Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
73.12.71.57
In Reply to: RE: 3.7 vs 3.7i posted by Bigjimsguitars on April 01, 2015 at 05:19:17
Not sure how you time correct in a passive speaker level crossover.
Follow Ups:
Well, you could....sort of, with a lattice network. Take a look at the tweeter section of the Tympani IV schematic.
However, that's not what's going on with the 3.7i setup. Jim is just speculating. :)
There would be no reason to time correct the 3.7 system anyways since it already is phase-coherent resulting from the drivers all being on the same physical plane.
Dave.
When it comes to things audio, phase coherence remains as a concept and not an attainable entity. Coherence means that two signals will have the same frequency, amplitude and phase at the same point in space (spatial coherence) or time (time coherence). The audio world completely ignores the amplitude requirement. But even so, does anyone think that two drivers -say a mid driver and a woofer - producing the same sine wave signal will result in coherent waves? While it is possible to design a crossover which maintains a constant phase difference, it is not possible to build one because of component error. But even so, the drivers are physically incapable of performing this feat. Think of speakers like light bulbs where the light is always incoherent. They are not mono beam lasers.
Some manufacturers offset the acoustic centers of their drivers to compensate for time differences but this has nothing to do with phase.
Nelson Pass wrote an interesting article on phase coherence years ago and even extolled the virtues of the Tympani 1-D and MG IIa but concluded that phase coherent speakers are not achievable (he did not have computerized DSP crossovers at the time but I don't think anything has changed).
I don't necessarily disagree with your statement. This is a tricky one for everyone to get their brains wrapped around. :)
You'll note that I labeled "phase-coherence" as more of a marketing term.
Consider the case of two identical drivers coupled effectively acoustically, but driven with two different frequency ranges (ranges within both their capabilities.) Assuming a properly implemented crossover, could/would that system exhibit "phase-coherence?"Magnepan's are large diaphragm systems with non-coincident drivers and if we really want to get prickly and technical with the technical goal posts, I would stipulate that it's impossible to achieve phase-coherence, time-coherence, linear-phase, square-wave reproduction, etc, etc, and any other lofty objectives audiophiles might think they can achieve.
However, what Andy picked up on was a good observation.
Good fun......much more interesting than ficus tree discussions.Regardless, I still don't think any of this is the basis for the 3.7i modification. But, just speculating on that. :)
Cheers,
Dave.
Edits: 04/02/15
Davey, neither the day nor the week, for that matter, are proper for me to express my feelings about the manner and style of your comments.
I can say that -- on my book -- Satie's input did not warrant being called what was said. I can also say that Jim's ficus tree discussion did bring a good thing or two to the table, as well. And so did Pictureguy here, along with others.
My hope and prayers are that you can review these comments of yours -- and analogous ones in the past -- and find more constructive ways to say things. Your effort in bringing out the best of what our collective experiences may have yielded, could distil quite a few gold audio nuggets, more fun for all...and more enjoyment for you also.
I don't have a problem with Davey blatting his horn at me, but thanks for the sentiment. I am well aware that when I don't specify "electronic" on "crossover" he would be free to interpret it as either the electronic or the acoustic result and that he will inevitably interpret it so that it would grate on him and require terse and severe sounding responses. He's the pro and I am the dabbler so I appreciate him correcting me when I am not specific enough or appear to be just plain wrong. That it comes with the occasional kick in the face sort of scolding is ok. I just wish his replies had more explicit explanatory content and not had riddles and mazes.
Well, you might remember I gave you a long, multiple-post (off line as well) explicit explanation of your misunderstanding of the Sanders cable A/B topic awhile back. Was I successful at convincing you?? :)
Maybe it's possible that riddles and mazes might spawn more objective thinking and less falling back to entrenched "audiophile ideas"?I don't necessarily enjoy banging my head against the wall trying to convince somebody who doesn't want to be convinced. Sometimes I throw my hands up and sometimes I don't. Sorry about that.
Cheers,
Dave.
Edits: 04/02/15
You don't need to convince me. What you are doing is educating all of us. Whether or not I end up agreeing is immaterial.
I continue to disbelieve in the idealized source models for current and voltage behind the Sanders test.I still think that it is a better idea to do the test with the cable connections disconnected and connected at both ends when switching from one cable to another. Even if the assumption of idealized source is maintained then the results would still be more definitive.
