I still do like the look and feel of my full paper set (except for #17 of course 😎). You still got yours, Scott?
We are talking about microturbulence. Direct dome output is straight off the end of face of the dome. The CD is squirted from the diaphragm edges inwards in most cases, through a compression along a phase plug. It is a major difference and is likely to produce an audible affect the sound.
Daiphragm friction at low levels and air friction at high levels.
CD,s are recognised with a different sound signature.
I want like you, the right answers. I do not agree with the simplistic denial. Loss of heat from the direct operation of the CD is waste energy, not just from the heating of the voice coil, but air friction especially with the higher frequencies. At high outputs both contribute. You can hear this when a CD is being pushed. Something must be going on. I would never use a CD for lower sound levels that direct drivers perhaps with a horn can achieve. Put up a proper scientific argument.
To quantify the comparisons an experiment could be set up. Did Earl do this with say a 4" driver like a Radian 950 into a log phase plug into 1 1/2" throat. How did he do it. Can you correlate audio energy output versus total energy going into the voice coil. MIT or another esteemed laboratory could set up and do all this.
Short waveguides do not sound throaty. Long waveguides/horns do with the sound reflecting across the throat intothe transition and out. HOMS
You and I know the throat is a problem, hence your foam idea. I used to use a bit of cotton wool in the throat. But it took ouit too much treble information
The point of it all is not us performing and arguing, but merely to say CD's do not do everything well. Anyone who starts with CD's will never know what a direct driver can deliver in months of varied listening. Remember Ariel is direct drivers, and Lynn came to horns later on with a fresh mind.
Are you saying there is no place for the direct driver? for audiophiles. What place would you put them in?
What matters is does one use direct drivers esp for tweeter or not. It should be calculated and tested for ones own personal application.
Last edited:
I heard the Summa's a few years back at the RMAF shows, with Earl's personal choice of equipment driving the speaker, from the CD player to the receiver used for the demo. Rooms at the RMAF aren't cheap; at the time, they probably cost at least $1500 for the duration of the show.
Even at that price, the room was very small, and that's something he couldn't control. The big rooms at the RMAF show can cost as much as $5000 and have to be reserved at least a year in advance. That said, I'm pretty sure what I heard was a fair representation of the intended GedLee sound, with very low coloration and studio-monitor dynamics.
The passive crossover in the Summa, to the best of my knowledge, is complex and the driving amplifier should have a low source impedance (a fraction of an ohm). This rules out the entire family of zero-feedback DHT-triode amplifiers. From a brief personal conversation at the show, Dr. Geddes personally favors selected transistor amplifiers, and if that's the best match for the Summa, I fully respect his choice. He's been designing loudspeakers for a long time and knows what he's doing.
That said, I'm not switching over to transistor amps except for casual listening of iTunes downloads (at 256 kbps) or 5.1 lossy-compression movie soundtracks. So whatever speaker I build is going to have to work with DHT-triode amplifiers. This means no DSP, a relatively simple crossover optimized for an amplifier output impedance around 1~1.5 ohm, and efficiency above 95 dB/meter. Aside from subjective tuning, those are the technical requirements.
What it be possible to re-design a GedLee speaker to work with these archaic amplifiers? Maybe. I'm not the person to do that. If Dr. Geddes recommends certain amps for his speakers, that's a decision that should be respected. The amp/speaker interface is a critical portion of the overall system design.
There aren't a lot of loudspeakers suitable for DHT-triode amplifiers. There's the Fostex, Lowther, and AER full-rangers, but that's not for me. Not interested in restoring an Altec Model 19, nor rebuilding a Belle Klipsch, and Tannoy isn't interested in selling raw drivers to the DIY market. There's also physical restrictions in how big (and heavy) the speaker system can be.
I hope this answers a few questions.
Last edited:
Hi Lynn,
I saw this post by an NS15 owner.. He has a pair of the late Allen Wright's PP 300B differential amps in the backgound. I'm not sure if he uses then with the NS15's. So feedback from him would be interesting if he drives them with DHT's:
http://www.diyaudio.com/forums/gedlee/257185-my-new-new-summa-speakers.html
A couple of comments here. First there are many different types of dome tweeters, soft dome or hard domes, inverted or not, silk or polyester materials, Al, Ti and Be hard domes. They all act differently but all have similarities. Every dome has some form of nodal breakups, it is a natural phenomena of how a dome work. Then you can add a phase correction plug in front of the dome. There is no general dome sound, they are all different. At the same time they do not require a horn for loading but have much lower transfer efficiency than any compression driver.
Now do I think that compression drivers are superior in every way, no, not at all. Compression drivers also have many issues, most of us know that, there basic superiority is in transfer efficiency. They also fall down in many ways. Falling high frequency output, no matter the diaphragm or throat size. Many compression drivers start to roll off hard above 12Khz and many can never produce the top octave well. Horn design is another issue. do you use a radial type, a round horn, a conic sectional horn or a hyperbolic or exponential expansion. They all have distinct sounds, there is no perfect solution.
Neither idea is perfect, there are compromises in both types of devices. For normal sized rooms with a more reasonable sized enclosure I would personally pick the dome tweeter, hands down I would pick that for best sound without having to go to great efforts to get a balanced sound.
Now to Lynn's argument that a Summa speaker should only be matched with a SS amplifier I don't follow that argument. The Summa is a high efficiency design so a lower powered amp should work, it doesn't take a lot of power to drive those speakers. The network shouldn't create such a problem that a well designed vacuum tube amp shouldn't work. I don't see any reason that with output transformers you couldn't match those to a Summa type speaker, that makes little sense to me. You can pick your output impedance with the output transformer so I don't understand that argument. Feedback or not in an amplifier shouldn't make the difference, what I am hearing is that really the non feedback amplifier is not very forgiving and why many have gone away from that for so long now. it is a very touchy type of amplifier with a very limited usability in most audio systems.
The Beyond the Ariel speaker and matching electronics are a very specific system designed concept, mixing and matching with other electronics can completely undo the synergy that has been developed, this is really a highly evolved and specific system that will fall apart when changes are made that don't match Lynn's design requirements. This is not a generic design, there is little room for modification if the original goals are to be met. Specific drivers must be used and exact electronics including tube type are required. Deviations would more than likely produce inferior results.
It would be like putting a brand new Corvette Z07 supercharged engine in a Ferrari. Yes it would work and the car would still be fast, but the end result wouldn't be what a Ferrari owner would expect!
Now do I think that compression drivers are superior in every way, no, not at all. Compression drivers also have many issues, most of us know that, there basic superiority is in transfer efficiency. They also fall down in many ways. Falling high frequency output, no matter the diaphragm or throat size. Many compression drivers start to roll off hard above 12Khz and many can never produce the top octave well. Horn design is another issue. do you use a radial type, a round horn, a conic sectional horn or a hyperbolic or exponential expansion. They all have distinct sounds, there is no perfect solution.
Neither idea is perfect, there are compromises in both types of devices. For normal sized rooms with a more reasonable sized enclosure I would personally pick the dome tweeter, hands down I would pick that for best sound without having to go to great efforts to get a balanced sound.
Now to Lynn's argument that a Summa speaker should only be matched with a SS amplifier I don't follow that argument. The Summa is a high efficiency design so a lower powered amp should work, it doesn't take a lot of power to drive those speakers. The network shouldn't create such a problem that a well designed vacuum tube amp shouldn't work. I don't see any reason that with output transformers you couldn't match those to a Summa type speaker, that makes little sense to me. You can pick your output impedance with the output transformer so I don't understand that argument. Feedback or not in an amplifier shouldn't make the difference, what I am hearing is that really the non feedback amplifier is not very forgiving and why many have gone away from that for so long now. it is a very touchy type of amplifier with a very limited usability in most audio systems.
The Beyond the Ariel speaker and matching electronics are a very specific system designed concept, mixing and matching with other electronics can completely undo the synergy that has been developed, this is really a highly evolved and specific system that will fall apart when changes are made that don't match Lynn's design requirements. This is not a generic design, there is little room for modification if the original goals are to be met. Specific drivers must be used and exact electronics including tube type are required. Deviations would more than likely produce inferior results.
It would be like putting a brand new Corvette Z07 supercharged engine in a Ferrari. Yes it would work and the car would still be fast, but the end result wouldn't be what a Ferrari owner would expect!
I thought that the discussion was about nonlinear distortion. Linear distortion is always quite audible.
Most loading in a CD comes from the compression ratio, but the waveguide does have some loading effect, especially one of finite length. As the waveguide gets shorter and shorter its load becomes more and more that of a truncated tube and not that of the theoretical infinite horn. So mounting a CD in a baffle is going to be very different than mounting it on a horn - mostly because of the short length rather than the flare. The directivity response will change completely however, even if somehow you found a way to negate the termination reflections.
I love these kinds of claims since there are no Summas in another country. If you remember correctly the poster then corrected to say that the speakers were Summa copies, not built by me - well that's not really a very good test now is it!?
Bill, you can of course put a CD on a baffle but what you get is not going to look anything like what you would get if it were on a waveguide. The waveguide is essential to its operation even if the load aspects are not. In waveguides the response will change less and less as the control angles narrows. It will be greatest at 180 degrees and above about 30 degrees will hardly change at all. 45-30 will have only a small change, 90-45 will be enormous. So sure you can test at 180 but I am not sure what the results would mean.
The tweeter does not have a compression ratio and hence a phase plug.
You don't completely understand that HOMS lower in frequency the wider the waveguide. SO for a given bandwidth a wider waveguide will have more HOMs than a narrow one.
If the attenuation is a problem then EQ it - that's what I do.
Direct radiators can never be constant directivity.
The crossover is designed for a low impedance amp. If the amp has a higher impedance then the crossover needs to change or the response flatness will suffer. I can and have done this, but not all amps with higher output Z have the same Z - so this gets to be a problem. I recommend no particular amp, but I do make clear that the crossover assumes a low Z in series. If the customer knows the impedance that they will use I can adapt the crossover accordingly. Its not rocket science.
Last edited:
PS. If you remove the phase plug from a compression driver, it becomes a dome tweeter. It is the compression ratio that creates the higher efficiency, but then this requires a phase plug. The phase plug then requires a waveguide to work right. Fortunately along the way one can achieve constant directivity, which, to me, is what it is all about - not efficiency.
And therein lies the question. What type of directivity or dispersion a person prefers. Some want extremely narrow dispersion and the head in a vice sweet spot, some want a wider window, so that many sitting on a wide couch can hear basically the same things in every position and then you have those who choose the omnidirectional sound of an MBL type of speaker where the room itself sets the stage. So there is no single answer, it is a personal choice. No one solution will fill all the desires, there are no right answers, only preferences.
ps. No need to remove the phase plug, just turn the CD around, take off the back cover and figure out how you want to tackle the phase plug and throat as the back chamber. Seen this done over 30 years ago in a studio application, nothing new there.
ps. No need to remove the phase plug, just turn the CD around, take off the back cover and figure out how you want to tackle the phase plug and throat as the back chamber. Seen this done over 30 years ago in a studio application, nothing new there.
Last edited:
As I said above, it doesn’t make sense to talk about turbulence micro or otherwise in conjunction with an oscillating motion. As Dr. Geddes correctly pointed out, vortices need time to develop. We’re not dealing with a linear flow at some velocity such as for example water in a pipe, where an obstruction in the stream can set up vortices. This is the propagation of oscillating motion in air molecules and ‘disturbances’ here are called diffraction waves. Micro-turbulence? That’s called Brownian motion of the air particles and if it would ever be a problem with CD’s it would certainly be so with dome tweeters in equal measure - just as is the case with diffraction.
Furthermore, If turbulence could be set up, it would have to be supported inside the space of the compression chamber and inside the phase plug, whose dimensions are in the range of 1 to 2 mm. The wave length at 5 KHz is 69 mm. What effect would a 1 mm vortex have to a wave length 69 times larger? I’ll let you figure that one out. [hint: a wavelength of 1 mm is 344 KHz]
You mention losses into heat in CD’s. This is at best an unsubstantiated claim. Collisions of air molecules with hard surfaces inside the driver are elastic and there are no lossy/absorptive materials to turn motion into heat. And please don’t come back with claims about skin friction and such, since again this is not linear flow and there aren’t any shear forces involved.
Let’s talk about wasted energy in CD’s. This is truly erroneous! Here, we have a driver type that on average is 30% efficient. IOW It turns 30% of the electrical energy into acoustical energy. The dome tweeter is at best 3% efficient. Do you really want to talk about losses and wasted energy in this regard?! – really?!!!
As for your claim on the voice coil heating losses, here again the thermal capacity of the CD’s coil, is an order of magnitude greater than that of a dome – after all, they’re built to support very high SPL’s and the accompanying heat issues. There was a thread here a few years back started by Dr. Geddes on the perception of dynamics in loudspeakers. There was much discussion on the effects of heat in voice coils and such. I suggest you find it and read it, it’s quite educational.
It seems you like dome tweeters. No one can argue with your perception. Enjoying what you like and thinking that it’s the best for you is one thing. Here however, you’re trying to reconcile your belief in the superiority of the dome with unsubstantiated claims in physics that frankly don’t make much sense. When your postulations are confronted, you claim ‘simplistic denial’. Unfortunately and as was said earlier, the burden of proof lies with the claim and you’ve made quite a few of them. I don’t mean to be confrontational and there’s no offense intended, but you don’t seem to have a good grasp of physics, at least not on this topic. Had you simply stated that domes sound better to you… well, what could I or anyone else say to that? Not much indeed!
Take care and enjoy the music.
For the Fun : And now they scan the turbulences of your own ear-horn :
https://www.youtube.com/watch?v=7iXfdWGDN7I
We should call it the Van Gogh's !
They are funny at AKG, I hope Quincy Jones at 82 ans can have ear correction yet and any tinitus !
Is it possible to have some passive horns on the head for each ear to correct a room response and have the ideal FRs like time domain reverb correction with it ? a sort of short horn but with scientific concepts : https://www.google.fr/search?q=ear+...oTCILVwMDUyccCFcSJGgodAwoMhA&biw=1391&bih=649
Or open foam plugs for ears
... Sorry I 'm totally lost in the difficult concepts of this thread but we start from a couple of low efficienty THD amp and high efficienty (>94 db) speaker.
https://www.youtube.com/watch?v=7iXfdWGDN7I
We should call it the Van Gogh's !
They are funny at AKG, I hope Quincy Jones at 82 ans can have ear correction yet and any tinitus !
Is it possible to have some passive horns on the head for each ear to correct a room response and have the ideal FRs like time domain reverb correction with it ? a sort of short horn but with scientific concepts : https://www.google.fr/search?q=ear+...oTCILVwMDUyccCFcSJGgodAwoMhA&biw=1391&bih=649
Or open foam plugs for ears
... Sorry I 'm totally lost in the difficult concepts of this thread but we start from a couple of low efficienty THD amp and high efficienty (>94 db) speaker.
Last edited:
You will have done all the necessary finite element modelling I guess. All horn profiles have their merits and demerits. I favour JMIC and OS round horns.
I read some misundertanding into some of the above. You did a good job on Summa, really good. Why not leave it there please. My SS amp would drive your Summa to perfection ,but I am too far away.
Yup, but sadly they are boxed in storage (..and have been for quite some time).
I'm not sure I have every issue though. (..wasn't 17 the online issue with Thomas Mayer's 3-Phase power supply? - loved that article.)
They came out on the news stand in the mid-90's so sporadically that it was almost like finding treasure - even if it was like 7 or 8 US for a magazine (..which at that time was mostly absurd).
Anyway,
Speaking of Thomas Mayer:
VinylSavor
FYI - finite elements are simply one way of doing a numerical simulation. Until recently FEA could not really do a very good job on waveguides. Back when I started studying them some 30 years ago FEA was in its infancy (in fact I was one of the first people to use FEA for acoustics - on my thesis - but the frequency range was limited to a few hundred Hz.) I had to develop numerical techniques that could span the frequency range required and this precluded any point allocation methods. Only a continuous closed form solution would have ever sufficed.
Today with a mega-buck FEA package and a very fast multiprocessor computer one just might be able to do an adequate job on a waveguide design. But then you would likely just come up with what is already know from other techniques.
Today with a mega-buck FEA package and a very fast multiprocessor computer one just might be able to do an adequate job on a waveguide design. But then you would likely just come up with what is already know from other techniques.
I understand that you cannot attenuate reflections or diffractions electrically but I am asking what do you actually measure?
What do you actually measure that is different than what would be measured if I used just a simply RC network?
The Golden Age amplifiers of the Fifties were all feedback amplifiers, with 20 dB of global NFB a typical figure. They were limited to this figure by phase margin considerations as a result VLF and VHF phase rotation in the coupling caps and the output transformer. A few amplifiers like the McIntosh had additional local loops to linearize the Class AB region ... McIntosh chose an operating point with a relatively small A region so the amp could provide more power than competitors. The Citation II was another example of a design with multiple feedback loops. The drawback with multiple feedback loops is a potential decrease in phase margin and a considerable increase in settling time; this was not that well understood in the Fifties.
Audio Research and other latter-day high-end tube amp manufacturers basically made Fifties-style amplifiers (PP pentode with global feedback) and added transistor-regulated power supplies. In the USA and UK, things stayed this way until the DHT people discovered what Jean Hiraga had been doing in Japan and France. This then set off a culture-war between the high-end mainstream (Audio Research et al) and the newcomers, which still echoes through today. Zero-feedback DHT is very much a niche product, separate and apart from the mainstream of high-end audio, never mind the iTunes-downloading mass-market.
When 20 dB of global NFB is applied, the output impedance of a tube amp is reduced to a fraction of an ohm, and suitable for any loudspeaker, aside from power considerations. With zero feedback, it rises to 1 or 2 ohms, and is not "current drive" in any useful sense. It's just a little higher. As Dr. Geddes mentions, it requires only minor changes in the crossover and the T/S alignment of the woofer. It's not current drive, which requires a radically different crossover, and makes any T/S alignment problematic.
Although most vacuum-tube amplifiers are transformer-coupled, raising the primary impedance isn't very useful for reducing the Zout, because so much power is then thrown away. That wouldn't be a big deal if the amplifier had 200 watts or more, but 200 watts is a huge amount of power for a tube amp, never mind one operating in Class A. Paralleling tubes has the awkward effect of moderately increasing high-order distortion, since the transfer curves are all slightly different and don't quite match, and stability considerations can also arise. (Pentodes can oscillate in the RF range if you're not careful when paralleling them. There are several commercial amplifiers with known problems in the 1~10 MHz region.)
The net result is that Class A tube amps are restricted to about 20~35 watts if PP, and 10 watts if single-ended. Transmitter tubes double or even triple this, but then you're building 1~2 kV supplies, and the output transformer becomes very hard to build. (Peak voltages in the output transformer windings are several times the quiescent value.)
The "classical" PP-pentode-with-feedback amps behave and measure very similarly; this reflects their similar architecture, and isn't surprising. DHT amps are all over the place sonically, partly because of their moderately high output impedance, but more so because the lack of feedback exposes the sonics of the underlying architecture and tube selection. Different tubes have different distortion spectra, and this is audible in a non-feedback circuit.
The typical 20 dB of feedback knocks down all the distortion harmonics, but it also creates new higher-order harmonics (but at very low levels) as a result of recirculation and mixing from output to input. The back-EMF currents from the loudspeaker also interact with the phase margin of the amplifier (at low and high frequencies). An amplifier that is stable into a resistive load may lose some of that stability when connected to a reactive loudspeaker load; if that happens, the recovery time from slewing events is increased.
Amplifiers with similar distortion figures (but different topologies) can sound very different, and additional differences can appear with different types of loudspeakers. Amplifier colorations can also mimic speaker colorations, and it takes a fair amount of swapping around (and measuring) to pin down the real source of a coloration.
To recap, the speakers I'm interested in are aimed at the Sound Practices crowd, who are proponents of the zero-feedback DHT approach. People who enjoy PP pentode-with-feedback Class AB amplifiers have a broader choice, and people who like Class AB or Class D transistor amps can build pretty much any speaker they want. The choice of amplifier sets the degree-of-freedom in the loudspeaker design.
Last edited:
Lynn,
Not to start an argument but what you have just posted seems to be the major reason that besides a few diehard audiophiles the vacuum tube industry is basically dead. It isn't even that it uses vacuum tubes but the very basic limitation of non feedback topology. We all know that even high efficiency speakers are a rare breed today, size matters to the masses and larger high efficiency speakers just don't fit the bill anymore. So you are in very rarefied company today. Even here in the diy side of things you are a very small minority. Personally i have no interest in class-A operation, it just has to limited a use outside of very few speakers today. Yes I could use something like that with the Altec speakers I have sitting in a large room but even those are so rare today that most people don't even realize that they are speakers.
I commend you on keeping up the good fight, I just think you have little company to keep in this endeavor. I've heard some of the expensive SE amplifiers with 7 watts output and perhaps a bit more and wouldn't give a plugged nickel for that kind of power. There is no way that is going to drive even my high efficiency speakers to any kind of real dynamic level without falling on its face. I've used them to make the audiophiles happy at the old Stereophile shows but strictly to power the high end with compression drivers where that low power was more than enough. I can't imagine using them for anything beyond this. It seems even 50 years ago that is why feedback tube amplifiers were developed, there just isn't a market for anything less. You are part of a very dying breed, I can only imagine what your amplifier would cost on the open market? One of the reasons that the high end audio industry is disappearing today, it is almost a non entity.
Since I have to market into this newer world of consumers I can't even think about a vacuum tube, it is not possible. I have to work with SS and digital make it as good as it can be. anything less than that is just commercial suicide.
Not to start an argument but what you have just posted seems to be the major reason that besides a few diehard audiophiles the vacuum tube industry is basically dead. It isn't even that it uses vacuum tubes but the very basic limitation of non feedback topology. We all know that even high efficiency speakers are a rare breed today, size matters to the masses and larger high efficiency speakers just don't fit the bill anymore. So you are in very rarefied company today. Even here in the diy side of things you are a very small minority. Personally i have no interest in class-A operation, it just has to limited a use outside of very few speakers today. Yes I could use something like that with the Altec speakers I have sitting in a large room but even those are so rare today that most people don't even realize that they are speakers.
I commend you on keeping up the good fight, I just think you have little company to keep in this endeavor. I've heard some of the expensive SE amplifiers with 7 watts output and perhaps a bit more and wouldn't give a plugged nickel for that kind of power. There is no way that is going to drive even my high efficiency speakers to any kind of real dynamic level without falling on its face. I've used them to make the audiophiles happy at the old Stereophile shows but strictly to power the high end with compression drivers where that low power was more than enough. I can't imagine using them for anything beyond this. It seems even 50 years ago that is why feedback tube amplifiers were developed, there just isn't a market for anything less. You are part of a very dying breed, I can only imagine what your amplifier would cost on the open market? One of the reasons that the high end audio industry is disappearing today, it is almost a non entity.
Since I have to market into this newer world of consumers I can't even think about a vacuum tube, it is not possible. I have to work with SS and digital make it as good as it can be. anything less than that is just commercial suicide.
I measure a reduction in the reflections and the diffraction - or is this a trick question?
I showed how delayed aspects of the impulse response were reduced even after the frequency responses were made to be the same. These aspects were almost invisible in the frequency domain but became apparent in the time domain. I did a paper on this at ALMA many years ago. Maybe you should look that up.
An RC network would lower the highs and thus change the whole impulse response - not at all the same thing.
Don't you see that the foam reduces the highs, the diffraction and the reflections. Then EQing the highs brings them back, but the diffraction and reflections stay reduced. What part of this aren't you getting?
Last edited:
Its always fun to have a rational discussion about things that matter, so thanks for asking!