| bobhayes |
Hi,
Does anyone know the driver characteristic responsible for the annoying 5kHz peak in the response of the Fostex 206E.
There are other peaks at 2.5kHz and 10kHz acording to the fostex plot, but they are not annoying me at all, maybe because all I can hear is this emphasised SSSSS!
It was evident before I cut the dustcaps out, it was equally evident with no dustcaps. I can't kill it with a cylindrical phase plug, and now my mind is dissolving into madness, and I can no longer see the light
!
Is damping the whizzer the solution, of does the nasty artifact originate elsewhere, perhaps a main cone break up mode?
I would also be curious to know which bump in the 206 impedance curve is the whizzer resonance, and the origin of the other (1kHz and 2.5kHz)
Apart from the 1kHz bump, it all looks a bit harmonic to me.. 2.5k, 5k, 10k.
Anyone shed any light on this for me? |
|
|
| MarkMcK |
Could you supply an image of the frequency response chart showing the resonance modes?
This would be useful in trying to answer some of your questions. Keep in mind that Fostex has a habit of overlapping cone and dustcap/whizzer resonance modes. Instead of one resonance, there may well be a cone problem and a whizzer problem at the same frequency.
Mark |
|
|
| BudP |
Bob,
How brave are you?
One of the guys following the Walsh thread got brave and did what you see in the attached pic. Here are his words.
"I re-enabled my Fostex drivers per your template posted at the diy forum. Easy to do since the blocks are acrylic and wipe off with alcohol. Apparently, I was off by about 18 sets of blocks! This time I enabled the inside of the cone and whizzer as well as the outside, except that on the inside of the whizzer I "Mambonied" felt triangles. I also, Mambonied the back of the cone.
Right now I'm experiencing sound as if it were not coming from the speakers. Sounds now appear to come out of solid objects. The sounds in the background are now clear. And the sibilance is now gone; violins and close miked female vocal recordings don't cut my ears off anymore. The musical picture is/sounds more complete. I think this is quite remarkable coming from a near full range driver"
Look at this thread for more info.
http://www.diyaudio.com/forums/show...89&pagenumber=3
Be glad to help you scare yourself, just to preserve your sanity. I will be treating a Lowther shortly .
Bud |
|
|
| planet10 |
Thanx Bud, that is a VERY illustrative. Thanx.
dave |
|
|
| Scottmoose |
| You guys are brave men. :devilr: Might have to give this a shot on my next pair of drivers. |
|
|
| chrisb |
Bud - this and the several other threads on which your EnAbl and the Mamboni process are discussed are dizzying to try to follow.
In regards to the photo of the EnAbled FE206, the colour of the cone is certainly not stock Fostex paper. Can you shed any light on the material used? |
|
|
| MJK |
Do you have any measurements of SPL response or electrical impedance before and after treating the driver?
It would be interesting to see what changes in driver performance occured. Intermediate measurements would also be interesting to be able to assess which modification produced which change in response. |
|
|
| BudP |
Hi Chris,
You guys know lot's more about what Fostex has used than I.
I will say that the semi translucent nature of the paper, as we see it, would be typical of a thin, calendered, low rag content, high wood pulp content paper, when the conformal coating, I, and the brave soul who did this work, use.
I am trying to entice him to join in so you can all grill him in detail but that will be up to him.
Here is the other pic he sent to me and it shows the thickness of the felt he used and also the inner cone pattern. Scary thick felt pieces from my perspective, but I really have no idea how bad the problems he had were and this might be just the ticket for whizzer cones.
Dave, is it time to start an EnABL thread, that begins to collect all of the different posts in one spot? Is that even possible? Hate to have Chris fall out of his chair just trying to keep track of me jumping about.
Bud |
|
|
| BudP |
Martin,
I do not have any quantitative data from this treatment. I have some data from other speakers and it is fairly generic, seems to have happened to all of the 2 or 3 hundred drivers I have gone through in 40 years of fooling around. Go here and look at the dissected CSD plots, at the end of the paper, from a "phenolic ring" cone tweeter as was popular in inexpensive speakers in the 80's.
http://www.positive-feedback.com/Is...andingwaves.htm
I will say that high frequency drivers will gain guite a bit more efficiency if you apply more than one coating of conformal /speed control material.