Some things (and electronic audio equipment) in the real-world can function pretty close to a theoretical model. (In this case a voltage-source.) However, it seemed that was a premise you couldn't accept so I understand why you continue to disbelieve.
The incorrect analysis about current splitting/sharing and various other conclusions were formed after you'd run off on a tangent. I made an attempt to bring you back but wasn't successful. Oh well.
I can't disagree with you that doing an actual physical switch of the cables might be more definitive...maybe...but by how much? Regardless, that option wasn't in the original premise of the Sanders evaluation technique. In fact his WHOLE objective was to set up a configuration where you wouldn't have to resort to that. :)
Discussions like all of these are of limited value if an audiophile can't agree to accept basic engineering principles. I find this somewhat amusing/ironic in your case since you are an engineer. :)
Cheers,
Dave.
We were always taught to keep in mind the simplifying assumptions and approximations we are using in order to have a useful equation and to make sure they apply where we use it - and if not then how to determine by experiment how much of a fudge factor should be used. Also, we were taught that actual results matter more than sticking to the application of equations. That is why where there is a chem E there is a lab nearby.
Well, did you actually perform the Sanders experiment and produce some actual results? :) At least I'm basing my opinion on a solid electronic engineering principle and previous/actual/similar experiments along those lines.
Maybe considering some alternate (non-audio) examples might persuade you. Off the top of my head, how about the power distribution system in your home for example? (Think about that one a bit.)
Anyways, WAY off topic for this thread. I'd still like to learn of some solid (incontrovertible) information regarding the "i" configuration of the 3.7's.
Cheers,
Dave.
JBen,
IMO Davey's comment about the ficus was not off-base. I refrained from commenting on that post but I can assure you that the presence of a ficus (real or fake) will have no effect on the room acoustics. The "ficus effect" is an urban myth. I did not read all the comments in the thread but unfortunately the OP got bad advice and bought a 2nd ficus. Davey also did not post because I guess he felt like I did that the subject was trivial. Perhaps the OP should have posted on the Tweaks forum where this kind of thing is more acceptable.
As far as his reply to Satie, I agree it was a bit blunt but he was correct and IMO Satie should not take offense as Davey said nothing negative about him. He only disagreed with his statement. I have to say that both Satie and Davey are valuable contributors to this forum and while Davey could have been gentler in his response, the facts remain regardless how they are expressed.
I don't think Davey will change his style and if you don't like his comments, just skip over them. I suspect to some extent that's what Satie does.
BTW before I get flak about the ficus, here is a picture of my room.
They look better than the bass traps they conceal. They also hide some of the actual diffusors. I bought them to improve the room, but not from the standpoint of acoustics. I'm perfectly satisfied with the artificial plants - as decorative devices.
Dave is a no BS kind of guy. I respect that. There's nothing I'd rather know than the truth, but we're all irrational in more ways than we'd like to admit (I obviously include myself).
By the way, my cables are routed across cheap insulators that I spray painted. It satisfies an anal-retentive desire for neatness, as well as discouraging careless walking in that area of the room.
Yes, if you look closely you can see some absorbers (DIY with Roxul 60 and burlap on wood frames) behind the pseudo-plants. I would have put real plants in the room but with my horticultural talents, I would have a lot of leafless twigs in pots within a very short time.
Assuming you really meant "time-coherent", Dave, I would say your definition is different to what, say, Vandersteen says?Sure the rebate for the ribbon is carefully judged to put it in the same vertical plane as the mylar driver which supports the bass & mid panels. But but but ... when you set up a pair of 2' wide panels in front of your listening seat, the centre-line of all 3 drivers cannot be the same distance away from your ears.
So you cannot get true time-alignment with 3.X ... or 20.X, for that matter. :-((
Regards,Andy
Edits: 04/01/15
Andy wrote:
" ... when you set up a pair of 2' wide panels in front of your listening seat, the centre-line of all 3 drivers cannot be the same distance away from your ears."
You may also find that the distance from top-center-bottom of the driver is even more out of each other than sideways. You can steer the lobe from the speaker by the crossover but it is difficult for speakers having their drivers side by side to combine this with a large sweetspot.
The vertical dimension is irrelevant because the wavelaunch is cylindrical, If it were not then you would have had terrible image smear well beyond "big violin". And the Scaena and Pipe Dreams speakers would have been a flop. Look at Griffin's paper
Time alignment is done by a number of DSP based maggie users on this forum at least on tympanis and there was at least one impulse response measurement of the result with great outcomes.
I thought that might catch the eye of someone. :)No, I meant phase-coherent.
Phase-coherent is different from linear-phase. Magnepan's can never be linear-phase....it's acoustically impossible.
The definition of "phase-coherent" has been bastardized by various manufacturers. It's actually become more of a marketing term at this point.Phase-coherence and time-coherence are two different things. One can be achieved with Magnepans but not the other.
Cheers,
Dave.
Edits: 04/01/15 04/02/15
Which is why the crossovers are never symmetrical. That way you can adjust delay in the XO filters to get time coherence while the drivers are at different distances. old maggies were setup to time align when facing nearly forwards, the new .7 ones apparently require face on toe in.
Thank you, SAtie, I was JUST ABOUT to chime in with that input. The asymetrical lobing that results in my 1.6s helps image and sweetspot.
Don't scoff at 90 degrees of difference!
Too much is never enough
You missed the point completely. Satie was referring to the crossover alignment not the polar response. (At least that's what he said.) :)
Obviously the polar response is asymmetric. The crossover alignment can contribute to the polar response, but on the defined design axis the crossover alignment is symmetric....at least in the case of Magnepan's.Everybody should easily understand the asymmetrics of the physical construction. Except for the 20.X, the speaker are even asymmetric front to back.
Cheers,
Dave.
Edits: 04/02/15 04/02/15 04/02/15
And Boy, what a setup monster as a result. Especially for persons like me, who don't read the directions.
My MG-1s set themselves up, being marked L/R and meaning it!
My 1.6s sure look funny with the speaker connections visible to the listener. Since flipped back around.
And yes, I realize that for crossovers with different orders, a time difference is introduced. My 1.6s for exampl with a 1st order high pass and 2nd order lowpass have the tweeter ahead by a few miliseconds.
The Stereophile 'road test' is VERY clear on this point.
Too much is never enough
Total nonsense.
All of the crossovers of Magnepan speakers are in fact symmetric. And even if they weren't, it wouldn't directly affect the phase-coherence of the result.
Dave.
The 90 degree difference between a 1st and 2nd order crossover constitutes 'time delay' and can be used in your favor. That's one reason base orientation make such a difference with my 1.6 panels. By that I mean the basic setup before fine tuning. Tweeters IN or OUT and mylar Front or Back. Once that's sorted out, you are good to go. I've EXTENSIVELY tried all orientations.
Too much is never enough
I'm glad you put 'time delay' in quotes. :)
This is a topic that confuses a lot of folks.These are really not first or second order crossovers. Do you understand that? Don't be confused by looking at just the electrical networks.....it's more complicated than that.
Dave.
Edits: 04/02/15
They are ELECTRICALLY 1st / 2nd, right? So how does the physicality of the driver change that?
What am I missing? interaction by sharing a driver?
How MUCH more complicated and how much difference? Replacing the speaker crossover with a line level SEEMS simply like making an electrical duplicate, but you are saying THAT may not be the case. Eleaborate, please.
Too much is never enough
I didn't say anything about line-level crossovers, did I?The order of the electrical crossover is only part of the design. The ultimate acoustic crossover is the objective. In nearly all speakers systems from all manufacturers this objective is a symmetric alignment with complimentary roll-offs for adjacent drivers. Crossover electrical topologies can take on all kinds of crazy configurations to achieve that goal.
Cheers,
Dave.
Edits: 04/02/15
I did not talk of the acoustic result (outside of the time issue) but the electronic portion, which - if you care to check - should have aligned in 3 way maggies to provide near time alignment when the speakers (pre .7 models) are near face forwards and oriented tweeters out, at distances of approx. 10-12' and in phase when tweeters are in. Measure impulse response at various toe in angles.
This post addresses your post above to Roger too.
A Maggie tweeter is approximately 4.5' high so at a 10' measuring distance there's an approximate 300uS difference in acoustic path length from transducer center to end (top or bottom.) That's actually a phase distortion and inherent in this design. This aspect of the "time alignment" can't be "corrected" with either fancy crossover design or electrical delays via DSP means or any other means.
You can toe the speakers to create similar path length distances for the horizontal offset of multiple drivers, but the vertical aspect remains.
You can measure impulse responses all day long on Maggies at a thousand angles, but it won't yield anything approaching the type of impulse response you could get from a theoretical point-source or even a real-world conventional speaker with point-source radiation objective.
We don't listen to Maggies for their great measured performance in the time domain. :) They have other attributes, obviously.
If time alignment...time coherence...linear-phase response...or any other goal like that is a box that needs to be checked, I suggest to start listening to headphones. :)
Cheers,
Dave.
I don't think there is a vertical component to the output of a line source as far as near field listening is concerned. Meaning that the contribution of the output from the top of the driver and the bottom of the driver does not reach the listener at significant contribution to SPL Those portions simply maintain the cylindrical waveform.
Griffin discussed this in a post here a decade ago.
http://www.audioasylum.com/cgi/t.mpl?f=mug&m=67308
Here is a live URL of his paper
http://www.diy-audio.narod.ru/litr/nflawp.pdf
This means that within the near field, which would be down to 500-600hz on a maggie at 10' there should be the possibility to obtain time coherence within the frequencies important for time domain perception (localization and thus imaging). If successful array extension by reflection off the ceiling and floor are contributing (which in most cases you don't have since the drivers have so little output in the vertical dimension) then you simply have a lower transition freq from near to far field.
As Helmhotz and Guitar Slim demonstrated with their DEQX time alignment of drivers to get rather nice time alignment with the impulse response within 0.1 ms for all the drivers together. The time smear of the line source is negligible and needs no correction by concave shaping of it. The departure from line to concave line produces more of a point source effect which works against image height rendition. One of the advantages of a line array over a box full of drivers covering different freq., time aligned or not
http://www.audioasylum.com/cgi/t.mpl?f=mug&m=192759
Well, the 300us exists. You can't explain it away. :)I understand Jim Griffin's take on this aspect of line-source radiation and chatted with him a few times regarding this. And I understand where you're coming from with your take this, but this is an inherent physical characteristic of the transducer.
A Stanley tape measure will show the relative differences in path length (time smear) do, in fact, exist. No acoustic measurements are necessary to observe/prove that.
Whether our human perception/processing masks this characteristic...our measuring techniques can't fully display it....or how much relative weight we place on this aspect of the radiation, is another topic. A pseudo-line-source like a Maggie speaker will exhibit a comb-filtering aspect to its response if measured in a free-field environment. I don't care what kind of user impulse-response measurements "disprove" that theory. :)I'm not making any bold statements here....just pointing out the obvious.
I believe "image height rendition" of this sort is "large"ly unnatural and that a point-source radiation is more realistic in a domestic environment. So, where does that leave us? :)
A left turn here (as you like to do often) to a different aspect of this. Obviously Maggies don't have this option, but you might consider why the power-tapering wiring configuration was adopted for the line-array design Jim put forward. :) I know he explains his reasoning in his paper, but think about that a little more. :)
Cheers,
Dave.
Edits: 04/03/15
I am not arguing that they don't exist, I am saying that for our psychoacoustic purposes the magnitude of the contributions from the time smear are not sufficient to overcome the precedence effect for localization, but that that it might contribute to the sense of "ethereal" images that line source planars often induce rather than the "solid" images from a simulated point source like a Vandersteen. Should have raised that when we were discussing it with Josh a few months back.
I should point out that Beveridge's elaborate waveguides solve the residual vertical radiation problem pretty much entirely and I don't recall their having an ethereal feel to their images.
What the impulse measurements show is that the time smeared contribution is not at significant levels.to contribute to SPL so would likely be ignored by our hearing mechanism's localization process. Fortunately the time smear is too short to affect our pitch determination, which is a far longer process.
If the image height rendition of line sources is "artificial" it sounds to me far more realistic than anything I have heard from a box - even concentric KEFs and Tannoys.
Pictureguy, thanks for saying what I was.
Jim
ARS VS-110
Customized Bottlehead Foreplay II
Preamp
Magnepan 2.5R's
B&W ASW 300 Subs
Ah Tjoeb Tube CD Player
w/Siemens E288CC's
MaggieMate X/O's from subs to
2.5R's
... while the drivers are at different distances.
Much better to do this using DSP (rather than a passive XO), methinks! :-)) I will report back at the end of the year. :-))
Andy
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: