The beginnings of this thread -- the 1st 68 posts have been copied out of 3 other threads.
Beyond the Ariel
OHM Acoustics "Walsh F" Speaker remakes
Help me kill a fostex artifact.
If some of the comments don't make sense it will be because they are out of their original context
This is the beginning of a rather curious post. I have been bouncing around a few threads saying speculative sounding things about a speaker driver and associated baffle surfaces treatment, that does not look feasible, on first approach.
A moderator has agreed to create this thread, that has all of the pogo-ing posts in one place so everyone can doubt at once.
What I am portraying in these posts is a concept for controlling edged and smooth taper terminus zones. These zones are responsible for the bulk of what is wrong with modern drivers. They cause reflected standing waves, circulating Raleigh waves and diffractive emissions at the final terminus of a "speaker system". Almost all of these reflected artifacts are transient in nature. This means they do dissipate and without an extended dissipation period, due to particular materials used.
That their dissipation over time has a profoundly corrupting influence upon following sound energy and it's resultant reflections is quite obvious, when you get them to stop. What is not obvious is how pervasive the corruptions are. Speakers should theoretically provide infinite detail, complete with "musician applied color" and instrument induced transients and their resultant decays, all of them. Speakers should reproduce the sound field recorded, with enough spatial cues and time to arrival cues, to allow our millions of years old, audible threat assessment system, with it's semi autonomous correlator, to finish recreating the sonic event being reproduced with a fluency and ease that does not tire us nor force us try to hear beyond some irritating noise that interferes.
This EnABL process, portrayed in the following post series, does remove these transient artifacts, produced by uncontrolled reflections and refractions. This is not a theory. It works. It is also so good at this removal that it will not seem possible, for someone who has not tried it themselves, to believe. The qualitative statements you will find, as the braver and more adventurous DIY'rs try it out, will simply not be creditable. "Nothing could be that good" will be the thought and nothing could be farther from the truth, either. There are levels of treatment, having to do with how much your skill at hearing corruptions has developed, and, how maniacal your activities are. But just the most rudimentary attempt will bring huge benefit to your reproduced sound. The real learning process, after a few treated speakers, will be in determining how much is enough.
I will be providing a guided tour through treating a pair of Lowther drivers. I know next to nothing about these speakers and their idiosyncratic sound, but I am certain to horrify those who are devoted to these devices. I apologize to you for this insult, but the person who owns the Lowther’s has asked politely and pretty forcefully that I proceed.
I will not be making any before and after Quantitative measurements. I am going to leave that to you folks who really do know what you are doing in that realm. Of course that means you will have to treat a set of drivers, or two, but I will not be upset about this. I will suggest that very little, of what is a profound qualitative improvement, is going to show up in the usual tests. At least that has been my experience, but, I am not qualified to do these tests anyway. You who must, should probably start qualifying the results of this process with a CSD plot.
There are two documents on line you can refer to.
The first is posted at Positive Feedback Online, Issue 21 at this url http://www.positive-feedback.com/Iss...ndingwaves.htm This will give you some ground to understand from and does provide a truncated CSD on a cone, "phenolic ring" tweeter at the end of the paper. Take special notice of the last pair of comparisons, this is what all of these words are in honor of.
The other is a patent document found at where ever you can obtain these from. I have never been successful, but then, I have the original hardcopy and so my motivation has been low. The patent number is 5,304,746 and it is currently in force, to the extent that I have millions of dollars just waiting to be spent on lawyers to enforce it.
So read on and enjoy,
(original post that above was inserted into follows)
Just in case you folks would like to step outside of the normal mass loaded cone termination regimen, please read this.
The entire process was developed on a pair of Ohm F's and turned them into point source radiators, in a spherical sense. Also eliminated the reflections from the materials joints.
Crank up the WD 40 machine... here we go.
The first thing done was to remove all of the internal foam damping material. Second was to "letter" this block pattern , on the inside, using a flat acrylic model paint, right out of the bottle . This paint was applied directly to the titanium surface, aluminum surface and the bottom edge of the paper using Speed Ball A series, 1 to 5, calligraphy pens (the ones with a bent tip). The pattern template is supplied by printing on a sheet of paper, with printing and scaling provided by Corel Draw 4 or later. Cut the round pattern out, apply it to a thin cardboard stiffener and you are ready to "letter". The scaled pattern needs to be a bit smaller than a direct size, so it settles deeper into the top of the cone than where you are going to apply the pattern. This pattern application location is the bottom edge of each material stage. The same process is applied to the outer surfaces, in the same places, though moved up a bit due to the glue overlap. In this case the ring printed by Corel needs to have the center cut out and it needs to be slightly larger than the applied pattern so that it will sink towards the bottom, underneath where the pattern is to be lettered.
The pattern does not need to be particularly precise so skill needs are rudimentary and if you can hold a calligraphy pen you can perform the needed activity. Best to learn on cheap paper speakers first. And, you will be amazed by how good these things really are.
Application of the conformal coating, on metal surfaces is only over the pattern, not the entire metal surfaces. For the paper you can apply it over all of the cone material, for much deeper bass as it will stiffen the paper, or just over the pattern for what has always been available.
Since this process adds mass only on the transverse wave boundary layer, the amount of mass needed is infinitesimal compared to what is needed to mass damp the entire driver surface. This is true regardless of materials or frequencies being controlled and works as well in infrasonic as it does in ultrasonic, though the pattern is comprised of really tiny blocks of printed material in the ultrasonic realm.
To experiment with dome tweeters and other small radiators you will need to acquire a range of technical lettering pens from the days of yore when mechanical drawings were done by hand on vellum paper. Since this is a scalar architecture, with the same number of blocks for any given ring diameter you will need some pretty fine points and to cut the acrylic paint with water by 10% or so.
Yes, I am the mad fool who developed this, so I can provide you with a Corel file should you be interested in experimentation.
The claims I can make are quite extensive but the one that I really like the benefits of the most are the resulting infinite resolution of micro dynamic signals and the addition of many dB of headroom before compression sets in. All of the information revealed by the process has always been there, just unintelligibly presented as noise, due to corruption by standing waves.
Also, what is really being done here is to speed up, or slow down, the wave front speed over the entire surface, to exactly that speed needed to emit a coherent wave front into the air, as the impulse traverses the driver surface. The white paper tests, shown on the last two pages, clearly indicate that more energy is getting off of the driver surface, before the final terminus of the tweeter in use.
Thanks for the info - and I personally would replace the phrase "mad fool" with "mad scientist". I'm a bit familiar with standing wave theory in airfoil designs - but that is a different application. In electronics and waveguide theory there are some applications especially used in phased array radar systems the traveling wave tubes - again a different application. But air can be considered as a "fluid" when considering the physics involved with compression and resistance to flow. The added problem with air is that it can be compressed where most fluids cannot. I understand your theory on the break-up of the standing waves with your EnABLE pattern which you have patented and I have taken the time to look over your patent at the USGPO. I took particular interest in the application of the pattern around the baffle of the speaker and on the surface of a dome shaped speaker - which might be of use in this type of project. I also took note of the math calculations used in the patent and considered doing just what you said when you stated to draw it up Coral Draw (or something of that sort). I was considering printing it out on a thin sheet of vellum and adhere it to the speaker cone but I was unsure as to what type of material was being used to perform the actual mechanical breakup of the standing wave. I was considering using some speaker dope and powered felt as a place to start and was also considering material to use on the speaker baffle - perhaps some cut up sections of felt pad. But 1st things 1st - I think Covey said that - Build a reference speaker as a base from which to work and compare and mamboni's design seems to be a very good starting point. I should also point out that I haven't been fooling around with audio since the mid 70's when I used to repair them. I used to have a complete set of those nice pens when I designed PWB's the old fashioned way - with Bishops red & blue tape on a precision grid - then I touched up the photo dropouts with the pins.
Isn't something similar to your EnABLE pattern being used to breakup rip tide currents? I seem to remember something like that in Popular Science. BTW - have you ever looked into an application for use with SONAR - might be something there.
damping vs damping
I have not used any materials except the acrylic paint to create the blocks with. I suspect others will work as this does seem to be a local mass interruption, in an area that is inherently low mass and high energy per sq CM. The paint has always performed well on every surface I have tried, and since I have been fooling around with this since 1973, there are a lot of treated surfaces by now.
That it works on incident surfaces is also quite interesting. Relieves the normal diffraction at the edges of boxes and other deflection plates. Use in horns and drivers is quite noticeable with more energy density in the air, very few reflection artifacts in the horns and a general lack of horn "cry" due to the standing waves that populate most horn flares.
With the average speaker box treated and all drivers also treated the box disappears from the sound field and no artifacts give rise to it's reappearance, and the perceived sound field is immense. Done properly there are no "hot spots" at any frequency and images with multiple frequencies or a range of frequencies extending across more than one driver are pin point in their placement, with respect to our ear / correlator / conscious comprehension, and do not wander due to frequency changes, even across differing types of radiators.
This is of course some kind of "speaker holy grail" too good to be true and you can be sure that Not Invented Here has been the only response I have received from those very knowledgeable folks practicing their own form of this black art. To very good results by the way as the EnABL process has far less effect upon modern speakers than it did 30 years ago. Now it pretty much just accomplishes the original goal of eliminating transient standing waves, by making it impossible for them to form in the first place, without altering any of the other tonal and transient events in any way I have been able to measure. Certainly distortion plots and the included traces of various even and odd order HD are not visibly altered.
I was cautioned many year ago by a naval engineer not to mention Sonar in conjunction with this process as the government would and easily could co opt it, without paying me a penny and enforce my silence about the process and it's benefits.
Thanks for your thoughts, it is refreshing.
You're EnABL approach to transverse wave termination is ingenious. Forgive my simplistic analysis, but your pattern of squares is behaving like a virtual diffraction grid, cause optimal phase cancellation in as much as the speed (and phase) of the wave are shifted as it traverses the stiffer lower impedance square-treated cone surface. Now, this in effect blurs the virtually cone-surround boundary, making the cone appear as an infinitely long transmission line. Your solution is elegant and simple. And I think I would be most interested in utilizing it in a future Walsh driver. I admit that I do not understand how the distribution of the squares is calculated for the woofer cone however.
I wanted to discuss the felt approach I implemented vis-a-vis a comment you made about it being mass loading. When I conceived the felt approach, the idea of adding progressive mass to the cone to achieve wave termination and simulate an infinite transmission line was the intuition. However, I discovered something unexpected - the felt fabric is incredibly light. In fact, after the felt treatment the Fs of the woofer is lowered no more than 1-2 hz. Yet the wave terminating ability of the felt is extraordinary. I believe there is something else at work here, far more important than mass: the felt consists of thousands of fibers. I believe the fibers are acting as a huge surface and providing for a much more efficient coupling of vibrating cone surface with air. Normally, the cone-air interface is a very inefficient due to the very different acoustic impedances of solid paper and air. The felt dramatically increases the effective surface area of the cone - many tines over. The best analogy I can think of is the small intestine, which to the naked eye is lined my a smooth pinky mucosal surface. But, under the microscope the surface is covered by billions of tiny hairs (villi), and ultrastructurally, each of these villi have thousands of microvilli. The result is that the effective absorption area of the small intestine is increased over 100-fold.
One might ask: if the felt is increasing the efficiency of acoustic radiation from the cone surface, why does the woofer not become louder per input watt? The answer is that all these thousands of fibers each cause random sound cancelling phase shifts and are behaving more as a low Q reactive impedance. In a way, the felt fibers are acting as thousands of tiny versions of your EnABL squares. Of course, the low Q results from the fibers coupling well with air and efficiently wicking acoustic energy away from the cone into the air.
Based on listening, the felt surface treatment seems exceptionally selective in damping out the resonant stored energy and perceived white noise that normally plagues the paper cone in its upper pass band. The felt has little effect on the woofer when operating in it's lower pistonic band. The result is an incredibly quiet soundstage and exceptional clarity of midrange separation. Of course, this must be heard first hand to be judged for oneself. Do you per chance live near Pennsylvania?
Mamboni / EnABL
I have no doubt that you are actually providing what my EnABL process simulates, an infinite boundary to a cone terminus, by actually providing an essentially infinite boundary layer.
Very clever, and the first specific instance of the use of this powerful lever, an effective control of the boundary layer, to enforce this effect I have seen other than the process I worked out so long ago.
A number of folks around the commercial side of the speaker industry are becoming aware of the problem with boundary layers and that the boundary layer is the only place to control it. As an example Dynavox uses filled polypropylene cones in their large diameter voice coil woofers. Of course this solution just eliminates any sort of boundary layer active control without actually eliminating the transient ringing, just it's duration.
I suspect that a combination of our processes would either be a spectacular success or an equally spectacular failure. I think we should collaborate.
Your description of the silence of the sound field around instruments, with no loss of reverberant information nor micro dynamic detail is very familiar to me. I listen to it almost daily, as I work with speaker cables and interconnect cables that can be tuned for how much dynamic color they preserve and, with excess, add to the energy passing through them. Both of these tools aid greatly in the design and vetting of the audio transformers that I am currently in the middle of the third round of R&D activity on.
Should we ask for a separate thread to continue this discussion on? I suppose we will have to have some private communications also.
I live in the Seattle area and am elderly enough that I do not travel much, so we may never meet. But what fun!
I do have an Auto Cad 14 lisp program already written, that will turn cone outer diameter, inner diameter and length between termini into a pattern set, that could be used for this purpose.
I was originally going to use this, laid out flat, to make a photo sensitive negative from, so that I could make a photo resist print on silk screen, use that to print on paper material, (Hemp?) cut them out and form them into cones. With a bit of edumacation the program works quite well.
I did forget to mention that applying the pattern at both terminus edges of a treated surface is very beneficial. I wonder what a pattern of Mambonis's neat felt, applied to the EnABL pattern would provide?
Your idea of a "tacky" template is quite a good one. The silk screen sock notion was always just too complicated and rife with Murphy crevices to really interest me. You seem to have enough enthusiasm to carry this off so let's try out your idea.
We need to find out what file format Knight prefer their information in. I will assume a vector file and I can produce most of them. What we also need to pursue is a correct template for the cone dimensions you are working with. Simple measurements of the diameter of the two openings and the length between will suffice for a first attempt. Then some small modifications to the thickness entry in the lisp program should give us a pattern for both sides of the cone.
A pattern applied to the "listening" side of a driver will provide about 70% of the control that treating both sides does. The energy pulse that traverses a free air diaphragm is reciprocal in nature. What you do to one side does have consequences for the emissions from the other side. This is true for all cone materials I have worked with, even the dead plastic diaphragms.
If you will email me at the location noted in my bio and provide pretty tight dimensions on the cone you have in mind, we can proceed. If the tacky template idea works out, it will provide an avenue for DIY's to tackle other cone types and sizes.
I was never able to successfully devise a program for projecting pattern blocks onto a dome surface, though there are still only two pattern rings needed. The solid geometry calculations just overwhelmed my poor brain when I attempted to translate the needed description to Auto Cad lisp.
So, feel free to charge ahead, I will be happy to provide intellectual, practical and even some financial support for the project. This is, after all, a public domain intellectual property and the only key I hold, how to make it work in reality, can be had for allowing me to participate in any commercial venture that arises. I am not a particularly greedy person and the patent runs it's course in six years in any event.
Mamboni specks a PIONEER W25GR31-51F 10" BUTYL SURR WOOFER (http://www.partsexpress.com/pe/show...tnumber=290-088) back in post #33. This woofer is desirable for a number of reasons but chief among them is it will go to at least 6 KHz - where most 10" woofers stop around 2K. Also the cone is of butyl construction which is important for its "sound". Not to mention that it is not an expensive speaker - so even if I destroy a couple while I'm fooling around it's not going to break the bank. I think that we may well need to treat both sides of the cone surface because using this woofer to simulate a Walsh device we will be listening to the sound produced by the back wall on the speaker cone and working around the spider most likely will block access to some areas of the rear surface. What are your thoughts on alignment requirements between the patterns on the inside versus the outside? I would think that the patterns should overlay each other and that misalignment could make things worse. On the other hand if the patterns are offset it might serve to breakup the edge reflections even more - and that could be a good thing! :D
Speaking of Walsh (and for that matter DDR, who make a similar transducer) have you shared your design with them? You don't need to answer that - I'm not trying to pry - it's just a suggestion. I would think that they might show an interest in the technology. I can see the "new and improved" ads now - "Our new speakers use the EnABLE technology (patented) that allows for - blah, blah, blah - your listening pleasure". At this juncture I'm not concerned with dome tweeters because I will be using a Heil AMT on the high end. I'm not interested in going commercial (famous last words) been there - done that - got the t-shirt. (One part of my business brain is sketching out a production line in the garage :rolleyes: - the retired part of my brain is saying NO - this is a HOBBY - HELLO :eek:.)
I'll send you a PM "test message" to establish contact and exchange some ideas - I don't want to hijack the thread. :no:
From the EnABL point of view you can ignore the spider, but the basket legs need to be dealt with. Usually just the conformal coating material will suffice. Once the wave front has terminated from the bending board as a coherent wave, it exhibits the same properties as "live" sound does, in that it will wrap around and cross over disturbances without tonal alterations becoming readily evident. I am reasonably certain that the wave front being emitted by a Mamboni treated cone will exhibit these same characteristics.
I suspect that a combination of EnABL and Mamboni treatments will work very well together. One "complaint" that I always had with a fully EnABL treated Walsh driver as you would find in a set of Walsh two speakers,or any other type of cone after treatment, was that I had to be extremely careful with the reflections from within the box as the driver becomes very transparent to those, since they are just as phase coherent as the emissions from the front side. I suspect that the felt triangles will help to suppress that back wave sensitivity.
|All times are GMT. The time now is 01:00 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright ©1999-2015 diyAudio