Other tests I have run, with an early generation Lincoln Audio test set up, show no observable alteration in the amount or distribution of harmonic distortion, throughout the first nine categories.
Bud |
|
|
| planet10 |
| quote: | Originally posted by BudP
Dave, is it time to start an EnABL thread, that begins to collect all of the different posts in one spot? Is that even possible? Hate to have Chris fall out of his chair just trying to keep track of me jumping about. |
Yes it is... in this case probably gather together copies of all the relevant posts into a single thread, and if you write an intro post once that is created i can make it show up 1st. A list of relevant existing post numbers and the thread they are in is a good starting point.
Like Chris the EnAble has been hard to keep up with... i keep getting further & further in the Ohm F thread because i really want to go thru it in one sitting.
dave |
|
|
| chrisb |
| quote: | Originally posted by BudP
Hi Chris,
You guys know lot's more about what Fostex has used than I.
I will say that the semi translucent nature of the paper, as we see it, would be typical of a thin, calendered, low rag content, high wood pulp content paper, when the conformal coating, I, and the brave soul who did this work, use.
I am trying to entice him to join in so you can all grill him in detail but that will be up to him.
Here is the other pic he sent to me and it shows the thickness of the felt he used and also the inner cone pattern. Scary thick felt pieces from my perspective, but I really have no idea how bad the problems he had were and this might be just the ticket for whizzer cones.
Dave, is it time to start an EnABL thread, that begins to collect all of the different posts in one spot? Is that even possible? Hate to have Chris fall out of his chair just trying to keep track of me jumping about.
Bud |
So you're guessing the more mottled and darker than normal "caramel" color to the paper are a result of your recommended conformal coating applied over the treatment blocks?
From personal experience with other drivers, I'd think that a wooden phase plug could certainly be worth trying along with the EnAble pattern treatment.
could you please PM me? |
|
|
| BudP |
Hi Chris,
Did the PM go through, I am never sure I have the procedures quite straight.
As for the color of the cone, I do not know. Just the shiny surface and semi transparent nature of the cone, you can see the Mamboni triangles through the whizzer cone, are familiar.
Phase plugs and other post emitter surfaces all seem to benefit. I have no experience with phase plugs per se but I can see what I would do for many of the ones I have looked at.
The main thing to keep in mind is that this is a three vector pattern and seems to need to be placed just before any major deviation in surface is encountered. So, if the phase plug is an oblate shape, just before its surface becomes tangential to the emitting longitudinal compression wave front, is the most useful place to put the pattern. At the start and actual finish, if there is a distinct one at all, are also places to treat. These will be very small in benefit compared to the one at the tangential circumference.
Once you get involved with this pattern you do tend to treat all surfaces, just to learn the benefits, and one of the benefits is a reeducation of your correlator. Well, "benefits" I suppose, as you will become a bit short on tolerance for drivers that are not as well behaved as the EnABL'ed units are. Something of a pain some times.
Dave,
How should I go about this collection process? Just find and note all of the post numbers in the various threads, copy them to an email along with the initial text you mentioned? Or do I need to copy and reformat the actual posts themselves and send that plus the beginning one?
Bud |
|
|
| planet10 |
| quote: | Originally posted by BudP
How should I go about this collection process? Just find and note all of the post numbers in the various threads, copy them to an email along with the initial text you mentioned? |
Thread URL, with post numbers in an email... i don't need the initial text. I will prepare the new thread, email you the URL, you can make the intro post, let me know you have and i'll make it 1st.
dave |
|
|
| ultrakaz |
A few comments about the EnABL'd Fostex.
