Guys, Guys no need to call names and shove dirt around.... This is my current project and you won't believe the dirt that is flying around!
OxBlue Construction Camera: Husky Stadium Renovation Webcam Presented by OxBlue & Turner Construction Company
BTY the actual line I used to close the sale was "I think he doth protest too much." A bit of a misquote from Shakespeare but it made clear the guy who was complaining the loudest was try to cover his own mistakes!
OxBlue Construction Camera: Husky Stadium Renovation Webcam Presented by OxBlue & Turner Construction Company
BTY the actual line I used to close the sale was "I think he doth protest too much." A bit of a misquote from Shakespeare but it made clear the guy who was complaining the loudest was try to cover his own mistakes!
Peter Baxandall gave a demo at a London AES meeting in the early 1980s showing how effective a loudspeaker used as a microphone can be.
John Atkinson
Editor, Stereophile
Well, like I said before, this is not the best thread to discuss really successful speaker design, and I don't care what your educational background is. Like most speaker box designers, everybody has their personal opinion as to what is important, and they choose to address the parameters that they consider most important. AND they usually tell every one that they converse with, how awful the guy down the street is, a potential competitor, and makes such lousy designs. That is why we HAVE to leave the real opinions to audio reviewers, who may not always be the best judges in the world, but at least they try to be accurate in their assessments of audio that they did not design, themselves. They also have a better chance of being successful at it than most of us.
Last edited:
Eddy currents in the pole piece caused by the moving and energized VC. They will look exactly like energy transferred to the air.
John
This is covered in a substitution schematics that I have shown here several hours ago. The gyrator is bi-directional. Any change at mechanical/acoustical side reflects to electrical side and vice versa. The electro-mechanical analogy and substitution is well known.
So what? PMA and Thorsten, your 'parsing' with each other and me, as well, doesn't do much more than confuse people with less experience. That is the problem. Most here do not have the experience to do really SUCCESSFUL speaker design (including me) yet everybody and his brother has strong opinions about it, and the efforts of others.
John, I think even you will run out of gray hair trying to predate the first use of a speaker as a microphone.
Thorsten, you sometimes write on issues that are not my immediate aoe, and on those occasions it has happened that I gave you the benefit of the doubt, but here you so miss the boat that I get the feeling there never even was a boat.
Take a driver with an impedance peak 4 times it's Rdc, which is lowish. Good woofers may go up by even 10 times their lowest point. Driven at close to Xmax on a voltage source, current draw goes down near resonance, compensating for the lesser energy required at that point for a given excursion. Put this driver on a current source, and at the resonance frequency this source will try to maintain the same current over a much higher resistance. Result: voltage goes up, easy math by how much. Total energy fed into the driver will go up by the same amount. Kaboom, VC flattened on one end or cone crumpled.
Furthermore, please read my post again. As soon as you place a resistor in front of a driver, you get a mix of voltage and current drive. That's why I pointed out that in most passively filtered more-way systems, there is such a mix between current and voltage drive for mid and high. You may even use it to correct a downward sloping tweeter, as I pointed out.
I also mentioned that current drive does something with Qes which you don't generally want to see in woofers or any driver for that matter. It is impossible to design enclosures for drivers with a Q approaching infinity. Low Qms = bad speaker. You don't want to lower Qts by applying mechanical brakes, which is what you do by lowering Qms.
Lastly, a loudspeaker on a voltage source becomes part of the amplifier feedback loop through back EMF. A very good thing, certainly for woofers. Unless you believe feedback is a bad, bad thing.
vac
Last edited:
Scott, you do not seem to understand WHY I ran the test. It was to show that one loudspeaker source could drive another loudspeaker from across the room AND make it into a measurable LOW Z inverse voltage driver voltage to the amplifier, and subsequently create IIM distortion, IF the amp was not designed with low enough open loop output impedance. This voltage would NOT be directly related to the forward voltage driving the amp normally, and therefore would be considered an ERROR voltage that would have to actively be damped by the feedback in the amplifier. Some amps are better equipped to handle this error voltage than others. Matti Otala had just presented a paper on this very subject, so I did it to show that he was on the right track. Besides, it was fascinating to realize that the 2 K-horns were talking to each other.
When we were kids we used speakers only, no mics, no batteries, for field telephones, playing war games.
Back EMF is the result of Lenz Law. This is a purely electromagnetic phenomenon.Lenz's law - Wikipedia, the free encyclopediaThe mechanical-acoustical loads imposed on the moving elements attached to the voice coil, do affect the potentially measurable EMF both qualitatively and quantitatively but they are not the cause of it.
Regards
George
No, I hit the nail right on the head (pataphysically speaking). K-horns are probably more reciprocal than most of the speakers folks here use. Try taking two ordinary 3 ways and point them into the room like you would listen (not at eachother) and while playing one of them measure the power the other is able to deliver to a load.
