Hmm ... quality of auto/transformers for horn MF/HF drivers ... good question. Since these are typically intended to be placed after the highpass filter of the crossover, there's no need for the core to be very large - and this has the advantage of reducing stray C's in the windings, moving up the highest frequency of the transformer. Small is good in this application. Bud Purvine, Dave Slagle, et al are the experts in winding techniques to optimize HF response to minimize ripple and group-delay errors.
Nothing wrong with the old Klipsch autoformers, but remember they were used with PA-style drivers that weren't that all flat to begin with, so minor ripples in the response wouldn't matter that much. In addition, the Electro-Voice T35 or T350 tweeters didn't go much beyond 15 kHz, so the ultimate HF extension of the autoformer wasn't that critical.
What I don't like about autoformers are the coarseness of the taps, while recognizing that tapped windings are a real chore to do, with all those leads coming out in a small space. What I do like is that it buffers the wild impedance swings of the horn/driver combination, so the passive crossover can see a near-resistive load.
Amplifier -> Passive Lowpass Filter -> 9~10 ohm shunt to ground -> transformer primary -> transformer secondary -> optional Zobel inductance correction -> driver.
The transformer multiplies the impedance of the driver/horn by the square of the turns ratio, and the 9~10 ohm shunt to ground flattens out most of the reactance, leaving a load (as seen from the lowpass filter) that is close to resistive (8~9 ohms). From the perspective of the driver/horn, it sees a source impedance that is divided down by the square of the turns ratio - basically, a fraction of an ohm. Most of the reactance of the driver/horn is buffered from the passive crossover, which sees a load close to resistive.
The alternative is to use an array of notch filters as part of the highpass filter, but if the impedance curve of the driver/horn is complex enough, you might need to have many notch filters, not just one - and tuning of the array of notch filters might be quite critical, requiring hand-selected selected capacitors to attain the required precision. I think I saw a modern JBL highpass filter that had three notch filters, for example. Add-in JBL's "charge-coupled" method (DC polarization of caps using a battery) and you can end up with a surprising number of parts in the crossover.
The really old-school approach is to simply ignore the bumps and ripples in the driver/horn impedance - but then, unwanted crossover interactions are a certainty. If you're going to go old-school, the even-order family of crossovers, particularly 12 dB/octave, probably have the least unwanted interaction. I'm guessing that's the reason that they're commonly seen in horn/driver crossovers dating from the Thirties, Forties, Fifties, and Sixties.
The Klipsch autoformer technique is quite old, going back to at least 1959. In original form it splits the highpass filter in two, with the series cap in front of the autoformer, and the shunt inductor in back of the autoformer. I haven't seen autoformers in Altecs or JBLs, though - they stayed with L-pads right up to the advent of multi-amped systems.
Nothing wrong with the old Klipsch autoformers, but remember they were used with PA-style drivers that weren't that all flat to begin with, so minor ripples in the response wouldn't matter that much. In addition, the Electro-Voice T35 or T350 tweeters didn't go much beyond 15 kHz, so the ultimate HF extension of the autoformer wasn't that critical.
What I don't like about autoformers are the coarseness of the taps, while recognizing that tapped windings are a real chore to do, with all those leads coming out in a small space. What I do like is that it buffers the wild impedance swings of the horn/driver combination, so the passive crossover can see a near-resistive load.
Amplifier -> Passive Lowpass Filter -> 9~10 ohm shunt to ground -> transformer primary -> transformer secondary -> optional Zobel inductance correction -> driver.
The transformer multiplies the impedance of the driver/horn by the square of the turns ratio, and the 9~10 ohm shunt to ground flattens out most of the reactance, leaving a load (as seen from the lowpass filter) that is close to resistive (8~9 ohms). From the perspective of the driver/horn, it sees a source impedance that is divided down by the square of the turns ratio - basically, a fraction of an ohm. Most of the reactance of the driver/horn is buffered from the passive crossover, which sees a load close to resistive.
The alternative is to use an array of notch filters as part of the highpass filter, but if the impedance curve of the driver/horn is complex enough, you might need to have many notch filters, not just one - and tuning of the array of notch filters might be quite critical, requiring hand-selected selected capacitors to attain the required precision. I think I saw a modern JBL highpass filter that had three notch filters, for example. Add-in JBL's "charge-coupled" method (DC polarization of caps using a battery) and you can end up with a surprising number of parts in the crossover.
The really old-school approach is to simply ignore the bumps and ripples in the driver/horn impedance - but then, unwanted crossover interactions are a certainty. If you're going to go old-school, the even-order family of crossovers, particularly 12 dB/octave, probably have the least unwanted interaction. I'm guessing that's the reason that they're commonly seen in horn/driver crossovers dating from the Thirties, Forties, Fifties, and Sixties.
The Klipsch autoformer technique is quite old, going back to at least 1959. In original form it splits the highpass filter in two, with the series cap in front of the autoformer, and the shunt inductor in back of the autoformer. I haven't seen autoformers in Altecs or JBLs, though - they stayed with L-pads right up to the advent of multi-amped systems.
I should add if you own a vintage Altec or JBL loudspeaker, the original L-pads can get badly corroded, especially in high-smog cities like Los Angeles (lots of ozone in the air). Corrosion ain't hifi - if the pot crackles as you rotate it, or has dead spots, don't clean it, throw it away and solder in a new L-pad. Makes a big difference.
If you've found a level-setting you like, you can measure the two legs of the L-pad and substitute discrete resistors. This eliminates the distortion from the mechanical contact of the wiper - and that is real enough, I've measured nasty-looking low-level distortion from mechanical contacts back when I was at Audionics. It only takes a small, nearly-invisible film of corrosion to create real trouble in the millivolt region.
If you've found a level-setting you like, you can measure the two legs of the L-pad and substitute discrete resistors. This eliminates the distortion from the mechanical contact of the wiper - and that is real enough, I've measured nasty-looking low-level distortion from mechanical contacts back when I was at Audionics. It only takes a small, nearly-invisible film of corrosion to create real trouble in the millivolt region.
Yes and no, Earl
For the yes-part I do have to apologise - seriousely!
As for the no-part its just not the appropriate way you (predictable) approach and qualify things.
Why not do a simple change in habits (from time to time
) and get interested by phenomenas - rather than theory - in the *first* place?
Michael
For the yes-part I do have to apologise - seriousely!
As for the no-part its just not the appropriate way you (predictable) approach and qualify things.
Why not do a simple change in habits (from time to time
) and get interested by phenomenas - rather than theory - in the *first* place?Michael
Lynn's comment about the coarse taps on autotransfomers made me think of a possible solution, well I think the solution would work on transformers with a primary and secondary but I'm not sure about autotransformers. For instance, if the taps on the secondary are 3dB per steps, have a set of fine taps on the primary that have .5dB steps. You would have coarse adjustments on the secondary and fine adjustment on the primary. This could cover a wide range of attenuation while providing small steps.
I have an old Teletype power supply that has the power transformer configured in this manner.
Bud? Dave?
I have an old Teletype power supply that has the power transformer configured in this manner.
Bud? Dave?
Gary,
If the stepped attenuator had a tap in the "primary" to drop the turns ratio half of a step, that would be the least difficult and expensive to do.
Keep in mind that Autoformers are still transformers and the secondary turns still act as an isolated winding does, with full phase inversion to the voltage and current, except this occurs in just one wire, rather than two.
The dielectric modification of the Electrostatic moment in the Antenna event within the core bounding box of the pass through window, is largely eliminated in an autoformer. So you give up some performance and gain some other, just like real life.
Bud
If the stepped attenuator had a tap in the "primary" to drop the turns ratio half of a step, that would be the least difficult and expensive to do.
Keep in mind that Autoformers are still transformers and the secondary turns still act as an isolated winding does, with full phase inversion to the voltage and current, except this occurs in just one wire, rather than two.
The dielectric modification of the Electrostatic moment in the Antenna event within the core bounding box of the pass through window, is largely eliminated in an autoformer. So you give up some performance and gain some other, just like real life.
Bud
The normal model of a transformers operation includes the core as being the primary transformation path of flux from the primary windings to secondary windings. Within the cores power to frequency realm this holds true. These ranges are approximately 400 Hz for commercial E/I core, 3500 Hz for 48% Nickel, 7500 Hz for 80 % Nickel and 18 kHz for amorphous core.
Above these frequencies the core material supplies a reluctance to the antenna event, between primary and secondary windings. This reluctance shows up as a rise time limit, to the formation of an EMF, within the window of the core. The antenna event is thereafter the primary transformation mode and that event is controlled by the electrostatic moments, during which the electrical field changes vector, with a flow of current following the collapse of this field. Dielectric materials between and surrounding the windings involved in the antenna event affect this transformation above the power bandwidth of the core and below it.
Commonly lumped as capacitive and inductive coupling, they both have their "distortion" elements, with the primary capacitive distortion arising from 400 Hz and up with the inductive distortion occurring across the power bandwidth of the core.
Autoformer's are subject to surrounding and self capacitance, when wound in a single tapped coil, but not to the antenna event capacitance, found in an isolation transformer. However, the countervailing currents within the wire, common to the primary and secondary do cause phase distortion across the frequency bandwidth, with the core bandwidth factor adding to these distortions.
The loss of the electrostatic moment of the antenna event can be a problem in autoformers, wound as a single coil. However, all of these are in very finely detailed coherence problems, when either isolation or autoformer devices are used. They are different enough to hear in an AB comparison, but it is difficult to portray either as "better" when both have the same level of design expertise underlying their construction.
We would need to involve John Attwod and his measurement tools and skills to quantify these qualitative differences and attach them to objective data.
Bud
Above these frequencies the core material supplies a reluctance to the antenna event, between primary and secondary windings. This reluctance shows up as a rise time limit, to the formation of an EMF, within the window of the core. The antenna event is thereafter the primary transformation mode and that event is controlled by the electrostatic moments, during which the electrical field changes vector, with a flow of current following the collapse of this field. Dielectric materials between and surrounding the windings involved in the antenna event affect this transformation above the power bandwidth of the core and below it.
Commonly lumped as capacitive and inductive coupling, they both have their "distortion" elements, with the primary capacitive distortion arising from 400 Hz and up with the inductive distortion occurring across the power bandwidth of the core.
Autoformer's are subject to surrounding and self capacitance, when wound in a single tapped coil, but not to the antenna event capacitance, found in an isolation transformer. However, the countervailing currents within the wire, common to the primary and secondary do cause phase distortion across the frequency bandwidth, with the core bandwidth factor adding to these distortions.
The loss of the electrostatic moment of the antenna event can be a problem in autoformers, wound as a single coil. However, all of these are in very finely detailed coherence problems, when either isolation or autoformer devices are used. They are different enough to hear in an AB comparison, but it is difficult to portray either as "better" when both have the same level of design expertise underlying their construction.
We would need to involve John Attwod and his measurement tools and skills to quantify these qualitative differences and attach them to objective data.
Bud
Would like to come back to my quickNdirty listening test findings regarding basket grounding as it might be of some interest regarding the discussion about ALNICO speakers.
Have adapted the arrangement in that I connected to the plates both side of the magnet with the plus and minus terminals via a single digit resistor (also serving as a fuse - just in case of) in each connect (checked if they *are* isolated before - which is not the case for the AMT for example).
Well at lest to my ears – the same improvement as with Bud's loop (hash reduction and its positive side effects).
This may give us a bridge to the subjective better rated ALNICO if we think of this material being conductive in contrary to the usual ferro magnets.
Maybe someone who has both types available may be so kind to check this out?
Michael
Have adapted the arrangement in that I connected to the plates both side of the magnet with the plus and minus terminals via a single digit resistor (also serving as a fuse - just in case of) in each connect (checked if they *are* isolated before
- which is not the case for the AMT for example).Well at lest to my ears – the same improvement as with Bud's loop (hash reduction and its positive side effects).
This may give us a bridge to the subjective better rated ALNICO if we think of this material being conductive in contrary to the usual ferro magnets.
Maybe someone who has both types available may be so kind to check this out?
Michael
Lynn Olson said:
Hi,
Did You already know, that horn MF/HF drivers have problems of their own? With any real device the top end shows itself a bit rude. How woud You relate the transformer stuff to intrinsic interferences? The latter coming from breakup modes in the dome and surround, but most prominent from path length issues (and true HOM to some degree). What is the proportion of that stuff related to psychoacoustics if the ultimate top end (frequency wise) is concerned?
One might use a trasformer. If so, why not as we are used to long for the ultimate? On the other hand. Does one need it? Example given:
driver + horn 110dB@1m with 2.83V
amplifier has signal/noise ratio of at best 90dB - You should measure it at all relevant levels taking the preamp into consideration, resistor noise, radio interference, hum from the supply etc.
=> Noiselevel of 20dB@1m up, and this noise is quite colored tonally, really annoing.
A resistor ladder as Earl Geddes suggests would help You out. It smoothes the impedance of the raw driver. Doing so it helps to apply a propper CD equalization. Besides of lowering the noise level at Your sofa.
It's really strange, how You Lynn and others complicate things. By now someone shows himself fascinated by the conductivity of AlNiCo. Has anybody already heard from up to triple demodulating rings in driver motors? Related to that, what is the use of conductive AlNiCo? Do You all really try to combine the emotional impact of former myths with top notch demands and comparable low budget?
have fun
soongsc said:
Hi,
You are sure? Eddy currents occure concentric to the pole piece. You can not shorten them by connecting top- and backplate. Eddy currents are addressed today by Faraday rings within the motor (BMS claims to use three). By the way with great success.
Distortion of 12" PA intended low/middrivers in the midrange seems to be a spurious flaw of cone resonances. High level, commonly out of band resonances tend to emphasize the radiation of very low level distortions. Having a resonant peak at ~2kHz with +10dB level would raise HD3 at 650Hz by that 10dB. That makes the difference between barely acceptable 1% and 3% that many would consider as to avoid at any cost.
cheers
When you connect the top plate to the - lug on the basket, this eventually is connected to ground in most amps. Even if this is connected to the hot side. The eddy current is more actively controlled rather than induced. I also looked into a Japanese desiged speaker, it actually had a dedicated ground lug along with the live and return ones.
When I asked one old driver builder here to try it, he came back and asked me "shouldn't we patent this?"
When I asked one old driver builder here to try it, he came back and asked me "shouldn't we patent this?"
soongsc said:When you connect the top plate to the - lug on the basket, this eventually is connected to ground in most amps. Even if this is connected to the hot side. The eddy current is more actively controlled rather than induced. I also looked into a Japanese desiged speaker, it actually had a dedicated ground lug along with the live and return ones.
When I asked one old driver builder here to try it, he came back and asked me "shouldn't we patent this?"
Hi,
As far as my knowledge holds eddy currents are shortened by itself. This is why it is called eddy. From that there is no use in connecting some building parts of the motor to ground. Ground doesn't mean anything in this context either. To address eddy currents it is most common to use Farady rings, sometimes called demodulation rings. The implementation is a bit delicate. 18Sound invented a costly AIC to cure that eddies. Do You really think they were all to dumb to realize a possibility like connecting "-" to the topplate?!
O/k, people DO hear it.
no offence
No, my aim in argumentation is not for an other proof about eddy current treatment.
There is something - in my observation at least - that bridges to Bud's attempts - and this one *for sure* has nothing to do with eddy currents cure of speakers (works at any amp stages as well as can be read) - and *might* bridge to the ALNICO issue.
I also was pointing out that deliberately inserting some low resistance to (what I think) is Bud's idea does not harm in the first place - putting some light on the electron reservoir theory.
Frankly - I'm not expert enough about ALNICO sonics (only one peace of gear that I was pretty happy with some time ago) to clearly judge if its more or less the same outcome - hence asked for experiences or validation from people that are possibly more familiar with "the sound of ALNICO" or would like to do a proof.
- For those who don't "believe" in such sublime differences - Ok -
- for those who are sold to the ALNICO factor it would be too much of a disillusion to confirm
- for those who would like to mimic ALNICO sound for cheap to nothing it could be revealing
- and for the scientist among us it could become one more "brainteaser".
Michael
There is something - in my observation at least - that bridges to Bud's attempts - and this one *for sure* has nothing to do with eddy currents cure of speakers (works at any amp stages as well as can be read) - and *might* bridge to the ALNICO issue.
I also was pointing out that deliberately inserting some low resistance to (what I think) is Bud's idea does not harm in the first place - putting some light on the electron reservoir theory.
Frankly - I'm not expert enough about ALNICO sonics (only one peace of gear that I was pretty happy with some time ago) to clearly judge if its more or less the same outcome - hence asked for experiences or validation from people that are possibly more familiar with "the sound of ALNICO" or would like to do a proof.
- For those who don't "believe" in such sublime differences - Ok -
- for those who are sold to the ALNICO factor it would be too much of a disillusion to confirm
- for those who would like to mimic ALNICO sound for cheap to nothing it could be revealing
- and for the scientist among us it could become one more "brainteaser".
Michael
The thing is, compared to the days when some of these widely held audiophile myths may have been reasonable concerns, science has advanced orders of magnitude. Yet, while many of these myths have been dispelled by today's science, the long time "authorities" of audiophiledom continue to reject the science. It reminds me of the days when then church defined what was acceptable science and what was sacrilege. Everyone knew that the earth is the center of the universe. To believe otherwise was heresy! 
I have a few drivers with the faraday ring, it did change inductance coil inductance which shows in the impedance change. I do agree that the same method can be used for different cures when design parameters are altered. But there are other ways for the same cure as well. So I don't think there is only one right way to skin the cat.xpert said:
I agree that the eddy currents short themselves out, but depending on shape, material, of conductor and it's relation with the alternating magnetic field, it reacts differently as well. I wish I had the math skills to solve the equations to see if there are better solutions.
mige0 said:good to see that alt least some remember the very roots of *science* are in religious belives - tells us something about ardent scientists ..
And the root of religious beliefs is in our need to explain (and control, as much as we can), the world around us. Science is not a collection of facts, but a method of uncovering the nature of physical relationships. Turns out, the method works pretty well for that purpose. And good (successful) scientists look first to the simplest explanation that accounts for the observed phenomenon.
As regards grounding the speaker basket: Bud sometimes talks in riddles that are incomprehensible to me. But I have observed on one occasion I was getting RF interference in one driver of my multichannel/multiamped system (it was, ironically enough, a christian radio station). It was on a 15" bass unit, driven by an AB transistor amp. I could modulate the sound by being close to the speaker, or touching it. Connecting the basket to the speaker common eliminated the interference. Could be a pool of special electrons. Or maybe something boring, like the basket acting as an antenna.
Sheldon