damped case, the low-end bass response is extended around the speaker’s resonance. No other measurements beyond the frequency response plots are provided and the author makes no subjective assessment as part of the experiment, but nevertheless, ends on an enthusiastic note about the technique for improving bass extension.

Now I know that this was written in 1959, but it struck me as this is something that is still advocated today, although mostly amongst proponents of full range drivers, which have a limited bass response. The tube heads think this justifies the high output impedance of their SET amplifiers.

Now can someone please explain to me how this concept of “current drive” isn’t a total kludge at improving bass response? I mean, the bass response of many speakers can definitely be extended here (but mostly only marginally), and a FR plot will show it, but ‘more’ bass doesn’t necessarily mean better bass, IMO.

What about stored energy, the loss of damping factor and speaker cone control? The benefits of a fast cumulative decay no longer matter? I’ve never owned a speaker who’s overall bass performance improved with a deliberate and excessive reduction in damping factor. An artificially boomy bass with a singing resonance isn’t more ‘Hi-Fi’ to my ears, and certainly isn’t to any objective standard.

A high-Zout amp seems rather sub-optimal option, in comparison to active equalization used in conjunction with a low z-out amplifier having adequate dynamic headroom.


Right?

[syn08]

That's an interesting question. I was thinking about transconductance power amps (driving current into a speaker load) some time ago, but never found any good pro/cons data.

Theoretically, the speaker electrodynamic (as in any motor) is current related, so current driving a speaker would make sense, but then I don't know about any relations with the acoustic properties of the entire setup.

[andy_c]

Now can someone please explain to me how this concept of “current drive” isn’t a total kludge at improving bass response? I mean, the bass response of many speakers can definitely be extended here (but mostly only marginally), and a FR plot will show it, but ‘more’ bass doesn’t necessarily mean better bass, IMO.

I'm pretty sure the current drive idea can be made to work well, but one would need a driver specifically designed for that.  If it isn't, strange things happen.

Below is a plot of an LTspice simulation of the subwoofer system that Pete and I were discussing earlier.  It's a Dayton (Parts Express) Titanic 15" sub in a 3 ft³ enclosure.  The plot represents the standard dB SPL vs. frequency at 1m distance with 2.83VRMS drive.  I've shown both current and voltage drive, and I've picked the current amplitude so that IRMS*RE=2.83 (where RE=voice coil DC resistance).  You can see that there's quite a bit of peaking when driven from a current source.

[ Guests cannot view attachments ]

To make the driver work properly with a current source, you'd need higher mechanical damping.  I don't know whether the end result would be better, worse, or a wash compared to the voltage drive case.  The LTspice simulation is attached below.

[GK]

That’s what I was thinking about mechanical damping, but such a requirement doesn’t gel well with speaker efficiency, for one.

That FR plot of yours sure shows the problem but is an extreme case though, as a lot of so-called “current drive” schemes really are only a poor approximation. A SET amplifier with a damping factor of 1 or 2 in reality can hardly be declared a current-drive amplifier. Same for a low damping factor SS amplifier.

The thing is, I just can’t see how electrically decoupling any driver in such a way to achieve a greater output at resonance can be done without compromising the transient response to some degree.

edit:
I'll scan and post that 1959 article this evening.

[andy_c]

That FR plot of yours sure shows the problem but is an extreme case though, as a lot of so-called “current drive” schemes really are only a poor approximation. A SET amplifier with a damping factor of 1 or 2 in reality can hardly be declared a current-drive amplifier. Same for a low damping factor SS amplifier.

Okay, I see what you're getting at.  Here's a more sane example.  I chose an infinite baffle design for this driver by just making the box volume 3e6 ft³ instead of 3 ft³.  With an amp having zero output impedance, this gives a high-pass function that's overdamped, having a Q of 0.383 (the driver's Q_TS).  Then I used a MathCad worksheet to adjust the amp's simulated output impedance to give a system Q of 0.707.  This turned out to be 3.54 Ohms.  I made a new simulation, and adjusted the open-circuit voltage of the source to match outputs at high frequencies.  This required a 5.85 dB increase in the open-circuit voltage relative to the zero Ohm case.  Here's the result.

[ Guests cannot view attachments ]

Now the improved bass extension can be seen.  Great new application for SET amps .

Anyway, the new simulation is below.

[GK]

The buzz words are “system synergy”. 

[GK]

Scanned the 2-page article but the full pages are not readable when shrunk to <=200kb. If anyone would find these pages exciting, email me.


 

[andy_c]

The buzz words are “system synergy”. 

  Or maybe the power paradigm.

"Background of Relatedness"

[GK]

 

Interesting that you mentioned an intinite baffle; that's what is mentioned in that article snippet I attached to my previous post. An overly damped bass driver is the is the only application that a quasi current (high impedance) drive would really make sense.

[andy_c]

Looks like nowadays they have drivers meant specifically for infinite-baffle applications.  Here's one with Q_TS=0.59 - half way between critically damped and Butterworth in free air.

The one from the previous simulation had Q_TS=0.38.