Why would you need some sort of motional feedback in a frequency domain where transducers will see a resistive acoustic impedance and cone excursions are generally very low? And a lot of them tend to have cone breakups in this range too.
since I don't master the English language I am forced to use the translator, the right sentence is that of JMFahey.
thank you all for the comprehensive answers
I think I was at that lecture, in London. Was it given by an engineer from Tannoy?
If so, the most memorable part was when severval AES member stepped up after, some with overhead slides ready, to show how they had done the same thing before over the previous couple of decades, and the problems they encountered. One of them was David Birt from the BBC Design & Equipment Dept. of SuperBal fame.
If it was the Tannoy one, I heard they had extensive early trouble in production with ironing out stability due to the considerable tollerances involved. Successful in the end, and being Tannoy, most likey a great product.
Edit: what I am saying is, it appears to be non-trivial even to the old and wise.
No idea, I only read the article. I think it was from the 1980's. I'll look it up when I have access to my attic again.
Member Soundbloke referred to a relevant Birt publication some years ago in another MFB thread, but unfortunately this was a JAES non-free (40 $ or so) article.
David Birt was a lovely guy. Bigger than life charater, with big silver grey sideburns, and shirts with large embriodery dragons on the back. That was in office hours. Outside it was BIG motor bikes.
Always had time for you with a big smile.
Always had time for you with a big smile.
Hello OP, that's the answer, from drewmc. But let me fill in a few foundational points that may have gotten lost in the abstract chat above.
1. Really stupid that woofers have their basic resonance within the music passband - which means that woofers have enormous distortion and other inadequate performance.
2. Like in the remainder of the known engineering universe, that can be addressed with corrective feedback. The benefits of diddling with the signal to the driver without feedback are trivial compared to feedback control.
3. The following points apply to woofers. When a researcher looks at a driver, they tend to think of it as full-range and always conclude feedback is great for low frequencies but not higher frequencies. So we are left wondering how things would work if they studied a purpose-built mid-range driver with feedback. Might be nice.
4. With a Rice-Kellogg woofer driver like used everywhere today, not feasible to conjure feedback except by sensing motion of the cone. That might be thought of as just sensing the behaviour of the dust cap - but that sure beats having no feedback. A few ways to do it, using an accelerometer glued to the dust cap is the leading DIY approach today (but I favour sensing the voice-coil with a Wheatstone Bridge for simplicity). Generations of DIYers have tried it and, when it works, the woofing improvements are almost always reported as exceedingly wonderful. But for a few reasons, impractical for commercial exploitation.
5. As others have pointed out, you can't just take a dual voice winding woofer and use one of the windings for feedback because of inductance from the other winding and low voltage output from the sensor winding. But as drewmc points out, you can make a higher voltage multi-turn coil as Rhythmic, Sony, and others have done in the past with good results.
B.
Last edited:
The Wheatstone Bridge design is an upgrade of current sensing. If you read the paper shown below, you will know just about everything you need to know about motional feedback, mysterious negative output impedance, and the Wheatstone Bridge method. OK to skip math at the beginning. Sad how little has changed since 1958.
Current sensing is simply when you use a resistor (say, .5 Ohm) in the ground return of the driver. Current sensing works great too because it testifies to back-EMF from from voice coil which reflects cone motion quite well.
The subwoofer forum has had many extensive threads on motional feedback.
Attachments
Last edited:
I looked it up, it was this article:
P. G. L. Mills and M. O. J. Hawksford, "Distortion reduction in moving-coil loudspeaker systems using current-drive technology", Journal of the Audio Engineering Society, vol. 37, no. 3, March 1989, pages 129...148. Paul Mills was appointed at Tannoy in July 1988, so the lecture and the article were probably about the same design.
There is a follow-up article from the same authors in the October 1989 issue, which mainly shows how complicated class-(A)B transconductance amplifiers can get when you want to ground one side of the load.
By the way, the first article mentions that motional feedback has been used as early as 1924: P. G. A. H. Voight, Improvements in or relating to thermionic amplifying circuits for telephony, UK Patent 231972, January 1924.
There is also no motion anymore at those frequencies.
(well technically there is, but not on the same way like you said)
The distortion comes either from non-linear behavior from the inductance. Le(x)
Which can be very easily tackled with a demodulation/shorten ring.
Plus it comes from the non-linear behavior in the magnet system (eddy currents).
I think that can be compensated with just current feedback?
Well or there are some issues with just acoustic mechanical problems.
But non of those techniques can solve those.
Last edited:
Two relatively new advances could push the use of motion feedback more into the mainstream. First, most modern magnet systems no include the shorting rings/ sleeves that reduce the voice coil inductance. There will also reduce the inductance of a pickup coil winding and allow stable feedback to higher frequencies. Higher than the breakup point of large woofers. The other trend is MEMS accelerometers. The small piezo sensor used on early servo woofers has poor response at low frequencies. The cell phone has brought us small, cheap, low noise accelerometers that could be used for motion feedback.
Another big factor is self powered speakers. If the amp is in the speaker system, it is not a big deal to add motion feedback. It may make more economic sense to correct a cheap woofer then pay the price of a good one.
Another big factor is self powered speakers. If the amp is in the speaker system, it is not a big deal to add motion feedback. It may make more economic sense to correct a cheap woofer then pay the price of a good one.
Thanks for link. Yes, seems to be Werner's paper.
He was at the once-great RCA lab (when Olson was there) and published more or less the same paper three times (counting this one too).
There are modern academic studies. But still hard to beat this simple demo with clear test results presented.
While the Wheatstone Bridge is an enhancement of the current sensor method, I couldn't say how much it helps.
He was at the once-great RCA lab (when Olson was there) and published more or less the same paper three times (counting this one too).
There are modern academic studies. But still hard to beat this simple demo with clear test results presented.
While the Wheatstone Bridge is an enhancement of the current sensor method, I couldn't say how much it helps.
Since some low frequency non-pistonic breakup is not hugely SPL-related, has anyone tried an "identical" speaker/driver as motional feedback? Not just voice coil feedback. The speakers should be boresight-facing each other. When both speakers are active, the push-pull arrangement reduces some nonlinear distortion. The question - when only one speaker is driven and the other is used as motion feedback, what happens? There must be ways to keep the combo from becoming a howling banshee.
Nothing complicated at all. Improved holland current pump circuit is simple for grounded load. I use such amplifiers, nothing complicated.