Planar Speaker Asylum

Welcome! Need support, you got it. Or share your ideas and experiences.

Return to Planar Speaker Asylum

Planar Speaker Asylum

RE: He is right but is referring to low level detail

64.252.147.184

In Reply to: RE: He is right but is referring to low level detail posted by Satie on September 20, 2012 at 20:51:50

Interesting. It seems to me that any effect from the depth of the channels would in the single-ended drivers be to an approximation invariant with respect to amplitude.

The acoustical impedance of a hole is dependent on velocity and so it seems to me possible that the damping of the diaphragm increases as a function of amplitude. But I'd expect this to cause compression rather than downward expansion.

If the nonlinearity of a dished driver, or the characteristics of the tensioned PET film itself, were an issue, I'd expect the same thing to be true of electrostatics, but that doesn't seem to be the case. Of course, the Xmax of an electrostatic is significantly lower than that of a planar dynamic. Again, AFAIK, diaphragm dishing leads to dynamic compression rather than expansion.

The movement of the diaphragm isn't a simple spring constant business since the Young's modulus of aluminum is on the order of 30 times that of Mylar. So the diaphragm isn't isotropic, it will prefer stretching horizontally into a cylindrical shape. But to stretch horizontally it will have to stretch vertically as well and since there isn't much mylar above and below the wires, the wires themselves will presumably stretch. The aluminum itself is presumably going to be within the linear range of the stress-strain curve. So I don't see an obvious source of downward compression here.

Finally, you have dynamic thermal effects -- the adhesive and PET presumably become more ductile as they heat up. The driver could become more limber when you turn the volume up. I found some stress-strain curves for polyester at various temperatures (Fig 1) and they're temperature dependent as you'd expect, but the tested temperatures are in the upper area of the operating range and not all that clear in the linear region:
http://www.4spepro.org/view.php?article=003911-2012-01-12 (Open in New Window)

Follow Ups:

RE: He is right but is referring to low level detail - Satie 15:21:24 09/22/12 (7)
You are looking at the top limits on excursion. The low level issue is different because it is an onset of motion issue where the wires don't deform/strain and the stretch in the mylar is the limiting factor along with the air resistance. The wires are simply stiff and prevent stretching in the vertical dimension at low amplitude.

As you pointed out, air flow resistance increases with velocity, thus limiting the motion of the membrane when departing the onset of motion so that an output sine pattern would flatten relative to the sine signal supplied to the driver. flow resistance is at its highest while the membrane is traveling through the 0 displacement plane.

Resistance to onset of motion is in the mylar/miloxane CLD and the stiffness of the wires that are acting as hard solids since the stress levels applied are insufficient to deform them. At he onset of motion there are viscous effects in the mylar since it is an amorphous polymer rather than an elastomer like rubber. The energy applied is largely lost to plastic deformation at onset of motion and what is left is damped by the air flow resistance. again we have the highest rate of strain as we move through the 0 displacement plane and we have spring resistance as we move towards the peaks on the sine waves. So this also acts to flatten the output sine waveform at low amplitudes/signal. Velocity is near 0 at the peaks.

Once amplitude is sufficiently high, the main resistive force is the air resistance to piston action. The wires are still not deforming under these relatively low forces and the mylar is stretching in a spring like manner no different than the surround on a dynamic driver. The air flow resistance through the channels and holes is insignificant, and the plastic deformation forces are insignificant relative to spring behavior.

For a reference here is a random article that has a good summary on these issues for amorphous plastics vs elastomers. Note the presence of retardation and relaxation effects and the fit of Voigt Kelvin type viscoelastic models to the behaviors and that there is little permanence to the effects of short lived stress.
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CCYQFjAB&url=http%3A%2F%2Fwww.iupac.org%2Fpolyedu%2Fpage36%2Fpage19%2Fpage21%2Ffiles%2FDMTA_1_fixed_A_1.doc&ei=MDZeUMTsDqHl0gGUnYGIBw&usg=AFQjCNFVg0COLIbbodlsYCQgfaZ_H_Q6-A&sig2=CnmTf6ZTcIQZtO21N6rwgQ
- http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd (Open in New Window)
- RE: He is right but is referring to low level detail - josh358 09:06:59 09/30/12 (6)
  "You are looking at the top limits on excursion. The low level issue is different because it is an onset of motion issue where the wires don't deform/strain and the stretch in the mylar is the limiting factor along with the air resistance. The wires are simply stiff and prevent stretching in the vertical dimension at low amplitude."
  
  I hope I'm not too off base here because my time has been limited and I haven't had time to consider the problem as thoroughly as I'd like to. But my take so far is this.
  
  In the regions of the stress-strain curve that we're concerned with aluminum follows Hooke's Law pretty closely, so I wouldn't expect a small-signal nonlinearity in the elastic behavior of the wires themselves.
  
  The electromagnetic forces on the wires are to a rough approximation (ignoring e.g. the curvature of the magnets) normal for small displacements, so in the absence of a diaphragm, there would be no deformation; however, as the surface to which the wires are attached dishes and becomes non-Euclidean, there will be elastic deformation of the wires to the extent that the diaphragm exerts a force.
  
  It is really the geometry of the field here that we are concerned with, rather than the shape of the diaphragm. These are not necessarily identical. I gather from the paper that PET is viscoelastic rather than elastic, so the forces exerted by the diaphragm won't be time invariant, but will exhibit creep. However, the models (Kelvin�Voigt, etc.) seem to be linear functions of time.
  
  So far, I'm having trouble seeing how small signal nonlinearity could result from viscoelastic behavior in the diaphragm. I'd expect some kind of displacement-dependent hysteresis effect, but I don't see how that could come about from time delay, as opposed to forex a non-linear function of changes in temperature or displacement.
  
  Damping is another issue. Air damping should be viscous/resistive and thus linear, but that may not be true of the internal damping of the Mylar, which could have a hysteretic damping component proportional to displacement. Whether that effect could be significant I don't know but it would seem to go in the wrong direction, anyway, e.g., it would tend to cause compression rather than downward expansion.
  
  force on them that isn't , and will behave I think in a
  - RE: He is right but is referring to low level detail - Satie 01:53:43 10/05/12 (5)
    The various elements in the stress strain relationships may be linear on their own but they are all different in scale and kick in at different rates of strain and at different strain levels. The main issues at low signal levels are the stiffness of the wires and the tensed diaphragm, and air resistance. It is simply there. Many have observed it.
    loosen the mylar, open up the board with larger perforations and narrower or thinner magnets and use foil and you have the makings of a more responsive system at low levels - something more akin to a ribbon.
    
    For most folks this simply does not matter. I am aware of it but I am not bothered since I listen at somewhat higher levels anyway.
    - RE: He is right but is referring to low level detail - josh358 05:35:30 10/05/12 (4)
      Well, yeah, I've observed it too, going all the way back to my 1-D's. I just don't know what causes it, or even of what it consists.
      
      As far as I know, the aluminum is operating in the linear region of the stress-strain curve. If it weren't, it would suffer permanent deformation, like a ribbon tweeter that sags after it's been overdriven. So there should be no hysteresis. So I don't think that's the issue.
      
      The magnets are insignificant, the holes dominate. Air damping from the holes is AFAIK resistive. It's caused mostly by eddy currents in the holes and it's a dashpot phenomenon. Again, I don't think you're going to get nonlinearity at low levels from this. Damping should affect transient response. The one thing I've read on enlarging the holes (in electrostatics) implied that it improved transient response. I assume that what you're after here is critical damping.
      
      I don't see a problem with the tensioned Mylar, either. The deformation is non-linear but if that were a problem, electrostatics would exhibit it too.
      
      So I'm down to two hypothetical mechanisms, a time-dependent and nonlinear one involving non-Euclidean deformation of the wires by the viscoelastic adhesive and PET, and thermal effects. Both I think would be consistent with the (subjective and maybe inaccurate) observation that foil exhibits less of the effect than wire.
      - RE: He is right but is referring to low level detail - Satie 13:54:51 10/05/12 (3)
        
        In Reply to: RE: He is right but is referring to low level detail posted by josh358 on October 05, 2012 at 05:35:30
        
        look at the young modulus for PETs :2-2.7 vs aluminum 69 GPa.
        
        The air resistance is linear to velocity/proportional to rate of strain, at 50 db spl at the membrane we are talking 0.02 Pa. peak membrane speed would be about 20 u m/s. We can figure out what the pressure equivalent of the perforated plate's resistance would be.
        
        What I am saying is that the limiting factor dominating resistance to diaphragm motion is different at various output levels. That any of them is linear does not matter since there is a different slope at each level.
        
        RE: He is right but is referring to low level detail - josh358 12:04:42 10/07/12 (2)
        
        In Reply to: RE: He is right but is referring to low level detail posted by Satie on October 05, 2012 at 13:54:51
        
        Varying stress/strain slopes won't matter if the materials are within their linear range. Aluminum should be, whatever the excursion, or it would deform permanently. The stress/strain curve of Mylar and I'm guessing the adhesive, also a polymer, is a bit more complex because of viscoelastic creep, but again, electrostatics don't have this problem, so it can't be as simple as the Mylar. So I don't see how any of these factors could give rise to downward compression in and of itself, so long as motion is linear and temperature is constant (which, of course, they aren't).
        
        RE: He is right but is referring to low level detail - triamp 14:14:29 10/12/12 (1)
        
        In Reply to: RE: He is right but is referring to low level detail posted by josh358 on October 07, 2012 at 12:04:42
        
        Some speakers give the impression of revealing more low level details by EMPHASIZING certain frequencies. Seems to me that if a detail is low level, that is to say, low in level, at low to moderate volumes this detail would be below the level of audibility- unless it's "boosted' in some way. Stands to reason that as you raise overall playback SPL that details that were below the audible threshold would be brought up to audible level....
        
        For example, if one of these "low level details" is 30 dB below the average level of a recording, and you are listening at a moderate level of 80 dB, then that low level detail will be reproduced at 50 dB in your listening room. A room is considered a 'quiet' room (i.e., a library) if it's background sound level is 40 dB. A typical living room would probably be a little higher in SPL. So, at a playback level of 80 dB these "low level details" would likely be BURIED IN ROOM NOISE unless the speakers had some treble peak or other emphasis.
        
        Seems to me that absent frequency response peaks or other coloration, a speaker would only revel low level detail when the overall level is brought up to a fairly loud playback level. Seems that this reviewer has decided that the relatively flat playback response of Magnepan speakers is actually a FLAW as far as he's concerned.
        
        If I want emphasis on some portion of the spectrum, I'll use EQ, thank you very much. I don't find it convenient to use choice of speakers as a TONE CONTROL.
        
        RE: He is right but is referring to low level detail - josh358 08:00:53 10/14/12 (0)
        
        In Reply to: RE: He is right but is referring to low level detail posted by triamp on October 12, 2012 at 14:14:29
        
        I suspect that there are several things going on. One is what you might call true detail, which seems to correspond to a cliff-like waterfall plot. Another is the false detail that you mention, from high frequency emphasis. Another, I think, is that some speakers have their self noise at relatively high frequencies and I think this can provide an illusion of detail. This is mostly true of electrostatics.
        
        Finally, there are the Fletcher-Munsen curves. A speaker that is boosted in the bass and the treble has in effect built-in loudness compensation and will sound better when the volume is backed off. But it will sound wrong at natural levels. The right place for loudness compensation is in the preamp or computer, where it can be applied appropriately (or would in a calibrated system, if recordings had an absolute level reference).
        
        And, of course, there may be some nonlinearity in the speakers themselves, as Satie points out. I'd love to do some measurements to see if I could verify that and track down the source. You could for example test for thermal effects by superimposing a small signal on a large one of a different frequency, and seeing what happens to the small signal. It should differ from IMD in that if there's for example a thermal effect, the small signal should change as a function of time.
RE: He is right but is referring to low level detail - JLindborg 12:31:21 09/21/12 (5)
I actually drove one of the Tympani 1C panels extremely hard just for test.
After a few seconds the wire visually expanded a lot.
I had previously secured the wires with new spray glue so that they would not be affected by the aged old glue.
The wire started to arc along with it's mylar so a wave pattern formed from top to bottom.

I have talked about this with Roger and promised a film/picture of this.
I have not yet come around to do this but I will.

SO, You guys who are power crazy and think that the Magneplanars can handle several hundreds of watts do not actually see what is going on behind the cloth.
It's all but good and HiFi. The mambrane has no more of it's original shape as intended and is in that instant a very poor transducer.

The one who succeeded was the one who didn't know it was impossible.
- RE: He is right but is referring to low level detail - josh358 13:47:12 09/21/12 (4)
  I suspect that wave pattern is always forming, it just isn't easily seen at lower amplitudes. After all, we're talking a tensioned (when driven) wire -- basically a musical instrument string. So it will have vibratory modes, just as the Mylar will. And the two will interact, as the vibrations in the Mylar mostly reflect off the impedance mismatch at the much more massive wire, and as the wire transmits, messily, its own vibratory patterns to the Mylar. And so forth. I think you could see the consequences of this modal behavior very clearly in the study that BigguyinATL did.
  
  Not to say that behavior isn't messier at high levels. Few consumer loudspeakers can handle high levels cleanly. It's the one area in which the sound I remember from the studio was clearly superior. However, kilowatt amps are used for their instantaneous peak capacity. You can't even get past 64 watts to the tweeter without blowing the fuse. With the 10-20 dB peak/average ratio of acoustical music, a kilowatt amp at full capacity will be putting out no more than 10-100 watts average. And I suspect that just as the ear is tolerant of moderate clipping on instantaneous peaks, it's more tolerant of compression and distortion in drivers. Dynamic woofers typically get into the 10% harmonic distortion range at the upper end of their usable output. Not pretty, but we seem to tolerate it until it gets so bad that you hear doubling because of the Fletcher-Munson curves.
  - RE: He is right but is referring to low level detail - JLindborg 15:05:38 09/21/12 (3)
    Well Josh, this was not a wave pattern of sound waves. These waves WAS permanent waves from the wires trying to get longer.
    So my comment stays firmly.
    
    I will post pictures this weekend.
    
    Cheers
    The one who succeeded was the one who didn't know it was impossible.
    - RE: He is right but is referring to low level detail - josh358 15:47:26 09/21/12 (2)
      Do you mean they stretched to the point where they stayed stretched? If so, I misunderstood you.
      - RE: He is right but is referring to low level detail - JLindborg 15:52:58 09/21/12 (1)
        
        In Reply to: RE: He is right but is referring to low level detail posted by josh358 on September 21, 2012 at 15:47:26
        
        Yes Josh! :) Exactly!
        It looks very odd and the first time I was afraid that it would stay that way but it returned to flat after a few seconds when I turned the sound off.
        
        The one who succeeded was the one who didn't know it was impossible.
        
        RE: He is right but is referring to low level detail - josh358 18:05:56 09/21/12 (0)
        
        In Reply to: RE: He is right but is referring to low level detail posted by JLindborg on September 21, 2012 at 15:52:58
        
        LOL, yeah, I'd have been scared shitless. :-)

Post a Followup:

FAQ

Post a Message!

Forgot Password?

Moniker (Username):
Password (Optional):
	Remember my Moniker & Password (What's this?) Eat Me
E-Mail (Optional):
Subject:

Message:	(Posts are subject to Content Rules)

Optional Link URL:
Optional Link Title:
Optional Image URL:
Upload Image:
E-mail Replies:	Automagically notify you when someone responds.

To view your new posting or follow-up, click on the RELOAD or REFRESH button on your browser.

This post is made possible by the generous support of people like you and our sponsors: