I have a question about your data. You mentioned that each driver needs to be specifically tuned. I understand en2 does this. How many interations of dot spacing and density were required before you:
1. Could see changes in the data.
2. Come to your final pattern.
From this, how much effort do you believe is necessary to "tune" a driver with the current practices?
Thanks.
1. Could see changes in the data.
2. Come to your final pattern.
From this, how much effort do you believe is necessary to "tune" a driver with the current practices?
Thanks.
I am trying to create a new way to explain it because the way you have been using leaves some of us flat. Please address what I stated?
Originally Posted by GeneZ http://www.diyaudio.com/forums/mult...nical-discussion-post2651221.html#post2651221
Just a thought.
I have been thinking about this question. What does EnABL actually produce in a driver? Could it be like the following?
Ever drive a car that has a solid rear axle? Not, an independent suspension? The solid axle with transmit bumps and thuds throughout the entire rear section of the car even if only the right tire was the one that hit the object. But, with an independent suspension? The bump would not transmitted through the entire back of the car and be felt much less.
Likewise.. Does not EnABL in essence break up the effect of a resonances of a driver, turning it into an independent suspension of sorts? In doing so, it minimizes (not eliminate totally)certain resonance problems that would otherwise be expressed over the entire driver at the point it would take place?
tuxedocivic
Soongsc had three main data sets within the work he did. Here are links to them. This should provide you with some further information.
http://www.diyaudio.com/forums/multi-way/100399-enabl-processes-26.html#post1227789 multiple rings phase vs frequency response
http://www.diyaudio.com/forums/multi-way/100399-enabl-processes-28.html#post1231568 dispersal of "break up modes" and resonance nods
http://www.diyaudio.com/forums/multi-way/100399-enabl-processes-226.html#post1427403 large blocks and dots applied
Bud
Soongsc had three main data sets within the work he did. Here are links to them. This should provide you with some further information.
http://www.diyaudio.com/forums/multi-way/100399-enabl-processes-26.html#post1227789 multiple rings phase vs frequency response
http://www.diyaudio.com/forums/multi-way/100399-enabl-processes-28.html#post1231568 dispersal of "break up modes" and resonance nods
http://www.diyaudio.com/forums/multi-way/100399-enabl-processes-226.html#post1427403 large blocks and dots applied
Bud
Genez
This pretty accurately describes what we have found, since using soongsc's work to figure out how to discover the major problem areas. We are not yet correcting the very small zones, as adding too many ring sets shoves the high frequency phase, or perhaps " time of arrival" too far up/forward and we end up with other problems.
SY's notion of using fractals to control these fine scale problems is one investigation. When dlr finds his perfect positions for controlling the gross difficulties that large distributed masses effect, the two investigations may be a very good combination and finally provide dispersal for all levels of resonant behavior.
Bud
This pretty accurately describes what we have found, since using soongsc's work to figure out how to discover the major problem areas. We are not yet correcting the very small zones, as adding too many ring sets shoves the high frequency phase, or perhaps " time of arrival" too far up/forward and we end up with other problems.
SY's notion of using fractals to control these fine scale problems is one investigation. When dlr finds his perfect positions for controlling the gross difficulties that large distributed masses effect, the two investigations may be a very good combination and finally provide dispersal for all levels of resonant behavior.
Bud
I feel your pain. Been in situations like that in regards to other things.
Question: "You and I and every other thing are a dependent arising, empty of any inherent reality" Tsong Ko Pa
Wouldn't being a dependent arising? Be a reality in itself?
Inspired by EnABL discussions, I started to look at various damping methods. It probably took me at least a hundred tests till I started to get a good feeling what was going on. Even after that, it took me probably 20 iterations to finalize on one specific design for a specific driver. The tests performed using the JX92S took probably 10 iterations without looking into mass distribution because I think the JX92S profile was pretty well handled that. I had some difficulty with stiffer aluminum drivers, and from Klippel scan measurements, it seems that there is no way of improving that kind of diaphragm without seriously effecting the sensitivity of the driver. This is when I started to look into structural stiffening patterns, one of which I had patented. Some of the JX92S metal patterns are just an example of what can be done.
If the driver profile is designed as good as the JS92S, then effort to further tune it would be much less. The basic design of the driver is the most important part. So how much effort it takes to tune a driver depends on the driver.
I am going to have to answer "somewhat but not exactly".
Theoretically, the pattern should break up the wave which travels in the axial direction so that you get waves in the circular direction. Once you get the waves in the circular direction it helps to damp the waves in the axial direction due to the cone material deformation interaction between these two modes. B&W uses weaving to create this effect as well. Then you have the mass of the patterns that also effects how the axial wave travels.
full circle???
