Motional and inductance impedance correction of loudspeaker
Sorry, i'm leaving this forum, and so will not be able to publish the promised article.
Sorry, i'm leaving this forum, and so will not be able to publish the promised article.
Last edited:
Hi Christophe,perhaps the following would be a start. Zobel network for parallel drive units - diyAudio. Compensating networks for a speakers fundamental resonance is something that I find unnecessary for bass speakers as the established routine of correct enclosure design with known speaker parameters and of course the amplifier damping factor has this aspect well covered. With tweeters,the xover network should be chosen to avoid the resonant frequency and ,anyway,trying to exactly match a a resonance with a compensating network is problematic,as the frequency is apt to vary with level,temperature and time. As an extra thought,it is always better to deal with (mechanical) resonances at their source and the wise course for any treatment is to err on a slightly under damped condition.By this,I am indicating that if there is an audible resonance with a speaker,adding a null filter at that freq. does not entirely solve the problem as a resonance can be excited by other signals.
First, thanks for your comment. it give an opportunity to answer at most of the usual questions and controversies on this subject.
The zobels are well know. But the usual simple way, based on Thiele & Small parameters L and R are not usually sufficient, because the Eddy currents makes the slope not linear.
I gonna explain how to design a compensating network near perfect up to X3 / X4 the upper range of any speaker. (important for passive filters)
Need one more resistance and one more cap. **
That is what is currently admitted...that i found untrue, both on listening experiences and measurements.
That is not my own discovery, but one of a very innovative and talented designer, François Delamare, from the "Maison du haut parleur" in Paris.
He had made to me, when he was designing the fantastic "Aeria system"* several demonstrations of it. All conclusive.
And i had continued to experiment this since this time in all my enclosure designs.. Both with active and passive filters.
I tried to explain why, amplifier side, it has a benefit to not have to deal with a non linear voltage/current ratio due to the resonance peak(s) of impedance. Whatever the kind of acoustical charge.
But it has, even with big cable's diameters with short lengths, low resistance passive filter inductances (don't forget those are in serial) , and ultra low output impedance amplifier a measurable impact.
I will try to find in my papers, some of the measurements i've made at the time (two decades ago). It is around two db level difference up to low medium. And you can hear the bass are obviously more separated, drier, 'faster'... in all circumstances.
As i said, it has a benefit too on the fundamental resonance of medium and treble speakers, as the filter will have a better linear response, because they see a constant impedance.
As always in this forum, i'm sure that it will be a lot of people to argue about this, without experimenting. I don't want to lose my time in argues any more here.
I will publish the method, free for anybody to experiment, and make their mind by themselves.
Free to others, blind believers in books and theory far from reality, to stay aside, with a less good sounding system.
Please, future contributors to this thread, let us concentrate on how (and not why), and reports about your results. Both in listening experiences and acoustic level measurements.
* MHP (la Maison du Haut-Parleur) : concepteur français de kits d'enceintes pour la HI-FI et le Home Cinéma...
** http://www.esperado.fr/images/temp/aeria_full_filterj.gif
Red: Zobel (The bass 47µf contain the zobel)
Green: Motional compensation.
Orange: Passive attenuator.
Yellow: Response curve correction.
Note: This filter is symmetrical. Bass filter 12db/oct, High pass filter: 18db/oct Cutting frequency:700Hz.
Please, be patient, and let me take some times to write this article.
Last edited:
I found that keeping the impedance constant (around 18 ohms) through the bass and lower midrange frequencies definitely helps my OTL work better with the (100db/w/m, 100 liter) Iron Lawbreakers. I have figured that this is because since the OTL sees a near resistive load here, the distortion structure is simplified and consistent through a range of frequencies which creates improved image and detail rendering. Incidentally my OTL design has a low output impedance similar to many solid state amplifiers, so response variations due to that are not an issue.
However, I did not use a parallel network to accomplish this, but actually employed an air core transformer also working as a bass driver series choke to translate the upper BR peak occurring in the 70 hz region to gain a couple db extra amplitude response in that region while removing the impedance peak entirely from the impedance characteristic, so virtually all power delivered from the amplifier to the bass driver is still translated into sound. At the same time, the air transformer function actually mutes the amplitude above 100 hz by up to 4 db, allowing a 101 db/w/m 8 ohm driver to work in a 100db/w/m 18 ohm speaker flat down to 50 hz with smoothly rolling off bass below that to 32 hz.
I decided that I preferred to allow the Iron Lawbreaker impedance to rise smoothly above 300 hz or so for because it lets my OTL amplifier sound several times more powerful on most program material than its 60W/channel value into 8 ohms which lets the World's Biggest Bookshelf Speakers approach live levels for many types of music. Of course, the OTL may be a bit challenged trying to do a 300W bass guitar solo at live levels, but it still does surprisingly well when pushed there - the OTL is clearly the best sounding amplifier I have ever heard in overload, considering that it flat tops just like a sand amp there with a test sine wave. The Iron Lawbreaker speaker/driver are JBL 2220A and Altec 288G-16, both alnico magnet.
However, I did not use a parallel network to accomplish this, but actually employed an air core transformer also working as a bass driver series choke to translate the upper BR peak occurring in the 70 hz region to gain a couple db extra amplitude response in that region while removing the impedance peak entirely from the impedance characteristic, so virtually all power delivered from the amplifier to the bass driver is still translated into sound. At the same time, the air transformer function actually mutes the amplitude above 100 hz by up to 4 db, allowing a 101 db/w/m 8 ohm driver to work in a 100db/w/m 18 ohm speaker flat down to 50 hz with smoothly rolling off bass below that to 32 hz.
I decided that I preferred to allow the Iron Lawbreaker impedance to rise smoothly above 300 hz or so for because it lets my OTL amplifier sound several times more powerful on most program material than its 60W/channel value into 8 ohms which lets the World's Biggest Bookshelf Speakers approach live levels for many types of music. Of course, the OTL may be a bit challenged trying to do a 300W bass guitar solo at live levels, but it still does surprisingly well when pushed there - the OTL is clearly the best sounding amplifier I have ever heard in overload, considering that it flat tops just like a sand amp there with a test sine wave. The Iron Lawbreaker speaker/driver are JBL 2220A and Altec 288G-16, both alnico magnet.
Last edited:
When I was at KEF we experimented quite a bit with conjugate networks and introduced the 104.2 with total conjugation (I believe Mike Gough was the primary engineer on that model).
It is not a great trick to come up with a total Zobel type network to deal with all features of a typical system impedance curve. It is more of a trick to keep it appropriately tuned in production, especially if you want to keep a very flat impedance curve, although small errors will still keep the impedance fairly flat.
I am not convinced that there is any great benefit to this. Certainly there is no universal benefit. Most amplifiers are quite happy to drive a loudspeaker load. A low source impedance (high so-called damping) will negate the effects of impedance variation. If your cables maintain a low effective impedance then this flat output voltage will be preserved. In terms of amplifier distortion, the lowest distortion with most amplifiers will come with a moderate inductive rise. That is, distortion will be lower with non flat impedance than it will be with flat impedance. This difference is fairly academic as the distortion levels are typically very low.
Some transistor amps are unhappy with high phase angle loads, expecially when there is a combination of low impedance and high phase angle. If current protection is part of the amplifier this is the when it is most likely to be triggered. Watching phase angle and the minimum real part of impedance might be worthwhile for some of those cases (and only when playing at very high levels).
That you have heard a demonstration of an effect, and benefit, does not mean anything outside of that particular case. You can not generalize that impedance conjugation is necessary, only that the one amplifier tested was unhappy with a variable load. (I'll give you the benefit of the doubt that no "placebo effect" was going on in this particular demonstration.)
Flat impedance, like extremly low distortion, flat phase and many other technical parameters, need to be placed into the category of "seems like a good thing but we haven't proven any real benefit".
David S.
It is not a great trick to come up with a total Zobel type network to deal with all features of a typical system impedance curve. It is more of a trick to keep it appropriately tuned in production, especially if you want to keep a very flat impedance curve, although small errors will still keep the impedance fairly flat.
I am not convinced that there is any great benefit to this. Certainly there is no universal benefit. Most amplifiers are quite happy to drive a loudspeaker load. A low source impedance (high so-called damping) will negate the effects of impedance variation. If your cables maintain a low effective impedance then this flat output voltage will be preserved. In terms of amplifier distortion, the lowest distortion with most amplifiers will come with a moderate inductive rise. That is, distortion will be lower with non flat impedance than it will be with flat impedance. This difference is fairly academic as the distortion levels are typically very low.
Some transistor amps are unhappy with high phase angle loads, expecially when there is a combination of low impedance and high phase angle. If current protection is part of the amplifier this is the when it is most likely to be triggered. Watching phase angle and the minimum real part of impedance might be worthwhile for some of those cases (and only when playing at very high levels).
That you have heard a demonstration of an effect, and benefit, does not mean anything outside of that particular case. You can not generalize that impedance conjugation is necessary, only that the one amplifier tested was unhappy with a variable load. (I'll give you the benefit of the doubt that no "placebo effect" was going on in this particular demonstration.)
Flat impedance, like extremly low distortion, flat phase and many other technical parameters, need to be placed into the category of "seems like a good thing but we haven't proven any real benefit".
David S.
Say, Speaker Dave - were you familiar with KEF's 'ACE' program? I'm using coconut charcoal (in sealed plastic containers with a desiccant) to 'expand' the Iron Lawbreaker's internal volume compliance based on their 'white paper'.
Btw, I once convinced a friend to move from Bose 901's to KEF 104's. I probably told him that he didn't want to listen to his sheetrock do the talking.
With identification by some of the audibility of higher order distortion products 80 or more db down from the overall signals in some cases, I tend to think such effects can be audible, except perhaps with amplifiers with extremely high feedback factors which some claim have sonic problems of their own. I know that I've owned several solid state amps that were well regarded in their times - Hafler DH220, Son of Ampzilla, Crown K1 and K2 and Dynaco Stereo 400, and they all sounded quite different from each other to me, except that perhaps the K1 and K2 were rather close to each other - it almost seems an open issue how much of that is load related.
Btw, I once convinced a friend to move from Bose 901's to KEF 104's. I probably told him that he didn't want to listen to his sheetrock do the talking.
With identification by some of the audibility of higher order distortion products 80 or more db down from the overall signals in some cases, I tend to think such effects can be audible, except perhaps with amplifiers with extremely high feedback factors which some claim have sonic problems of their own. I know that I've owned several solid state amps that were well regarded in their times - Hafler DH220, Son of Ampzilla, Crown K1 and K2 and Dynaco Stereo 400, and they all sounded quite different from each other to me, except that perhaps the K1 and K2 were rather close to each other - it almost seems an open issue how much of that is load related.
Last edited:
So, please, make your own tests to convince (or not) yourself.
You will certainly understand that i don't want this thread to be spoiled with "non believers" vs "believers" fights, like it happens recently in an other thread with pro and con Current feedback.
I'm happy to know you had worked at Kef. On my side, I had the luck to exchange a lot with Raymond Cooke, in the 70th, while i was in charge of the electro acoustic developpment in the R&D department of a big French company at this time. We were involved together to design monitors for ORTF.
Since the first time we talked together with my friend François Delamare, we had made more than 50 different enclosures, (around 10 on my side) closed or bass reflex, with big or little Loudspeakers diameters, cone, dome as well as horns, including my big studio monitors. Each time, the benefit was so OBVIOUS that both of us will never design a system without such compensation networks.
On my personal enclosures, [ sub bass with JBL 46cm +two ways columns made of 31cm+jbl driver charged by a spherical waves horn] i am on the way to achieve an active filter (modified DCX2496 based): The compensation network will surely be the only remaining thing from my actual passive filter.
Be sure i'm not a "believer". I make my mind in listening and measuring ( Brüel & Kjær), and, as a sound engineer too, i have copies of many of my own masters i know very well to figure out what is right and what is wrong.
Again, i ask respectfully all contributors to reports their EXPERIENCES with such networks, and not their "Believes". Thanks.
[edit] To be totaly honest and precise, i have no experience with the benefit of such networks with active filters,as all our previous work was with passive. I will report my experience comparing with and without when my set-up will be finiched.
Last edited:
We talked about it a bit over at the Classic Speaker pages. I have some curves there.
Failed stuffing experiment - Page 2 - Acoustic Research - The Classic Speaker Pages Discussion Forums
I wasn't involved in it but I did measure some later and it seemed to work pretty well. The volume increase of the charcoal was pretty good but it seemed difficult to effectively fill the cabinet.
Regards,
David
Hi, speaker dave,
First, i hope my previous answer does not look as disagreeable, it was not my intention. I hope you understood the reasons of my reaction: we are both passionated, and i believe we follow the same path, when something new look interesting: try-it, measure, try to understand why and figure out what benefit we can get with it.
If it is interesting to figure out the reason of some improvement we have experienced (and verified) it is non productive and discouraging to have to fight with "non believers" (thje subject is not to believe or not) , see what i mean ?
I just discover Kef has found a better damping material than fiberglass (i fear your measurements). This look very interesting.
On my side, i use a complementary trick: i glue walls of damping material in the middle between the enclosure walls. It divide the overall volume in 8 little volumes and increase a lot the damping, as it acts where the speed of the air particles are the fastest. If you can keep half of those little volume near close, at the back of the enclosure, it has a real benefit during high transients, and produce to the basses the same kind of benefit i found with motional impedance correction networks. Basses sound "faster", "lighter" and better damped.
First, i hope my previous answer does not look as disagreeable, it was not my intention. I hope you understood the reasons of my reaction: we are both passionated, and i believe we follow the same path, when something new look interesting: try-it, measure, try to understand why and figure out what benefit we can get with it.
If it is interesting to figure out the reason of some improvement we have experienced (and verified) it is non productive and discouraging to have to fight with "non believers" (thje subject is not to believe or not) , see what i mean ?
I just discover Kef has found a better damping material than fiberglass (i fear your measurements). This look very interesting.
On my side, i use a complementary trick: i glue walls of damping material in the middle between the enclosure walls. It divide the overall volume in 8 little volumes and increase a lot the damping, as it acts where the speed of the air particles are the fastest. If you can keep half of those little volume near close, at the back of the enclosure, it has a real benefit during high transients, and produce to the basses the same kind of benefit i found with motional impedance correction networks. Basses sound "faster", "lighter" and better damped.
I have also done your stuffing partitions approach. A good way to kill the standing waves without overstuffing the cabinet.
I'm not trying to hijack your thread and would like to see your approach and results to conjugate load matching.
Also nice to meet a former friend of Raymond's. He was a gentleman of vast experience and connections.
Regards,
David S.
I'm not trying to hijack your thread and would like to see your approach and results to conjugate load matching.
Also nice to meet a former friend of Raymond's. He was a gentleman of vast experience and connections.
Regards,
David S.
Last edited:
It will take some times for me to find back my old measurements (if they are still somewhere): No PC at that time. I have to join François too in order to see if he had conserved some datas, on his side. He is no more in the audio business and it is a pity.
That, indeed the first thing you thing when you remember Richard. It is a common quality of all the "gurus" i had met in my professional life. A friendly attitude, open mind, curiosity and modesty. Some kind of noble attitude probably due to this endless quest of Grail 🙂
I too have looked into AC, porosity is certainly important, akin to the difference in regular wool felt padding and needled, albiet much greater surface area. Issue I had was how long it will last. Having lived in two different coastal areas where humidity and salt air is a serious issue.
Thought I would have tested by now but still have lingering doubts in doing such. Would be a shame to design a system and discover you had to change out once a month because of.
Thought I would have tested by now but still have lingering doubts in doing such. Would be a shame to design a system and discover you had to change out once a month because of.
Don't see that this is necessarily required depending on the speaker's design. In fact, when I did my (closed box) HT speakers that use coconut charcoal, I used a high effectiveness desiccant that can be individually replaced for each speaker without disturbing the dusty charcoal. The idea is that the desiccant will maintain a near optimal low humidity since it appears to draw the water vapor more aggressively than the charcoal does, at least based on my research. This is the type of material that is used in multipane windows to keep humidity low between the panes.
Of course, a commercial speaker manufacture like KEF wouldn't resort to this, for obvious reasons, but a little maintenance by an audio hobbyist is not verboten.
Last edited:
Sorry but I don't think you understand what high humidity and salt air can do. If the salt air can eat at my hand tools it'll eat at the speakers and everything inside. Salt cannot be removed via heating to burn it out of the carbon, woud require distilled water to remove and rebaked to reactivate. Friends use Rid It in the cartons and replace once every month or two. They're filled to the brim with water when replaced and everyone here has AC, it's not an option. 🙂
Example are the normal screws and nails used for decking, Cannot use them near the coast, they are sold at the store and many ignor the warnings, discover their decks fall apart in less than 5 years due to the corrosion. Epoxy coat isn't bad, galvanized is worthless surface treatment, while stainless and brass are preferred. What's good on a boat are the only things that last here.
Example are the normal screws and nails used for decking, Cannot use them near the coast, they are sold at the store and many ignor the warnings, discover their decks fall apart in less than 5 years due to the corrosion. Epoxy coat isn't bad, galvanized is worthless surface treatment, while stainless and brass are preferred. What's good on a boat are the only things that last here.
Last edited:
Hi, Greebster -
I'm really glad I specified cement board siding on my house in northern Illinois, although the builders hated wearing out their saw blades and the cement 'dust' from cutting. I understand cement board is popular in Florida due to the weather you describe. A single coat of outdoor latex has held up nearly perfectly for 13 years on my cement board (except for some localized staining). Unfortunately, the contractor (and myself) also used cedar trim and railings which have had to be repainted several times. I also have had some good luck with covering the cedar railings with many layers of UV protected phenolic resin which have held up relatively well for years. It turns out that even polyurethane 'spar varnish' is too rigid and sensitive to UV rays to work nearly as well. With the phenolic, you just sand the scruffy irregularities and put on another couple layers.
Probably, in your environment, this approach wouldn't stand a chance with paper coned speakers unless they were closed box and the driver cones were 'sealed' against moisture intrusion.
I'm really glad I specified cement board siding on my house in northern Illinois, although the builders hated wearing out their saw blades and the cement 'dust' from cutting. I understand cement board is popular in Florida due to the weather you describe. A single coat of outdoor latex has held up nearly perfectly for 13 years on my cement board (except for some localized staining). Unfortunately, the contractor (and myself) also used cedar trim and railings which have had to be repainted several times. I also have had some good luck with covering the cedar railings with many layers of UV protected phenolic resin which have held up relatively well for years. It turns out that even polyurethane 'spar varnish' is too rigid and sensitive to UV rays to work nearly as well. With the phenolic, you just sand the scruffy irregularities and put on another couple layers.
Probably, in your environment, this approach wouldn't stand a chance with paper coned speakers unless they were closed box and the driver cones were 'sealed' against moisture intrusion.
Last edited:
Hehe they used a circular saw to cut it, now that's funny! They make a special scissors for cutting this stuff, no dust at all! Easy as pie and no burned up blades (or saws) to boot! 🙂
Most speakers (not all) I have developed so far have no Zobel network, and AFAIR only one speaker deals with the impedance peak in the bass. I seldom think it's useful to design a good filter, in many cases you can live fine without.
A Zobel network is useful in traditional 2-ways. It helps to keep the load constant under high excursion of the woofer (Le varies with excursion), therefore reducing some intermodulation. Instead you can use steep elliptical filters, which are not so prone to changes due to load variations. But they are not so popular.
With real 3-ways, such a Zobel is not really necessary, as the midrange does not face such high excursion.
Of course, unless you use a modern amp with low output impedance. That's what feedback is used for, make the amplifier independent from the load.
If you prefer high output impedance valve amps, then you have to linearize the speakers impedance.
Yes, we can live fine without. But all my experiences had proven, to my ears as well as measurements, that we live BETTER with.
It is like everything in high end audio: There is no precise frontier and each improvement even outside of what some consider as enough for audio (bandwidths, slew rates, distortion levels, noise levels, phase's linearity) provides a better listening experience.
You will notice that, in the filter i pointed to as an example, we used a symmetric filter. It looks weird. It provide too an audible transparency increase.
As an audio designer since 1970 as well as a sound engineer, i use the following method: Understand, try, design, build, listen, measure, understand.
With great suspiciousness about my listening ability and memory.
If something is correlated between measurement and listening experience, i try to better understand how our ears works. If something is not immediately measurable, i try to make more measurements to see where is this change i notice in an 'no doubt' way. I take something as 'effective' in only one condition: repeated experiences in various environements and times.
Those inductances compensations had brought effective listening improvement each time we (i'm not alone) had added-it in more than 50 different enclosures of all kind.
I just found a very web site about loudsepeakers:
Les menus du chapitre : Filtres
The author shares exactly the same point of view, and provide in-line tools to calculate zobels as well as motional impedance compensation. He provides too measurement results, and a big database of loudspeakers T&S parameters, so, i think my work to publish calculation methods will be near useless... or easier. 🙂
Last edited:
What I discovered using Zobel is the change in phase higher up when trying to match to that mid. Best results always resulted a more extended range where the phase tracks wider in bandwidth at the crossover point. When removed to reduce cost and we are talking some big low DCR coils here, became difficult as this phase tracking bandwidth was reduced and shifted off, most of the time resulting in changing the type of slope required to do properly.
- Status
- Not open for further replies.