"...all you need to do..." hee hee... 
Ok, the Conclusions section basically says that the HOM frequencies are a function of the diameter of the exit hole size of the driver in question, and independent of the horn that comes next. Smaller exit size means higher frequencies before a HOM will be able to occur. It also says that some HOMs that are predicted are not excited in practice - although this part is a bit foggy on a fast read - probably, or I suspect, there is some background information that is "known to those proficient in the art" that explains this, but I don't recognize or know.
So, this gives rise to a number of potential "cures" and questions regarding Dr. Geddes "solution" to the problem.
The first question that comes to mind is why is a big large foam plug required if the finding shows that HOMs occur only back at the exit area of the driver and are unaffected by the use of a horn thereafter. The author states that over 50 horn/driver combinations were tested to verify this result. It would seem to me that a short small foam plug would suffice, and give the same result, without the significant HF rolloff and loss of sensitivity?
Perhaps Dr. Geddes has discovered other HOM modes that extend lower in frequency than those found in the throat area, but he has not yet explicitly shared this finding, afaik.
In the paper there is a somewhat ambiguous and unclear reference to the "internal membrane". I am not sure if the author refers to the diaphragm of the BMS driver (a membrane) or the virtual membrane of the FEM mesh surface being simulated (the planar wave front)?? This makes some difference, since it has different significance depending on if it is the planar wave or the diaphragm itself.
_-_-bear
Ok, the Conclusions section basically says that the HOM frequencies are a function of the diameter of the exit hole size of the driver in question, and independent of the horn that comes next. Smaller exit size means higher frequencies before a HOM will be able to occur. It also says that some HOMs that are predicted are not excited in practice - although this part is a bit foggy on a fast read - probably, or I suspect, there is some background information that is "known to those proficient in the art" that explains this, but I don't recognize or know.
So, this gives rise to a number of potential "cures" and questions regarding Dr. Geddes "solution" to the problem.
The first question that comes to mind is why is a big large foam plug required if the finding shows that HOMs occur only back at the exit area of the driver and are unaffected by the use of a horn thereafter. The author states that over 50 horn/driver combinations were tested to verify this result. It would seem to me that a short small foam plug would suffice, and give the same result, without the significant HF rolloff and loss of sensitivity?
Perhaps Dr. Geddes has discovered other HOM modes that extend lower in frequency than those found in the throat area, but he has not yet explicitly shared this finding, afaik.
In the paper there is a somewhat ambiguous and unclear reference to the "internal membrane". I am not sure if the author refers to the diaphragm of the BMS driver (a membrane) or the virtual membrane of the FEM mesh surface being simulated (the planar wave front)?? This makes some difference, since it has different significance depending on if it is the planar wave or the diaphragm itself.
_-_-bear
wxa666 said:
Title of this thread: "How is HOM measured?"
Personally I don't want to measure HOM. I find if all possible steps are taken to minimize diffraction and reflection in the waveguide, at the throat and mouth terminations, and diffraction on the enclosure itself, the sound quality is improved enormously.
I learned this from reading Geddes's work, did some experiments on old, badly designed, horn speakers using foam - very cheap, easy to do. My feeling after this is that Geddes probably has optimally designed speakers. Therefore logically, either buy his speakers or follow them as a DIY model.
However, when I think about measuring HOMs, what I see in my mind's eye: is a vertically suspended wire with a tiny mic capsule that can clip onto precisely marked intervals on the wire; this ensemble sited a foot or two or three in front of a waveguide/horn on a turntable. One could record whatever signals put through the horn and map and process them however they want.
This is definitely in the realm of quick and dirty - the envisaged basic setup may need refinement but might be good enough. It's certainly not beyond means of an enthusiastic DIYer.
wxa666 said:
You said it yourself: horn coloration is in some measure attributable to the presence of HOMs, and while all designs exhibit them by definition, there are some which have more, and some less, and while OS may generate the least, it is apparently possible for OS to perform poorly if the termination is poorly implemented or it is mated with a driver inherently generating high levels or with a sub-optimal transition at the driver/waveguide interface.
Absent a convenient means to objectively quantify HOMs directly, we look to other measureable characteristics for indications of anomalous behavior which might distinguish what is good from what is not so, and with respect to my own work, the particular one you mention has been cited, the presence and magnitude of ripples in the response, an ultimate subjective criterion being, at least in some minds, whether a particular design reduces HOMs (and other anomalies, presumably) to the level of inaudibility in comparison to quality dome tweeters.
Thus, your suggestion that recent improvements in non-OS diffraction waveguide designs may have reduced the practical import of HOMs by reducing them to insignificance in this respect via other means would be consistent with my own view, and apparently that of Geddes as well, that in terms of both quantitative and qualitative outcomes, there exists a continuum, and OS, which we also now recognize has its own set of issues, is not the only worthy approach.
I believe we are in agreement -- not everything besides Geddes waveguides sucks so badly as he and others claim but are unable to prove....
wxa666 said:
Funny bones indeed
– but I don't think you concept applies.If you look at what is written and measured about HOM in ducts (where it in fact makes the most sense for me to refer to it as that specific subset called "HOM") - its *not* exactly at our will what wave front shape we call the fundamental – simply because there is a law involved that allows for higher order modes only up from certain frequencies for given (cross area) dimensions.
If you look at my sims or do your own, you will realise that this basic law applies to horns as well – more complicated due to less uniform shape of boundaries of course.
Also there is a distinct difference to note with overlaid interference patterns (ie "HOM") versus the fundamental - whatever its very shape may be.
All interference products – or better put – the parts that cause the interference with the fundamenta -l do *not* alter the gain of the horn, as all of this diffracted / reflected waves *do not* interact (not strictly true for the diffracted parts that continue in the main direction as – by definition - they lost part of its energy).
They do alter what we measure at a certain point in space though - by mere constructive / destructive interference.
In contrary - different assumptions of the fundamental wave shape *do* alter what's coming out from BEM simus – and if we would have the opportunity to alter the shape of the wave front we inject into the throat at will, we possibly could demonstrate it by measurements as well.
Also read Bjorn's excellent paper on this subject.
This way – as far as I understand it now – the big (several dB) ripples in the gain (acoustic impedance) of a horn we are familiar with are – at first hand - a property of the mouth / throat area and how you like to join 'em (the contour).
If we would like to see the impacts of (more subtle) diffraction / reflection / delay / interference we have to dig some orders of magnitude deeper into impulse response and acoustic impedance measurements IMO
Michael
bear said:
I'd say, diffraction is created at *any* point of the horn contour where the main wave front is *bent any further*.
If this happens (strongly enough to attract our attention) at cross sections larger than the driver / horn joint – well, you get the picture....
(thats why I'm with Earl in this very aspect regarding Makarskis paper)
Hence the whole *HOM* semantic is of little use as it becomes as complicate with and without the use of "HOM" concept - and even less useful, if we look at the "non interaction agreement" of interfering wave fronts.
Michael
Patrick,
"I am going to post plans for a foam plug for the 18 Sound XT1086 this week."
I'm very curious to see how you do this.
I'm making another foam mold, which I'll slice into 1/2" thick layers as patterns to cut out the 1/2" thick foam I have.
But I'm not sure how to cut the layers near the mouth where the angle is very shallow.
"I am going to post plans for a foam plug for the 18 Sound XT1086 this week."
I'm very curious to see how you do this.
I'm making another foam mold, which I'll slice into 1/2" thick layers as patterns to cut out the 1/2" thick foam I have.
But I'm not sure how to cut the layers near the mouth where the angle is very shallow.
noah katz said:
Noah, what kind of foam did you use for your mold? The insulation stuff that comes in a spray can? And what did you use to keep the foam from sticking to the horn?
ThomasW uses an electric kitchen knife for rough shaping and a beard trimmer for the fine work.
doug20 said:
Doug, it's not regular foam, which would block most of the sound. It's reticulated foam that has the most of the web melted or dissolved away. It's used for filters and occasionally for speaker grills. 30 ppi (pores per inch) seems to be what everyone is using.
http://www.swisstropicals.com/Poret Filter Foam Pricelist.html
http://www.foambymail.com/Hi-FlowFoam.html - scroll down to speaker foam
Elias said:
.. as I mentioned it earlier. That simple. The problem is: what is the basic mode?
ZilchLab said:
I disagree. I never said that. The idea of HOM is what makes You nervous, not HOM for itself.
ZilchLab said:
To measure "HIGHER ORDER MODES" it would be convinient to know what is defined (!!!) as the basic mode ...
ZilchLab said:
Yes we do. There are many horns / waveguides around that do very well in any respect even regarding "HOM".
I tried the foam in a can, but two tries failed - big air pockets.
I got some more of the 2-part expanding foam (X-30 from Tap Plastics) that I used the first time and tried again last night; haven't opened it up yet.
I used mold release compound; expensive ($13/can), but really works.
I think I tried Vaseline the first time, or maybe some grease, but it didn't work very well.
I got some more of the 2-part expanding foam (X-30 from Tap Plastics) that I used the first time and tried again last night; haven't opened it up yet.
I used mold release compound; expensive ($13/can), but really works.
I think I tried Vaseline the first time, or maybe some grease, but it didn't work very well.
Hello,
It depends from what distance the measurement is done. You can use plane wave as basic mode if you can accept small curvature error. Depends on what accuracy you are aiming at. Maybe spherical is better as if speaker is a point source.
Also what should be defined is the area samples are taken. Maybe size of a head?
And the grid size, maybe 1mm as wavelength at 20kHz is 17mm.
Well, that's awful lot of data already..
Honestly, I don't think anyone will measure HOM using this method. However, let me be positively surprised!
- Elias
It depends from what distance the measurement is done. You can use plane wave as basic mode if you can accept small curvature error. Depends on what accuracy you are aiming at. Maybe spherical is better as if speaker is a point source.
Also what should be defined is the area samples are taken. Maybe size of a head?
And the grid size, maybe 1mm as wavelength at 20kHz is 17mm.
Well, that's awful lot of data already..
Honestly, I don't think anyone will measure HOM using this method. However, let me be positively surprised!
- Elias
wxa666 said:
mige0 said:
I'd say, diffraction is created at *any* point of the horn contour where the main wave front is *bent any further*.
If this happens (strongly enough to attract our attention) at cross sections larger than the driver / horn joint – well, you get the picture....
(thats why I'm with Earl in this very aspect regarding Makarskis paper)
Hence the whole *HOM* semantic is of little use as it becomes as complicate with and without the use of "HOM" concept - and even less useful, if we look at the "non interaction agreement" of interfering wave fronts.
Michael
Perhaps I am not understanding what you intend to say.
As far as I know the mere bending of the horn contour does not in and of itself create "diffraction", if the bend is for example the exponential bend of an exponential horn, or even the less regular bend of the Geddes horn that's not diffraction in the usual sense of the word.
This would render the conic horn HOM free?
Or are all expansions creating "diffraction" in your world view??
A sharp discontinuity is generally considered to create diffraction... as in "diffraction lens" or "diffraction slot".
The bending of a plane wave to resemble a spherical wave does not in and of itself require that other energy (HOMs) due to reflections be incorporated into the wavefront.
IF the suggestion is that along the entire (or a significant portion) of the horn's length there are modes of reflection that have significant impact and intersection upon the wavefront, at multiples of the fundamental frequency, afaik this has yet to be established or perhaps even hypothesized by any researchers, or via other empirical measurements. (doesn't mean that it hasn't)
I'm all "ears" wrt this...
_-_-bear
noah katz said:
My guess would be that it would be a signal generated by a driver that presents a planar wavefront at the exit of the driver's thoat.
I would expect that interesting frequencies might be where the diameter of the exit corresponds to 4, 2, 1, 1/2 and 1/4 wavelength. Although this does not seem to be a condsideration in the Markarskis paper's methodology.
_-_-bear
ZilchLab said:
wxa666 said:
O.K., you said that others said it, then, Geddes presumably, as he continues to say it today, absent means of demonstrating HOMs other than via subjective evaluations of the result of theoretically minimizing them:
wxa666 said:
Are you suggesting that the relationship is fictitious, or merely itself indeterminate?
bear,
actually it does. Basically, any horn/waveguide with a mouth creates higher order modes. Even Geddes's OS waveguide produces HOMs but, he calculates, it produces the least amount of HOMs.
See this:
http://www.gedlee.com/downloads/Appendix_A.pdf
It's an appendix to a chapter where the heavy lifting is done. But you're a smart guy and can get the drift, I'm sure.
Geddes's waveguide papers used to be on a German DIY site
'Acoustic Waveguide Theory' AES 1989
'Acoustic Waveguide Theory Revisited' 1993
YGM
actually it does. Basically, any horn/waveguide with a mouth creates higher order modes. Even Geddes's OS waveguide produces HOMs but, he calculates, it produces the least amount of HOMs.
See this:
http://www.gedlee.com/downloads/Appendix_A.pdf
It's an appendix to a chapter where the heavy lifting is done. But you're a smart guy and can get the drift, I'm sure.
Geddes's waveguide papers used to be on a German DIY site
'Acoustic Waveguide Theory' AES 1989
'Acoustic Waveguide Theory Revisited' 1993
YGM
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.