If you're not trying to convince anyone here, why are you posting in the first place?? You post data with no supporting documentation or explanation, then conjure up an "explanation", subsequently referring to it in later posts as irrefutably demonstrated as fact.
Since this is the technical thread, there SHOULD at least be some form of peer review allowed, even if it doesn't meet your liking.
John L.
If you're not trying to convince anyone here, why are you posting in the first place?? You post data with no supporting documentation or explanation, then conjure up an "explanation", subsequently referring to it in later posts as irrefutably demonstrated as fact.
Since this is the technical thread, there SHOULD at least be some form of peer review allowed, even if it doesn't meet your liking.
John L.
There is no perfect positioning of any pattern, for two reasons.
First, every driver is different. No pattern nor any combination of patterns will work the same on any two drivers. As others have pointed out, it's just as likely that any specific pattern, no matter the pattern, will be detrimental.
I'm glad to see that we have another case of someone actually providing measurements . The results contradict the claims of specific improvements that have been said will undoubtedly occur. No smoothing of resonances, essentially no change in dispersion, just typical unit-to-unit variation. As mentioned, the change could be interpreted as being detrimental, though in the perception of the differences, preference is a factor. What one person calls an improvement, another may say otherwise. I have no issue with that, if someone likes the change, fine.
Second, the only way to find a truly optimal (read as not perfect) application, would be to have access to something such as a laser interferometer or a Klippel system. Even in those cases, it would be a very iterative process for any existing driver. Short of that, it's guess-work trial-and-error and an exceedingly large number of measurements. All of it. If an improvement could be made. I've had cases where no improvement could be made. The better the driver, the more likely this is the case. Transducer engineers aren't all amateurs as are most of us.
There just is no one-size-fits-all.
Dave
I do not use the EnABL process, but rather the discussion inspired me to look into what is going on. If people showed me the same data I posted, that would be enough for me to know what I need to do. I see no value to treat this as a peer review process. I have already explained what the patterns are supposed to do, if people believe it, fine; if not, it's fine also. I share experience only because the initial discussion inspired me, and the results from the inspiration were very good for me. When I see data that is close to what I have already presented, then probably it's time to take the discussion deeper.
I noticed that several companies are now using various types of weaved materials. The speakers I now listen to for nearfield have weaved Kevlar. Warfedale, Mordaunt Short, and Audioengine are now all using a weave.
An externally hosted image should be here but it was not working when we last tested it.
Yup.
We do know how to find the important surface locations for pattern rings for specific drivers. The original pattern locations, at diaphragm edges, were effective but being able to locate the on surface positions is much more effective. Subjective and anecdotal as that improvement has been to date, even with what data we have. Looking at Soongsc's elimination of the major resonance node in the Jordan driver we can see what is potentially available from investigation.
Bud
Stock
enabled
I know the impulse is related to CSD. Is there something here?
DLR, I can't figure out the window thing. When I have more time I'll tinker. My measurements do get spikey above 10khz. Perhaps background noise? The speaker SPL would still far outweigh any background noise afaik.
enabled
I know the impulse is related to CSD. Is there something here?
DLR, I can't figure out the window thing. When I have more time I'll tinker. My measurements do get spikey above 10khz. Perhaps background noise? The speaker SPL would still far outweigh any background noise afaik.
We know what an ideal response would look like. So when we look at these responses we can notice that the first peak location change as well as the increase in amplitude where it should be damping off. These are caused by delayed reflection of the waves. So we see that EnABL is changing the reflection. If you can get the response to taper off in time instead of seeing those reflective spikes, then you are getting better. For John K, this means less salt and less load on the heart. 
We know what an ideal response would look like. So when we look at these responses we can notice that the first peak location change as well as the increase in amplitude where it should be damping off. These are caused by delayed reflection of the waves. So we see that EnABL is changing the reflection. If you can get the response to taper off in time instead of seeing those reflective spikes, then you are getting better. For John K, this means less salt and less load on the heart.
Nothing can be concluded about the location of the first peak since, among other things, HOLM Impulse is a single channel system so there is no ref pulse. We also know nothing about the positioning of the mic and how the start of the window is set. About all that can be concluded is that both impulses are pretty poor and that the Enabled case is certainly not an improvement. Neither is an indication of accurate reproduction. You would likely see more improvement by spraying the cones with rubberized automotive under coater.
True. I am not saying that those are improvements though. Just using those as an example for explaining changes that occur. Even though HOLM Impulse is only a single channel system, and the way it positions time zero might be questionable; if we look at the full impulse on a same measurement setup, the changes that occur can still be compared.
I curious as to whether anyone had investigated fiber rubbing in these cases.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.