Last edited:
*amateur* here
I have some practical experience with an amplifier with an output-impedance
of ~10 ohms (Le Monstre with different Fets)
Assuming a perfect current-source, the output-voltage is proportional to the
impedance of the loudspeaker.
It was quite obvious listening at lf, due to the impedance-peaks (basreflex)
the output-voltage went up producing LOUD bass.
Also when you attached the amp to the loudspeakers when an input-signal
was already present, a distinct click could be heard since the output-voltage
immediately went down.
Lowering distortion in the midrange by using current-drive has been mentioned here (including measurements) :
http://www.diyaudio.com/forums/multi-way/198408-raising-driver-qts-you-cant-tuna-fish-10.html#post2755619
regards,
Klaas
I have some practical experience with an amplifier with an output-impedance
of ~10 ohms (Le Monstre with different Fets)
Assuming a perfect current-source, the output-voltage is proportional to the
impedance of the loudspeaker.
It was quite obvious listening at lf, due to the impedance-peaks (basreflex)
the output-voltage went up producing LOUD bass.
Also when you attached the amp to the loudspeakers when an input-signal
was already present, a distinct click could be heard since the output-voltage
immediately went down.
Lowering distortion in the midrange by using current-drive has been mentioned here (including measurements) :
http://www.diyaudio.com/forums/multi-way/198408-raising-driver-qts-you-cant-tuna-fish-10.html#post2755619
regards,
Klaas
Hi,
To be honest, I used to give people who wrote complete nonsense the benefit of doubt when I was younger and even on occasion attempt to correct them and to teach them. Nowadays I am generally happy to let them continue in their error unless they the feel the need to contradict me directly with totally ludicrous stuff...
No. BL(x)
Plus, your scenario makes no sense in the first place.
And of course current drive does not change drivers excursion capabilities or it's thermal power handling or the power without clipping available from the Amplifier.
If I drive the speaker with current I must either pre-equalise (linkwitz transform can be used) the signal to attenuate the bass frequencies which would otherwise be boosted tenfold (20dB) otherwise or otherwise the mechanical problem (resonance) needs to be dealt with by mechanical solutions (like Eckmiller did).
So if I have my system with essentially the same small signal frequency response for current drive and voltage drive I get the same SPL and cone excusrsion switching between the states, except current will cancel power compression and Le(i) modulation distortion.
So your whole scenario makes only any sense to start with you took a given system, drove it in the lab near Xmax with a Voltage source amp and steady tones and then on the fly switched to current mode WITHOUT taking the neccessary measures to produce a sensibly working system. Of course on a complete moron would even attempt that.
BUT EVEN THEN nothing goes kaboom, unless you exceed the thermal power handling of the driver. In that case it smells bad and smokes, but still no Kaboom. So you can join Marvin the Martian in complaining about the lack of kaboom, but there it is. You may believe as you wish, but the facts happen to be what they are.
What is a "mix"? It's neither fish nor flesh.
If a Driver Designer decides to design their drivers so they need a certain source impedance other than zero but significantly less than say 100 times the voicecoil DCR it is not current drive and not "mixed drive".
I design all my speakers (those for personal use) to tolerate a typical open loop tube Amp, as I happen to like such Amplifiers. Does that mean I design them for current drive?
First, not much is impossible. And it is in fact quite possible to design enclosures for drivers with a high Qt. It was actually done in many old radios and the like... In fact, Bob Carver's "amazing" or whatever it was called speaker used drivers with a Qt of over 3 IIRC...
Well, now could you elaborate precisely WHY I would not want damp a mechanical resonance using a fairly linear mechanical system to do so and why it is actually a GOOD IDEA to use a highly non-linear electromagnetic system to do so?
This whole "Low Qm = bad" is by far greater hogwash than the whole "any form of negative feedback is bad" stuff.
Low Qm simply means that the mass spring system resonance in the driver is well damped. Would you buy a car without shock absorbers? Yet in a speaker the lack of these becomes a virtue!? Pleeeease.
I believe nothing particularly about feedback.
But I do know that normally it is not a good idea to deliberately add a significant non-linear elements into the feedback of a looped feedback system, unless of course you want to build a a distortion pedal, in which case a pair of anti-parallel diodes is a better choice than a speaker driver.
It gives more distortion that way.
Ciao T
To be honest, I used to give people who wrote complete nonsense the benefit of doubt when I was younger and even on occasion attempt to correct them and to teach them. Nowadays I am generally happy to let them continue in their error unless they the feel the need to contradict me directly with totally ludicrous stuff...
No. BL(x)
Plus, your scenario makes no sense in the first place.
And of course current drive does not change drivers excursion capabilities or it's thermal power handling or the power without clipping available from the Amplifier.
If I drive the speaker with current I must either pre-equalise (linkwitz transform can be used) the signal to attenuate the bass frequencies which would otherwise be boosted tenfold (20dB) otherwise or otherwise the mechanical problem (resonance) needs to be dealt with by mechanical solutions (like Eckmiller did).
So if I have my system with essentially the same small signal frequency response for current drive and voltage drive I get the same SPL and cone excusrsion switching between the states, except current will cancel power compression and Le(i) modulation distortion.
So your whole scenario makes only any sense to start with you took a given system, drove it in the lab near Xmax with a Voltage source amp and steady tones and then on the fly switched to current mode WITHOUT taking the neccessary measures to produce a sensibly working system. Of course on a complete moron would even attempt that.
BUT EVEN THEN nothing goes kaboom, unless you exceed the thermal power handling of the driver. In that case it smells bad and smokes, but still no Kaboom. So you can join Marvin the Martian in complaining about the lack of kaboom, but there it is. You may believe as you wish, but the facts happen to be what they are.
What is a "mix"? It's neither fish nor flesh.
If a Driver Designer decides to design their drivers so they need a certain source impedance other than zero but significantly less than say 100 times the voicecoil DCR it is not current drive and not "mixed drive".
I design all my speakers (those for personal use) to tolerate a typical open loop tube Amp, as I happen to like such Amplifiers. Does that mean I design them for current drive?
First, not much is impossible. And it is in fact quite possible to design enclosures for drivers with a high Qt. It was actually done in many old radios and the like... In fact, Bob Carver's "amazing" or whatever it was called speaker used drivers with a Qt of over 3 IIRC...
Well, now could you elaborate precisely WHY I would not want damp a mechanical resonance using a fairly linear mechanical system to do so and why it is actually a GOOD IDEA to use a highly non-linear electromagnetic system to do so?
This whole "Low Qm = bad" is by far greater hogwash than the whole "any form of negative feedback is bad" stuff.
Low Qm simply means that the mass spring system resonance in the driver is well damped. Would you buy a car without shock absorbers? Yet in a speaker the lack of these becomes a virtue!? Pleeeease.
I believe nothing particularly about feedback.
But I do know that normally it is not a good idea to deliberately add a significant non-linear elements into the feedback of a looped feedback system, unless of course you want to build a a distortion pedal, in which case a pair of anti-parallel diodes is a better choice than a speaker driver.
It gives more distortion that way.
Ciao T
I was planning to do a short piece on measuring loudspeaker frequency response by using reciprocity calibration. You just drive one one loudspeaker and use the other as a microphone. Then you switch the driven with the measured unit, and run the curves again. Divide the result by 2 and you have the response curve!
Oh yeah that is exactly how they calibrate measurement microphones!
Oh yeah that is exactly how they calibrate measurement microphones!
Hi,
That is what those fluffy sports socks where invented for. Stuff them in the ports.
At the moment I am trying to decide if I like my system better with the speakers connected to the 16 Ohm Tap (around 6R source Z) and socks or connected to the 4 Ohm Tap )around 1.5 Ohm source Z) of my Amp and without socks.
As the rest of the impedance is basically a flat line (1st order series crossover correctly implemented does that naturally) only the bass tonality changes. The 16 Ohm Tap is louder and I can play louder but I think I need more socks...
Well, as the late Steve Jobs used to tell people when they complained that his products did something unexpected if they did a certain thing: "Then just don't do that".
Nice data, thank you.
Seems the data goes about as far back as my experiments with current drive. I had access to the RFT development division anechoic chamber and got similar results, I rougly remember 12dB lower midrange 3rd HD...
With almost 20dB of feedback applied to the woofers and current feed on these Bass Distortion also was quite low and bass went ridish low...
So, it seems you can tune a fish after all...
Still, what worries me these days is getting all these different ...(x) from Mr. Klippel's poster improved in drivers I am going to use in products...
Ciao T
That is what those fluffy sports socks where invented for. Stuff them in the ports.
At the moment I am trying to decide if I like my system better with the speakers connected to the 16 Ohm Tap (around 6R source Z) and socks or connected to the 4 Ohm Tap )around 1.5 Ohm source Z) of my Amp and without socks.
As the rest of the impedance is basically a flat line (1st order series crossover correctly implemented does that naturally) only the bass tonality changes. The 16 Ohm Tap is louder and I can play louder but I think I need more socks...
Well, as the late Steve Jobs used to tell people when they complained that his products did something unexpected if they did a certain thing: "Then just don't do that".
Nice data, thank you.
Seems the data goes about as far back as my experiments with current drive. I had access to the RFT development division anechoic chamber and got similar results, I rougly remember 12dB lower midrange 3rd HD...
With almost 20dB of feedback applied to the woofers and current feed on these Bass Distortion also was quite low and bass went ridish low...
So, it seems you can tune a fish after all...
Still, what worries me these days is getting all these different ...(x) from Mr. Klippel's poster improved in drivers I am going to use in products...
Ciao T
The coupling issues are fraught with problems they must be exactly known. Also reciprocity calibration is not what you describe, it uses three microphones. Why would two identicle speakers not give the same response when switched?
http://bruel.ru/UserFiles/File/Review1_98.pdf
- Status
- Not open for further replies.