First to answer the original query of the thread, based on my experience with big (10db+) whizzer induced peaks, I think the best way to reduce them is to insert a notch filter. The second way it to diffuse the peak with a big plug, light bulb sized or bigger like the one used on the Beauhorn. The energy in the frequency is still there but is mitigated by the diffusion. I can't find the link but, there is a website where a guy experimented with different shapes and measured them (see also Tony Gee's site). If I recall correctly the best was the mushroom shape at diffusing the peak. The third way is to decouple the driver from the baffle. It is possible the 5kHz peak is baffle exaccerbated.
The Fostex lit states that the cone is made from banana pulp. I applied a few coats of C37 lacquer (the genuine stuff from Enomosser) a few years ago. Over time the color of the cone changed to what it is now, brownish like an aged banana. The result of lacquering was a dulling/smoothing of the sound from an already relatively smooth sounding Fostex driver, 168 Sigma. After reading Bud's and Manboni's descriptions, I thought I'd give it a try especially since I was not using the drivers. The end result is nothing short of fantastic.
In retrospect I wish I would have taken a basic frequency response measurement to see if the actual response was affected. FWIW, the response now is far from flat, but no "big" peaks and infinitely listenable.
If I were to make a guess, I don't think that either treatment is the solution for a big peak at 5kHz. I think that the treatment will make a good sounding driver sound better, and perhaps an okay sounding driver good, but will not act like a notch filter. |
|
|
| bobhayes |
Thanks for all the input.
BudP, I am going to have a good read through the standing wave link. Cheers for that.
But
Not quite that brave yet. Maybe I will be if I can't find a more subtle solution!
ultrakaz,
I think this might be my initial fix for the problem, since I have tried damping the whizzer with blue tack in various weights and locations including dots on the whizzer, whizzer edge, inside the VC, etc. The only difference was lower overall treble sensitivity and increased system bass associated with the change in moving mass, along with the associated raise in system Q. The bloody peak remaineds. I'll experiment more with mass loading later (mechanical tone control!)
Maybe it could be a serious main cone mode since the whizzer is mechanically coupled with no steep roll off? I wouldn't have thought so since the main cone should still be well down at 5 kHz.
I killed the slight sibilance of my 166E's with the simple installation of Dave's phase plugs, but I think the 206 is a harder nut to crack.
Get that thread going, so we can read all the experiences, and maybe draw some real correlations to treatments and effect on frequency response.
Thanks agaibn for all the input.
Keep it rolling.
Bob |
|
|
| G |
| quote: | Originally posted by ultrakaz
A few comments about the EnABL'd Fostex.
First to answer the original query of the thread, based on my experience with big (10db+) whizzer induced peaks, I think the best way to reduce them is to insert a notch filter. The second way it to diffuse the peak with a big plug, light bulb sized or bigger like the one used on the Beauhorn. The energy in the frequency is still there but is mitigated by the diffusion. I can't find the link but, there is a website where a guy experimented with different shapes and measured them (see also Tony Gee's site). If I recall correctly the best was the mushroom shape at diffusing the peak. The third way is to decouple the driver from the baffle. It is possible the 5kHz peak is baffle exaccerbated.
The Fostex lit states that the cone is made from banana pulp. I applied a few coats of C37 lacquer (the genuine stuff from Enomosser) a few years ago. Over time the color of the cone changed to what it is now, brownish like an aged banana. The result of lacquering was a dulling/smoothing of the sound from an already relatively smooth sounding Fostex driver, 168 Sigma. After reading Bud's and Manboni's descriptions, I thought I'd give it a try especially since I was not using the drivers. The end result is nothing short of fantastic.
In retrospect I wish I would have taken a basic frequency response measurement to see if the actual response was affected. FWIW, the response now is far from flat, but no "big" peaks and infinitely listenable.
If I were to make a guess, I don't think that either treatment is the solution for a big peak at 5kHz. I think that the treatment will make a good sounding driver sound better, and perhaps an okay sounding driver good, but will not act like a notch filter. |
I have to say that I agree. I have recently installed phase plugs which definately improved the balance of the driver but the notch filter is still needed to tame the shout in the midrange. I use a low damped SE tube amp with mine so I am going to reduce the L to .6mH and the parallel resistor to 2 ohms. I am using the alternate BR cabinet from the 206E PDF. Not optimum but nice none the less. Strongly considering building Mr. King's "virtual cabinets". |
|
|
| BudP |
ultrakaz,
Thanks for posting and settling the question of the rotted banana cone colors.
Your comment about the 5 kHz peak in the driver is spot on. That is only addressable with a notch filter. If you were to see it in a waterfall it would look like a mountain ridge with small hills on both sides. The pattern would mitigate the length of time the ridge remained formed and would control the other sympathetic resonance's but that is all.
I am certain Mamboni's process is better at controlling this sort of thing than EnABL. Running those resonant ridges into jungle territory is bound to confuse and weaken them far more than the EnABL process will do.
That is why I am on about using both for every instance that they fit and you just proved that they work together even in the areas I was worrying about, the small light cones. Now I guess I will have to try some domes to see what happens there.
We have to make sure that the Mamboni process does not get steamrollered by EnABL hoopla. I really don't think that an all EnABL treatment would have worked nearly as well for you.
You are a brave man.
Bud |
|
|
| bobhayes |
I wish I had some sort of proper measuring system set up.
It would be great to buy a shed load of the cheapest $2 drivers I could, and try differing techniques, and measure the effects.
Does anyone know of any software that runs on mac?
Also, as a matter of interest, does anyone know the speed of sound in paper.
m/s or ft/s etc are fine, I can work with both. I know the fostex's are banana pulp, but it will give me a rough idea.
Does (as the standing wave link states) the speed of sound in a material change with frequency? I thought it was pretty constant, or did I read the M Corrington graph incorrectly? (about half way down the page on the right.)
There are also some other things in the article that aren't clear. Seems like a blend of technical thinking and magic! |
|
|
| bobhayes |
Cheers Dave,
I only got OS 10.2 though.
It might get a bit pricey if I have to buy the mac software too. A bit of a wallet shock for a 'frugal' guy like me! |
|
|
| planet10 |
What Mac & how much RAM?
dave |
|
|
| bobhayes |
G4 powerbook, OS 10.2.8
I don't know much about how these things work, so I will give you what the machine says.
867MHz processor
L2 cache 256K
DIMMO/built in 128MB
DIMMO/J31 DDR SDRAM 512MB.
Does this mean anything to you? |
|
|
| bibster |
| quote: | Originally posted by bobhayes
Does this mean anything to you? |
Means at least you'll need some external audio device, as those iBooks (I have the same) do NOT offer any line in ou mic. in...
(I offered myself a Edirol UA25 for on the road...)
Cheers, Paul
PS: Compu as-such is powerfull enough for 10.4 AND FuzzMeasure...
PPS: FUzzMeasure allows you to make 2 simultanious (or whatever it's writte like) measures in the demo version.... so you can only compare 2 measurement, but hey... to try before you buy...
Paul |
|
|
| MarkMcK |
Before you proceed with any of this I would suggest that you ask for some quantitative documentation. Just posting the Fostex ad sheet is not enough. There has been enough work done to show that Fostex reports an idealized version of actual performance.
Please have a little healthy skepticism.
Best to all,
Mark |
|
|
| planet10 |
| quote: | Originally posted by bibster
Means at least you'll need some external audio device, as those iBooks (I have the same) do NOT offer any line in ou mic. in... |
That particular PowerBook should have audio I/O attached is the spec from MacTracker.
So all you need is FM, Mic, Cables & find a way to get Panther on it (you don't really have enuff RAM for Tiger). Now Apple doesn't sell Panther anymore so maybe eBay?
dave |
|
|
| bobhayes |
What's panther? and FM? Frequency modulation?
Maybe it would be simpler and far cheaper to buy a PC for these things, as there is a great deal of freeware available for PC's, and absolutely none for mac!
Mathcad, Autocad, Tubecad, numerous spreadsheets and .exe files etc.
Everything I want to use is PC based, and I'm stuck here with my other halves notebook. Depressing. |
|
|
| planet10 |
| quote: | Originally posted by bobhayes
What's panther? and FM? Frequency modulation? |
Panther = Mac OS 3.x
FM = Fuzzmeasure
| quote: | Maybe it would be simpler and far cheaper to buy a PC for these things, as there is a great deal of freeware available for PC's, and absolutely none for mac!
Mathcad, Autocad, Tubecad, numerous spreadsheets and .exe files etc.
Everything I want to use is PC based, and I'm stuck here with my other halves notebook. Depressing. |
Certainly is a lot more quantity in Windows....
1st the best PC (not the cheapest -- those can be had for free) is an Intel mac running Parallels.
MethCad granted, but all their competitors have ported to the Mac they will be forced to soon. AutoCad is a far to expensive 3rd rate piece of.... , spreadsheets will all run in Excel on the Mac except the ones using visual Basic (Excel & Word were both out on the Mac years before they were ported to the PC)
please excuse the rant....
dave
|
|
|
| vitalstates |
Steve has asked me to tell you that the cheque is in the post Dave.
lol
Ed |
|
|
| planet10 |
The 1st Steve that popped into my mind was Cresswell... (he works for Apple too)
I'm particularily biased on the AutoCad subject... i have been dealing with it snce it was version 1. A large portion of my client base is Architects, and where i live some 75% of arcitects are on macs and most of them using VectorWorks (even thou a as close as possible port to Windows has existed for years, Windows gets in the way of getting work done). Now VectorWorks is, IMHO, quite superior to AutoCad and for the price of a legitamate copy of AutoCad you can but VectorWorks & the machine to run it on (& even have change left over if you want to be frugal)
With a lot of very good Architects in Victoria exporting their work, a lot of fresh technicians are imported, so i have seen many times guys (& girls) with only AutoCad experience being hired and put in front of a Mac with VectorWorks. For th 1st 1-4 weeks they complain bitterly about how things are different, after that they start to mellow, and by about 6 months they are expressing wonder that they ever put up with AutoCad.
dave |
|
|
| Scottmoose |
Good point -I'd forgotten SC does some work for Apple.
I'm not surprised about the superiority the Macs have in terms of their OS over regular PCs. Benefits of proper development, and integrated at that. My next machine will be a Mac, as you suggested, with a virtual Win setup lurking around when required for MathCad. |
|
|
| BudP |
Hi Bob,
"There are also some other things in the article that aren't clear. Seems like a blend of technical thinking and magic"
Now that you have had a few days to simmer those objections to finding some technical thinking, would you discuss your questions about the magic?
Please understand that even after 40 years of messing with this, I still have difficulty with it being creditable. Questions just give me another look into what is really going on, and most importantly, from the point of view of another mind.
Perhaps this little bit should also move over to the collection thread because I am sure there are a lot of people with questions. Not the least of which is, how to measure what is changing, even how to describe it, in a sensible enough fashion, that someone can formulate an idea about how and what to measure.
Bud |
|
|
| vinylkid58 |
| quote: | Originally posted by planet10
Now VectorWorks is, IMHO, quite superior to AutoCad and for the price of a legitamate copy of AutoCad you can but VectorWorks & the machine to run it on (& even have change left over if you want to be frugal) |
Once upon a time, Autocad was the only game in town, whether it was used for mechanical/arch/civil engineering and design, unless you had a lot of money to spend. Nowadays, there are so many better/cheaper packages available, depending on what discipline we are discussing, I'm surprised Autodesk is still afloat.
Jeff |
|
|
| planet10 |
| quote: | Originally posted by vinylkid58
Once upon a time, Autocad was the only game in town, whether it was used for mechanical/arch/civil engineering and design, unless you had a lot of money to spend. Nowadays, there are so many better/cheaper packages available, depending on what discipline we are discussing, I'm surprised Autodesk is still afloat.
Jeff |
They are still afloat because they established a large base when the were the game ... this is the situation (same thing that made M$ Word the #1 WP)... you know nothing of computers... you ask someone you (think is) an expert, and that personal who has only ever experienced one piece of CAD software says... AutoCad.
dave |
|
|
| larkinrulez |
OT
He...interesting topic, for I'm a CAD-Administrator in a company with 40-50 AutoCad-Licenses.
I love AutoCAD and it will be the CAD-Standard for the next decades.
Of course there are cheaper CAD-Programs and they are good, but AutoCAD is like building your own boxes...it's like a DIY-CAD that can do everything you teach to it...a little LISP a little VBA or C# and it's a great base for your own CAD without any limitations. That's what I'm doing for living and I love my job :)
Edit:
OK...there is one thing I really don't like about AutoCAD. No support for Linux :( |
|
|
| bobhayes |
HI BudP,
I have never tried it or even heard about it before, but I'm just not convinced from the information given that the EnABL process (just the blocks), is actually having any audible effect on the driver. I'm sure it will have a minor effect, just not audible. Bear in mind I am not a scientist of any fashion, so maybe my pea brain just can't handle the data, but I don't understand the following:
| quote: | | An effective solution to the problem of transient standing waves on transducers is to apply a low mass pattern of diffraction blocks to the skin effect surface. |
| quote: | | This pattern must allow relatively free passage to a transverse wave in any direction. When placed near a terminus the EnABL pattern eliminates reflections from the terminus back through the pattern, due to the skin effect energy density within the pattern openings. |
I just think that if the returning reflection is of such low amplitude to be arrested by these blocks, it wouldn't really be a problem in the first place. Why does the high compliance surround not absorb the wave? At what frequencies is the pattern supposed to be effective?
I have seen many cones with a ridged surface, or with one or two raised concentric rings. Do these work on the same principle?
Also, wouldn't this process have more effect on radial modes, if the blocks were glued in a pattern from centre to edge? There is no terminus or surround to prevent radial modes propagating.
| quote: | | The thin coating increases energy density in the boundary layer of the membrane and a properly formulated coating will allow the transverse energy, in skin effect, to exceed the speed of sound through air. |
I thought that the speed of sound in paper, or any material denser than air, was greater than that in air. I would have imagined that this coating, having more density would counter the effect of the blocks, increasing cone internal damping, which I would have thought would add considerable MMS, changing lots of other driver parameters.
Why not use a driver made from a different material such as PVC, GRP or CRP instead?
| quote: | | Coating the membrane overcomes phase transfer dislocations, found on all boundary layers, by allowing a more uniform energy transfer per given area of radiation. |
I would have thought any phase transfer dislocations would be inaudible or immeasurable at anything over a few inches from the driver, where the wave becomes spherical in the bass and lower mids, and directional thereafter due to the complete mush of interacting waves off axis. Would the coating not simply alter these dislocations, rather than 'overcome' them? Surely the coated driver cannot be immune to the phenomena.
| quote: | | This interleaving of fricative sound interferes with time signature recognition and a reduction in depth of field information is the result. |
What is depth of field information? Is it resolution or dynamic range or something else?
| quote: | | The difference in output comes from greater effective emitting area and less resistance to impulse acceleration in the boundary layer. |
So the cone is stiffer? That would make the drivers HF rolloff earlier wouldn't it? (Voice coil coupled to less flexible cone would imply the VC impedance rising at a lower frequency than that of the original cone driver, due to seing more mass, more quickly.)
So the traversing wave is moving faster? How does this benefit the driver? Since the wavefront propagating from the cone nearest the voice coil moves through the air very slowly compared to the wave in the cone, the emitted wave front will be initially concave (assuming a normally mounted driver) at all frequencies that the cone is capable of reproducing (due to size limitations), regardless of what the cone is composed of (within reason) IMO.
| quote: | | These claims are demonstrable and measurable. The enclosed frequency response plots are from FFT processing of an .006ms duration square wave pulse test, with gates set at 0.00 and 5.5ms. |
I make that a 57.333kHz square wave. My hearing is duff over 16kHz.
I don't know what the gates are, but they are set at (high) infinity Hz and (low) 62.5kHz?
There are no labels on the y graph axis, is it SPL dB?
There is no doubt the graphs show changes to the driver, but I cannot see the relevance of 0.19mS (1.8MHz), 0.35mS (983kHz) and 1.19mS (289kHz) in audio transducers. (unless they were hypersensitive to RF instability or something.)
Sorry if I got it all wrong, but it seems to be 'talking up' the actual problem, and I'm not sure the EnABL process (the blocks part) described is effectively combatting any audible driver cone anomalies. I think the coating is the major contributer to any measuable or audible changes.
I'm hoping to get a cheap test setup together soon, to measure changes in response for different mods to generic drivers, but I have no ability to test and interpret such complex cone resonance issues, so I'll still be non the wiser.
That said, I will try these mods out and compare, because after all, it's not a test setup, computer or microphone that will be listening to the drivers, and there is no better judge than the ol' ears
For any change made to the characteristic of a transducer, it is purely subjective as to whether it sounds 'better', as there has never been, and never will be a 'perfect' one (even if there was, everyone would argue it was 'too' perfect). For all the technical advantages in using modern materials and processes, some people would still not trade their paper drivers for anything else. This in itself speaks volumes. Everyone here knows all about compromise, and for every positive action, their will always be a trade off.
If there is discussion of this topic elsewhere, please let me know, so maybe I will get a better grip on what is happening as I still don't fully understand.
Regards, Bob. |
|
|
| Svein_B |
First let me say that I have not studied the EnABL process.
Some weeks ago I did however do some simple measurements of the Fostex FE-126e. This driver has some nasty peaks around 7 kHz.
First I compared a new driver with a coated driver: The coated driver is unanimously reported to sound considerably better than stock. It did however not show any noticeablel change in the measured response (at leat not at higher frequencies).
Based on hints from MarkMck regarding location of resonant modes, I tried to add small felt and cork pads around the perimeter of the driver, 10 mm from edge. This did show results. It seemed that many small pads were better than a few larger ones.
Attached is a measurement I saved. This is with only 4 felt pads. 8 small cork pads (10mm dia) were even better in flattening the HF peaks. I show this only as example of measurement. I did not have the time for a proper listening comparison.
NB: The coating, using ModPodge, is important before doing these experiments as it will allow self adhesive pads to be applied and removed without damaging the cone.
SveinB |
|
|
| planet10 |
I wondn't really expect the mod-podge treatment to show up in the FR -- its job is to reduce cone self-noise....
In my mod the damar pattern is an attempt to help quell the resonances Mark pointed out... i know they aren't totally effective, because i can still see them in the impedance plot (i have to get a laptop assembled to put FR measure kit into place)... we are just about to go another round of experimenting on the cones, i'll give this a try.
dave |
|
|
| JimT |
The 2.7 kHz peak was bothering me the most. I tried the felt pads on the cone about 5/8" in from the edge. This did reduce the peak, but also tended to reduce the upper treble and I think it caused new dips and peaks. I did not want to lose the upper treble so I removed the felt pads. I got a large piece of some heavy felt, about 1/2" thick, and cut a hole for the driver with triangles that projected past the point where this 2.7 kHz resonance came off the cone. I taped this to the speaker baffle. It looks a lot like what Bert Doppenberg did on one of his speakers. If you go to http://www.t-linespeakers.org/FALL/bd-pipes.html and then go down 1 screen you will see what this looks like. It sounded horrible. I reduced the BSC resistor to about half and it now sounds better than ever. The 2.7 kHz peak is gone. While it now sounded very good the 5kHz peak was much obvious. The 5kHz peak seems to come from a resonance well down the cone. I tried the "98 cent tweak", speaker stuffing wrapped between the whizzer and the cone, and this calmed it down some. I have to take better measurements and refine these but hopefully this may be of some help.
Jim |
|
|
| BudP |
Hi Bob,
I missed your excellent challenge ot the EnABL process a week ago. When thread posts are collected into one spot, as they were in the EnABL + Mamboni thread, my link and email notification for the original threads disappeared. This is the first 1000 words or so of my answers to your questions from a week ago. It is also posted in the EnABL thread and all of the rest of my answers will only be posted there.
| quote: | I have never tried it or even heard about it before, but I'm just not convinced from the information given that the EnABL process (just the blocks), is actually having any audible effect on the driver. I'm sure it will have a minor effect, just not audible. Bear in mind I am not a scientist of any fashion, so maybe my pea brain just can't handle the data, but I don't understand the following:
quote: |
| quote: | "An effective solution to the problem of transient standing waves on transducers is to apply a low mass pattern of diffraction blocks to the skin effect surface.
quote:
This pattern must allow relatively free passage to a transverse wave in any direction. When placed near a terminus the EnABL pattern eliminates reflections from the terminus back through the pattern, due to the skin effect energy density within the pattern openings." |
| quote: | | I just think that if the returning reflection is of such low amplitude to be arrested by these blocks, it wouldn't really be a problem in the first place. Why does the high compliance surround not absorb the wave? At what frequencies is the pattern supposed to be effective? |
You may think in anyway that you wish to, we all do so. Please have another look at the paper posted on the Positive Feedback site.
http://www.positive-feedback.com/Is...andingwaves.htm
Scroll to the last graph, right at the end of the paper. Look carefully at the graph and notice that the untreated cone shows evidence of energy that is still radiating from the cone surface. It is still radiating in discreet peaks and troughs, within the time period that our correlator uses for placement of and distance from threat analysis.
Then notice that the untreated cone has ceased to have energy in the same format. There are still some peaks of emitting energy but most has dropped below the test sheets limit. You may think that all of the activity is still going on and I deliberately cut off the chart scale, to prevent you from seeing it. If you were to run this test on your own, you might prove to yourself that I am indeed attempting to hood wink you.
In actuality, the processed cone has just gotten more energy into the air, in a shorter period of time, than the untreated cone was able to. You can see evidence of this in the graphs. Look at the comparison of the first curve sets. Notice that in some discreet frequency points, the treated cone is 9 dB more efficient than the untreated cone. No question in my mind that the conformal coating, used as a boundary layer enhancement is effecting this event.
Now look at the middle graph from the Positive Feedback site. The two cones are emitting energy at about the same rate, at the same point in time. Different frequency points in this decaying energy plot show differences in storage between the two, for differing frequencies, but the amount of power being emitted is within an equivalent level. If the coating was the only effector here, then this graph should also show a noticeable difference in level, everywhere across the frequency band, because a more efficient emitter surface should work more efficiently, regardless of energy levels being worked with.
Something else must be at play here. This is the same tweeter in both sets of tests. The same test set up, same microphone, same computer etc. etc. One day later though. All that seems left is that the pattern is enforcing an increase in energy transfer into the air during this middle graph period and having done so, ceases to allow energy to return in the boundary layer and provide the same “ringing” emission that the untreated cone is producing. There are no other differences between the two test points. So, more efficient emission into the air from a combination of pattern and boundary layer enhancement is the answer, for the paper cone tweeter.
Change the diaphragm materials to poly carbonate, aluminum, titanium and suddenly things are quite different. Same pattern performs the same function, but the conformal coating is reduced to just the area where the pattern blocks are applied or not used at all. Here is where the effectiveness of the pattern shows up. Certainly this pattern increases the boundary layer impedance. Just as certainly, the materials with less boundary layer and higher internal longitudinal wave speed do not require the conformal coating, but do respond to the pattern with the same “character” as found in a drivers where the conformal coating is required.
Bud |
|
|
| planet10 |
The felt star is mine, shown installed on my set of Bert Doppenburg ML-V. The star was adapted from something that GM came up with (for the 40-1354?)
A prime number of points (13) with the edge of the star cut at ~ 45 degrees so that the shape is opening up, and there are no 90 degree surfaces.
dave |
|
|
|
