[dlr]

First, I just thought I'd make it clear to others that only your first two responses were replies to my points. The remaining were for Mark, I believe. You forgot to differentiate that.

dlr,

Or am I wrong in my understanding of the material of the phase plug?

It's NOT a phase plug at all. It's just a dust cap. The shape is a little longer (or taller) than most. It yielded the best overall response of all those that were tested. He was just being derogatory. There is a picture of the driver in my avatar.

Ah. I could not see enough detail in any photos to make that clear. I had earlier mentioned that a non-metallic one would still eliminate the air space resonance underneath, with which you agreed (middle of page 3). That implied a true phase plug attached to the pole-piece, therefore my continued assumption that it was a true phase plug. That is in fact not the case, since it's not a phase plug. It would help if you would make corrections such as that at the time, to prevent any further confusion.

With this additional information (not a phase plug), it changes the perspective a bit. Some might think it to be a fake phase plug, since for all intents and purposes it does not operate as a phase plug at all, but certainly has the appearance of being one. I thought it to be one due to the limitations of the photos on-line. My original assumption that it was a real phase plug was wrong, as you very emphatically stated above. Thank you for correcting that mis-perception now. I see John's comment as not being the egregious slight that you took it to be. It's not a real phase plug. Maybe if he had simply said that, rather than use the word fake, you wouldn't have been upset.

So it's as you state, a dust cap. A somewhat unusually shaped one, yes, but it offers no benefits of a real phase plug. The one benefit would apparently be to slightly alter the output at higher frequencies since it is part of the moving mass along with the diaphragm, but its shape may aid in the upper ranges.

Yes, while I agree with you here, what is seen at 5.5K is not a simple "rising response". It is very characteristic of a resonance. I'm not aware of any other way for a very symmetric peak such as that to exist in a raw driver. If you disagree, please explain exactly what creates that peak, specifically in the structure of the driver.

Yes there is a peak there caused by a resonance. But the decay rate of the resonance is very fast and little stored energy is there. The decay rate is faster and looks cleaner than the other two drivers tested and compared to. All looked good, but mine was the one made to look bad in the comments. Again his attitude towards me may have had everything to do with that.

Here's where I have to disagree on the analysis. I see that peak to be the worst of the three (although all three have issues) for two reasons. First, it has the highest Q. The others can be much more easily corrected in the crossover. Once corrected in this fashion, the resonance is no longer an issue, since the stored-energy issue reverts to that of the bandpass of the driver/XO combination. The easier it is to trap, the less complicate the XO may need to be.

Second, any peaks in the raw driver response will tend to magnify all harmonics of any distortion that fall on or near the peak. Motor generated distortion is not controllable by the XO, other than one that may limit excursion at the low end. Any distortion that is generated will be magnified by that peak. It's an undeniable fact of drivers. That's of course why some do not like hard-coned drivers. Cross lower and the linear distortion may be reduced in some ways, but the non-linear motor distortion cannot be. Any music signal in the middle of the passband that has linear harmonic distortions that coincide with the 5.5K peak will be amplified.

If that peak is around 5.5K, then it will coincide with any signal that has any harmonic component that is a multiple, e.g. motor distortion that is at 1833Hz will have a third harmonic component precisely at 5.5K.  Any music that has content at 2750Hz will have some amount of second harmonic distortion precisely at 5.5K.

The other drivers will suffer as well, so it would take some distortion measurements to distinguish which one is better controlled. You can't just look at the FR or CSD to determine this, you know that.

The more this is analyzed, the more I think that picking a fight with John was ill-advised.

[JoshK]

Second, any peaks in the raw driver response will tend to magnify all harmonics of any distortion that fall on or near the peak. Motor generated distortion is not controllable by the XO, other than one that may limit excursion at the low end. Any distortion that is generated will be magnified by that peak. It's an undeniable fact of drivers. That's of course why some do not like hard-coned drivers. Cross lower and the linear distortion may be reduced in some ways, but the non-linear motor distortion cannot be. Any music signal in the middle of the passband that has linear harmonic distortions that coincide with the 5.5K peak will be amplified.

If that peak is around 5.5K, then it will coincide with any signal that has any harmonic component that is a multiple, e.g. motor distortion that is at 1833Hz will have a third harmonic component precisely at 5.5K.  Any music that has content at 2750Hz will have some amount of second harmonic distortion precisely at 5.5K.

Thanks, I did not know that!  That is interesting.  I guess that is why Jon Marsh is so ademit (sic) about HD being >-50db below.

[jon_010101]

I told myself I'd stay out of this one, but ...

Here's where I have to disagree on the analysis. I see that peak to be the worst of the three (although all three have issues) for two reasons. First, it has the highest Q. The others can be much more easily corrected in the crossover. Once corrected in this fashion, the resonance is no longer an issue, since the stored-energy issue reverts to that of the bandpass of the driver/XO combination. The easier it is to trap, the less complicate the XO may need to be.

As far as I can see, that is a lower-Q resonance than those measured in the other drivers (esp. the Vifa's, which goes on for many cycles).

Keep in mind that High Q would imply low damping rate and long decay times.  Q being, specifically, "the number of cycles required for the amplitude of motion to reduce to 1/e^pi of its original value ... indicating that [for a large Q] it takes a longer time for the oscillations to damp out" (Morse, Vibration and Sound, p.25)

 

[dlr]

I told myself I'd stay out of this one, but ...

Here's where I have to disagree on the analysis. I see that peak to be the worst of the three (although all three have issues) for two reasons. First, it has the highest Q. The others can be much more easily corrected in the crossover. Once corrected in this fashion, the resonance is no longer an issue, since the stored-energy issue reverts to that of the bandpass of the driver/XO combination. The easier it is to trap, the less complicate the XO may need to be.

As far as I can see, that is a lower-Q resonance than those measured in the other drivers (esp. the Vifa's, which goes on for many cycles).


Keep in mind that High Q would imply low damping rate and long decay times.  Q being, specifically, "the number of cycles required for the amplitude of motion to reduce to 1/e^pi of its original value ... indicating that [for a large Q] it takes a longer time for the oscillations to damp out" (Morse, Vibration and Sound, p.25)

 

I understand what you're thinking, but in reality those other two have two resonances somewhat closer together. Taken together they can more easily be smoothed than the single Q of the M130. In addition, the M130 also has two closely-space, lower Q resonances at about 8.5K and 9.5K that tend blend, even if the smaller one looks to last a bit longer. My experience has been that it's easier to tame the others. What's important is not the CSD decay, it's what can be done to tame the peak such that a target SPL is achieved. If that can be done, there will not be any indication of a resonance in the CSD after application of the crossover.

They all can result in good lowpasses, I believe, but a resonance as seen in the M130 tend to be a bit harder to control tightly.

One thing I haven't seen anywhere is a set of off-axis response for the three. That would help to show just what the impact is of the resonances.

[Danny Richie]

I disagree dlr, on which is the worse of the two evils.

If the crossover will bury the peak 25db down or so and it has a short decay time (no ringing) then it is tougher to hear. If the peak is lower in magnitude (in level) but the decay rate is longer it will have more of a ring to it even if it is down in output to the same level as the other.

[Danny Richie]

So none of the "experts" wanted to respond to the distortion measurements that I posted? Non-experts? Is Hank the only one? Everyone wants to say how to measure but no one wants to read the measurements?

[konut]

The more this is analyzed, the more I think that picking a fight with John was ill-advised.

 
I try to look at both sides of an arguement/discussion. Here are some questions to ponder. Is there anywhere on the GR Research site that claims a phase plug? How would Mr. Krutke feel about being called a fake speaker designer? Who's front yard is being 'stunk up'?
Mr. Krutke is free to post any information and opinion he chooses. I hope he continues to do this as his measurements are a useful data point to consider. Many have benifited from his site. I've seen him engage in discussions on diyaudio but chooses not to here. Thats fine. Mr. Richie is free to respond to any information he percieves as innaccurate. I hope he continues to do this to clear up any confusion that might occur among his potential customers. Thats fine too. Everyone is free to make up their own minds as to what to conclude. Whats the problem here? My perception is that a lot of people want to think in terms of absolutes. No shades of grey, thank you very much. Thats fine as well.
You characterise this as a 'fight'. I don't see any fight, all I see is a discussion of ideas. We're both entitled to our opinions.

[jon_010101]

I understand what you're thinking, but in reality those other two have two resonances somewhat closer together. Taken together they can more easily be smoothed than the single Q of the M130. In addition, the M130 also has two closely-space, lower Q resonances at about 8.5K and 9.5K that tend blend, even if the smaller one looks to last a bit longer. My experience has been that it's easier to tame the others. What's important is not the CSD decay, it's what can be done to tame the peak such that a target SPL is achieved. If that can be done, there will not be any indication of a resonance in the CSD after application of the crossover.

They all can result in good lowpasses, I believe, but a resonance as seen in the M130 tend to be a bit harder to control tightly.

One thing I haven't seen anywhere is a set of off-axis response for the three. That would help to show just what the impact is of the resonances.

I don't really see how multiple higher frequency resonances are any more desirable than a single one, which will still be nicely above the crossover frequency.  I would agree that all can be dealt with easily.

If crossed over properly, there should be no indication of persistent ringing for any of these drivers in a linear CSD of a completed loudspeaker... but this does not fully preclude them from nonlinear excitation, as you noted previously.  Your criticism pertaining to the harmonic excitation of these resonances appears challenged by the fact that it applies to the other drivers under consideration (and most drivers in existence), which here exhibit even longer decay times.

[markC]

Woofer 2 should sound more pleasing to the ear as odd order distortion is lower in amplitude than woofer 1. What do I win?

[JAD2]

Let me introduce myself.
I AM AVERAGE JOE.
I AM A AUDIOPHILE.


A Average Joe is- An ordinary person, especially a man.
A Audiophile is-  A person having an ardent interest in stereo or high-fidelity sound reproduction.
Nothing more, nothing less, it holds no stature and is being applied.

I really wanted to get that straight since the ideals slung around here and other forums of this nature dont know what either of these are. Everyone likes to sling around their own definitions as to credit or discredit someone, or some ideal.

Now based upon the original topic posted.
Danny has every right to call out Zaph's Audio, why?
Well anywhere you go on this Internet to ask questions, ask for aid/help in DIYing your own speakers, one site/person is "always" suggested as the place/person to go to. It is heavily implied in one form or another that Zaph's is a godly authority of such knowledge. There are others, but no one is suggested as heavily as he is, I know this personally from being the average Joe asking.
If Zaph discredits, slanders, posts false information and so on ,even with little disclaimers given, Danny has every moral and legal right to question him as to the content of what he posts.Oh it isnt pretty, alot of this whole ordeal is childish, but its totally in essence correct him doing so. If Zaph doesnt reply it could be a way to help legally protect what he has done and keep it to a minimum, or, well I wont go there.

Now onto a different topic ideal, this whole measurement deal really to the average Joe, shouldnt mean squat, but you's try to sway people into this being correct.

Now back to the Audiophile thing.
Forgive me, but isnt the ideal to replicate music as it would be live?????
First off, all this measurement data in driver and crossover topology we see being posted all over, is it hear-able to the human ear?????? Dont forget, ALL!!!!!
I got another test for people to consider and it falls back onto the Ideal to replicate music as it would be live.
I did this test myself, its simple.
I plugged my Fender into my Stereo Amp and played it through a set of recommended by Zaph speakers and the set I built that wouldnt pass anyone measuring tests to be recommended to use.

Then plugged the same Fender into its Marshall, that plays through the Peaveys I own.
Want to guess the results????
Remember, the idea is to replicate the sounds as close as possible to the original.

I'll give ya a hint.
The Peaveys also wouldnt pass with your credibility and measurement test either.


If all amplified sounds are played through perfect balanced, prefect measurement curved drivers, with perfectly smooth crossovers, everything I read on all these forums would be true. Other than that, all you doing is correcting to the way YOU LIKE to hear it and not replicating the sound as intended or as close as it should be. Your altering the outcome to YOUR personal preference.

On that note, metal drivers, no matter how well they measure or dont, DO NOT come close to replicating accurately.
Paper ones are much closer!!!!!

[dlr]

I don't really see how multiple higher frequency resonances are any more desirable than a single one, which will still be nicely above the crossover frequency.  I would agree that all can be dealt with easily.

It's not that they're more desireable, only that a single trap that can handle both reasonably well would be easier to do passively, being somewhat equivalent to a broader single resonance. At that point, there will be little left of the resonance.

If a XO is designed that simply pushes the resonance(s) down into the stop-band as Danny suggests and is often done, then the resonances remain, only down in level. I don't think any of us can accurately predict the audibility of any of the cases for this. It would take designing the XO and then measuring the distortion of the result. That is seldom done by anyone, DIY or otherwise, at least as far as what is made publicly available. Raw driver distortion measurements are somewhat limited in their use due to this.

If crossed over properly, there should be no indication of persistent ringing for any of these drivers in a linear CSD of a completed loudspeaker... but this does not fully preclude them from nonlinear excitation, as you noted previously.  Your criticism pertaining to the harmonic excitation of these resonances appears challenged by the fact that it applies to the other drivers under consideration (and most drivers in existence), which here exhibit even longer decay times.

Yes, but the only pertinent issue is what is the distortion is after a properly designed crossover is applied. It a trap is used for the out-of-band, then the linear distortion may indeed be better in one vs. another. However, the only way to know what the situation is for the non-linear distortion aspect is to measure the drivers. All of them. Very little at all can be gleaned from the FR/CSD of any of these with regard to non-linear distortion. In general, though, high magnitude peaks tend to exacerbate it. The Q will affect the result, but only distortion measurements can provide that data. We can only speculate without that.

[Danny Richie]

Woofer 2 should sound more pleasing to the ear as odd order distortion is lower in amplitude than woofer 1. What do I win?

You helped and that is more than I can say for many. Your input is noted, but Jad2 wins the prise.

On that note, metal drivers, no matter how well they measure or dont, DO NOT come close to replicating accurately.
Paper ones are much closer!!!!!

He listened and learned something. 

[JAD2]

Woofer 2 should sound more pleasing to the ear as odd order distortion is lower in amplitude than woofer 1. What do I win?

You helped and that is more than I can say for many. Your input is noted, but Jad2 wins the prise.

On that note, metal drivers, no matter how well they measure or dont, DO NOT come close to replicating accurately.
Paper ones are much closer!!!!!

He listened and learned something. 

Yeah I listened and learned.
I was told metal drivers "IF" implemented right are the cats behind. Thats where I listened part 1.
Part 2 under actually listening, I learned even implemented as per the HT guru's, there harsh, fatiguing at reference levels and beyond and sound more like I put a speaker in a tin can enclosure. But dont tell the HT guru's that and you've decided to build speakers out of junk that I was told to throw out, ebay or the likes, because that gets you banned!!!!

[tubepudz]

For my own enlightenment. If one listens to music at 86 dB (example of something reasonable, every day use), how low does a distortion problem need to be before it is inaudible.  Is there a ratio (HD/playing level) or fixed value (20 dB, 30 dB, etc)?

If there is a performance issue and it is understood, does not change in time, or does not change as a function of operational level then usually life is ok.

[Ettepet]

Several hours ago I thought this topic should have been kindly but firmly brought to a halt.  Now it seems a whole new angle on the matter has been found.

[Danny Richie]

Just one correction from an earlier post. My 260 pound ex-business partner contacted me about the jumping up and down on the woofer frame. He says it didn't even flex.

So I took an old frame off the shelf and tried stopping on it myself. I don't weigh as much as Jeff did but it didn't move.

So then I jacked my car up and set the back wheel down on top of it. It not only supported the weight of my car but it didn't even flex. Afterward it still looked fine. No stress marks or anything. Now the voice coil former shoved the cone forward and smashed the cone, but the frame still looks great. 

[Ettepet]

This should be part of standard testing procedures!

[ooheadsoo]

I plugged my Fender into my Stereo Amp and played it through a set of recommended by Zaph speakers and the set I built that wouldnt pass anyone measuring tests to be recommended to use.

Then plugged the same Fender into its Marshall, that plays through the Peaveys I own.
Want to guess the results????
Remember, the idea is to replicate the sounds as close as possible to the original.

I'll give ya a hint.
The Peaveys also wouldnt pass with your credibility and measurement test either.


If all amplified sounds are played through perfect balanced, prefect measurement curved drivers, with perfectly smooth crossovers, everything I read on all these forums would be true. Other than that, all you doing is correcting to the way YOU LIKE to hear it and not replicating the sound as intended or as close as it should be. Your altering the outcome to YOUR personal preference.

On that note, metal drivers, no matter how well they measure or dont, DO NOT come close to replicating accurately.
Paper ones are much closer!!!!!

Big problem here.  When you play your guitar, you're trying to create sound, not recreate.  Try this one on for size:

1) Plug your fender into your marshall/peavey stack
2) mic your stack and record it
3) rebroadcast the recording from your marshall/peavey stack and compare that to the sound being rebroadcast from one of krutke's or Danny's speakers or whatever.

The point is that you're not supposed to play your fender through gear designed for reproduction vs. production.  They have completely different design goals.  This says nothing about your preference for metal or paper drivers.

[pmel]

I have a couple of anechoic chamber based questions.

Just so you know where I am coming from I will mention that I work in the EMC testign feild for a living.  I work up aroud 2- 50 GHz though so some of my knowledge could be inncorrect for the audio frequencies.

I have never been in a chamber that clamed to meet any estabished quidlines for testing below 100 Hz.  The reason for this is due to the wave lengths just get to long to continue to build things big enough to minimze reflections well enough to meet minimum testing regulations. 

For example at 100 Hz the wave lenght if around 12 feet I believe.  The reflective absorbers, or cones would also have to be around 36 inches long.  The last chamber I was in that could go down to 100 Hz was 20X34X40 ft if I rember correctly.  (these nubers are all approx, I am at home, and don't have access to any of my reference material on this.  Should be pretty close though. 

If we take that all the way down the audio spectrum to lets say 30 Hz these numbers are ridiculously high.  The wave lenght is now around 36 feet, the cones need to be over 20FT long, and the chamber woudl be big enough to park a 747 in.  I have only seen 1 fully functional chamber this big in my life, and it did not even meet minimum specs under 50Hz do to a poor room design.

So here are the questions.

Are you speaker measuring guys possibly talking about "semi", or "hemi" anechoic chambers?

Are the the specs for audio work that different from the EMC world that you can build the chambers smaller?  for instance I have to be able to prove absorbstion at freq of interest below 99.%, if it was only 70% for audio the size of the room, and cones could shrink greatly and remain somewhat accurate.

Or are you doing the old if you ain't got enough room find a deserted runway, and a flag pole trick?

This is for learning sake only.  I ain't trying to start another war.  I have been contemplating a very simple measurements setup for help in some of my own designs, and about half of the people I wanna ask are in this thread.


*edit, just wanna say, I like your car crush testing Danny nice way of using what you got on hand.*

[PaulHilgeman]

For example at 100 Hz the wave lenght if around 12 feet I believe.  The reflective absorbers, or cones would also have to be around 36 inches long.  The last chamber I was in that could go down to 100 Hz was 20X34X40 ft if I rember correctly.  (these nubers are all approx, I am at home, and don't have access to any of my reference material on this.  Should be pretty close though. 

Thats about correct.  The good thing though is that they are actually very well behaved as frequency goes down, so useful data interpretation is possible, but requires certain specified set-ups.

The old runway and flag pole trick is what I use and find to be very effective.  There are actually some advantages to this, as I know that there is no reflection at all.  Acoustic foam always reflects something!!!

-Paul

P.S.  Danny I understand if you delete my post, no problemo!