Re: Dr Geddes,
I'm not sure, that comment comes from my basic understanding of line sources etc. In odd dimensions (1, 3 5) the impulse response from a source is compact, while in even dimensions (2, 4) it has a tail. This comes from the basic math. A line source is basically a two dimensional source. (Interesting, this is why gravity waves - 4 dimensions - have to have a tail!) Its easy to see why in the case of a line source (not so easy for gravity!!). At any point in space, the contribution from parts of the source further and further away have to arrive at later and later times. The longer the line, the longer the tail. So even a flat piston has a small tail. However, put this flat piston on a waveguide and it will not have a tail anymore, because it will appear to have been created from a point source.
You can shade the line array and elliminate the tail (probably on along a line however), but that takes complex shading (time delays), amplitude shading alone will not do it.
There is a lot you can get from the basic physics of a situation.
chrismercurio said:
I'm not sure, that comment comes from my basic understanding of line sources etc. In odd dimensions (1, 3 5) the impulse response from a source is compact, while in even dimensions (2, 4) it has a tail. This comes from the basic math. A line source is basically a two dimensional source. (Interesting, this is why gravity waves - 4 dimensions - have to have a tail!) Its easy to see why in the case of a line source (not so easy for gravity!!). At any point in space, the contribution from parts of the source further and further away have to arrive at later and later times. The longer the line, the longer the tail. So even a flat piston has a small tail. However, put this flat piston on a waveguide and it will not have a tail anymore, because it will appear to have been created from a point source.
You can shade the line array and elliminate the tail (probably on along a line however), but that takes complex shading (time delays), amplitude shading alone will not do it.
There is a lot you can get from the basic physics of a situation.
Re: That was fast!
Marketing of the designs has been a disaster - probably comes from my vast experince in this subject
. There simply is no way that I could make a viable product out of these designs myself. Thats why I have given up. The only Summa's are in homes and I can't give out those names for fairness.
The "big guys" have created a situation where a speaker that is not heavily marketed has no chance in the marketplace. This keeps the small intruder out of the picture. Its a very smart move on their part. Devastating for the consumer who wants a real option in the marketplace however.
chrismercurio said:
Marketing of the designs has been a disaster - probably comes from my vast experince in this subject
. There simply is no way that I could make a viable product out of these designs myself. Thats why I have given up. The only Summa's are in homes and I can't give out those names for fairness.The "big guys" have created a situation where a speaker that is not heavily marketed has no chance in the marketplace. This keeps the small intruder out of the picture. Its a very smart move on their part. Devastating for the consumer who wants a real option in the marketplace however.
Troels JA8008
Here's another candidate for the midbass: the JA8008 HES
Troels Gravesen who you probably know from his excellent website:
http://www.troelsgravesen.dk/Diy_Loudspeaker_Projects.htm
has developed a high efficiency driver: the JA8008 HES
Specs can be found here:
http://jantzen-audio.com/download/PD_JA8008_Oct07_v1-f.pdf
It seems to be an excellent driver according to this enthusiast review from someone who has used the JA8008 HES: http://www.troelsgravesen.dk/JA8008_response.htm
Some more info: http://www.troelsgravesen.dk/JA8008.htm
Here's another candidate for the midbass: the JA8008 HES
Troels Gravesen who you probably know from his excellent website:
http://www.troelsgravesen.dk/Diy_Loudspeaker_Projects.htm
has developed a high efficiency driver: the JA8008 HES
Specs can be found here:
http://jantzen-audio.com/download/PD_JA8008_Oct07_v1-f.pdf
It seems to be an excellent driver according to this enthusiast review from someone who has used the JA8008 HES: http://www.troelsgravesen.dk/JA8008_response.htm
Some more info: http://www.troelsgravesen.dk/JA8008.htm
gedlee said:
That is why I never recommended a medium format Line Array in a normal venue, where a Nexo Alpha or PS 15 System could do just fine. Nobody ever listened to me against the notion that the all hot Line Array thing was the new Swiss knife. But hey, it looks cooler hanging than stacked, doesn't it?
Re: Re: Re: Dr Geddes,
Actually I didn't discover this. It falls out of the "Green's Function" approach in physics. The Green's Function is basically an impulse response, but unlike an impulse, which is only one dimensional - time - a Green's Function can be defined in any number of dimensions. So my comments came straight out of my "Methods in Theoretical Physics" class at school.
I just sent a pair of waveguides to Germany.
kstrain said:
Actually I didn't discover this. It falls out of the "Green's Function" approach in physics. The Green's Function is basically an impulse response, but unlike an impulse, which is only one dimensional - time - a Green's Function can be defined in any number of dimensions. So my comments came straight out of my "Methods in Theoretical Physics" class at school.
I just sent a pair of waveguides to Germany.
gedlee said:
I completely agree. I don't know why this is controversial - it is obvious from inspection that the multiple arrival times of a vertical line array must be non-coincident, even with theoretically perfect radiators. The larger the vertical array, the more the response at the listening position is spread out as a series of impulses - and must also suffer from comb-filtering for the same reason.
A drawing of the arrival times on a piece of paper reveals that - nothing more than simple geometry. Multipath is troublesome enough to eliminate from diffraction (and internal cabinet reflections) in conventional speakers (horns or direct-radiators), why multiply the problem with a driver array?
At low frequencies (say, below 200 Hz), not much of a problem, since the wavelengths are so much larger than the differential in arrival times. But for frequencies higher than 1 kHz, problems with line arrays in the time, frequency, and spatial domains become much more difficult to resolve.
The only two complete solutions have serious drawbacks:
1) A concave curved array will only be accurate at one distance and one lateral bearing angle (one point in space), and badly in error everywhere else. It also bothers me on an esthetic level - concave sound emitters are extremely rare in the natural world, and I would expect a concave emitter to have a weird, hard-to-describe coloration, or perhaps an odd spatial distortion.
2) The Quad ESL63 solution of a phased-array, similar to military radars, has the drawback of simulating the virtual-image point source via multiple delay lines - this requires many amplifiers or a very complex crossover/delay line. Doing so accurately is a process of successive approximation that trades complexity of implementation against accuracy of pulse response - you could get into a million-dollar loudspeaker if you pursued this to the logical conclusion.
3) A partial solution is an amplitude-shaded line array with the outermost radiators sequentially low-passed - which begs the question of the inherent efficiency of the center radiator, which (by definition) sets the efficiency (and headroom) of the whole array. At this point, we are only one step away from a MTM-array multiway loudspeaker system, and two steps away from a traditional MT-array loudspeaker.
I think where people go astray is thinking the waves from the individual point-source radiators somehow combine into a cylindrical wavefront - but they don't. They pass right through each other. True, at the very lowest frequencies (comparable to the size of the array), there is mutual coupling and a resulting increase in efficiency, but this falls apart as the wavelengths become shorter. What starts out as an efficient loudspeaker at the lowest frequencies become dominated by comb-filtering and time-dispersal at higher frequencies - and these are errors that are degraded by equalization, since EQ spreads out the time-dispersal even more.
Combining high efficiency, ample headroom, and a compact pulse response with rapid time-decay characteristics is a hard nut to crack - the failure of the prosound world to address this over the last sixty years speaks to the scope of the problem.
The last serious attempt in the professional world is probably the Shearer theatre horn of 1935 - a long long time ago. The popularity of tap-dancing in movie soundtracks demanded a solution - and I have to admit getting 1 millisecond precision is pretty impressive for 1935. The ingenuity of the Tannoy Dual Concentric deserves an honorable mention as well - impressive technology considering the WWII date of the design. Since then ... well, the soundtracks are (much) louder, and the mass-market-mandated explosions, car crashes, phaser whizzes, and dinosaur thumps come at you from more directions.
Hello Lynn,
how about the Unity horn? I wouldn't call them point source, but even so it seems they are pretty well regarded for their coherence.
And you've mentioned having some 6nd410 midranges. Did you have the chance to listen them? I am asking because I'd like to mate one with a waveguide but I don't know nothing about the subjective performance of these and I am not affording to experiment right now...
Good luck with your project!
Aside from the technical aspects of the Unity, which theatre loudspeakers or commercially available studio monitors use them? I look at these applications and I see JBL dominating the world market, with a sprinkling of other brands - EAW, Meyersound, Westlake, Dynaudio, Genelec, PMC, ATC, and a variety of others.
Only a few professional products mention the impulse response, such as the MeyerSound X-10, and even then, we don't get to see what it looks like 15 and 30 degrees off-axis. The performance of the constant-directivity waveguide will have a big impact on the off-axis impulse response.
Only a few professional products mention the impulse response, such as the MeyerSound X-10, and even then, we don't get to see what it looks like 15 and 30 degrees off-axis. The performance of the constant-directivity waveguide will have a big impact on the off-axis impulse response.
Lynn Olson said:
Beyma 102Nd (very speedy),
http://www.beyma.de/index.php?id=126
http://www.beyma.de/fileadmin/seiten/download/pdf/Beyma_professional/102Nd.pdf
Beyma 12MW NdE (low distortion, soft roll off),
http://www.beyma.de/index.php?id=179
http://www.beyma.de/fileadmin/seiten/download/pdf/Beyma_professional/12MWNdE.pdf
or probably - if not crossed too low - two vertically stacked Beyma 6G40Nd (excellent directivity)
http://www.beyma.de/index.php?id=120
http://www.beyma.de/fileadmin/seiten/download/pdf/Beyma_professional/6G40ND.pdf
matching the directivity of the Raal if the horn experiment fails for one reason or another
None of them under hung VC anyway.
Low (Doppler-) IM basically is the product of low frequency excursion (big Sd for given LF-SPL) and the frequency span to the point of measurement as outlined by John Kreskovsky earlier - and easy to calculate with the spreadsheet I have put together – not so much brand dependent at the very first glance.
Greetings
Michael
Lynn,
http://www.diyaudio.com/forums/attachment.php?s=&postid=1481781&action=thumb&stamp=1207965148
Are these available somewhere or is there more info on the specifics of these amps.
Also, I am SET 845 fan, does this fit my cord in comparison?
http://www.diyaudio.com/forums/attachment.php?s=&postid=1481781&action=thumb&stamp=1207965148
Are these available somewhere or is there more info on the specifics of these amps.
Also, I am SET 845 fan, does this fit my cord in comparison?
SunRa said:
Hello SunRa,
The 605 Nd is my personal favourite – but is somewhat peaky at the high end and does not have the excellent directivity of the 6G40Nd up to nirvana.
Same for the 18 Sound - with regard to directivity (they also do not supply distortion figures to compare with).
Besides that the 6G40Nd is a very powerful speaker – providing an impressive 110 dB at 1m in a 400mm (16") OB as a single driver down to 250Hz.
Greetings
Michael
Thanks, Michael, for the links to the Beyma drivers. The 10 and 12-inchers look good, with the 12MW Nde looking like an interesting alternative to the Altec 414's - efficient, very smooth response, low distortion, all good things. The 6-incher is rougher than I'm comfortable with, unless the HF notches and peaks are artifacts of test cabinet diffraction.
The major divisions on the Beyma graphs are 5 dB, and the entire FR curve of the Ariel fits inside a 5 dB window, with most of the local deviations being smaller. I'd like the FR of the completed system to fit in a fairly tight window, and with a minimum of notch and shelf filtering.
As for commercial availability of the Karna amplifier, maybe some day, if I could license the design and chase out the remaining technical issues - that alone is a 1-year project or more, and one that I've set aside for now pending the completion of the new speaker system.
845 SET amplifiers have very serious issues with driver-tube distortion. The distortion of the 845 is very low, in the 300B league, but the 120~150 volt bias demands drive voltages twice as large as the 300B, and there's just as much Miller capacitance as a 300B. In other words, a difficult-to-drive, highly reactive load at higher frequencies (potential slewing problems).
This a creates a requirement for a driver section with less distortion than the output tube at grid-drive levels around 200V rms - a small, ultralow distortion Class A power amplifier in its own right. No known RC-coupled driver can accomplish this due to the slope of the load-line and power losses in the plate-load resistor. Choke, transformer, or dynamic loads can do this, but only DHT's achieve the requisite low distortion levels - and it seems pointless to use a low-distortion output tube if the driver distortion swamps it out.
So-called distortion-cancellation schemes using driver tubes with complementary distortion to the output do not account for the upper harmonics, which are the ones that really matter. (2nd harmonic dominates THD measurements but is nearly inaudible. The magnitudes of the upper harmonics - 3rd, 4th, 5th, etc. - are much smaller but are also much more noticeable.)
If you like the sound of the 845 (SE or PP), seek out a design that uses a DHT driver and choke, transformer, or active loads for the driver. No RC-coupling, sorry. It's easy and cheap to do but has 3~5 times the driver-tube distortion, not a small difference. That probably eliminates most commercial 845 amplifiers, and many DIY designs.
The major divisions on the Beyma graphs are 5 dB, and the entire FR curve of the Ariel fits inside a 5 dB window, with most of the local deviations being smaller. I'd like the FR of the completed system to fit in a fairly tight window, and with a minimum of notch and shelf filtering.
As for commercial availability of the Karna amplifier, maybe some day, if I could license the design and chase out the remaining technical issues - that alone is a 1-year project or more, and one that I've set aside for now pending the completion of the new speaker system.
845 SET amplifiers have very serious issues with driver-tube distortion. The distortion of the 845 is very low, in the 300B league, but the 120~150 volt bias demands drive voltages twice as large as the 300B, and there's just as much Miller capacitance as a 300B. In other words, a difficult-to-drive, highly reactive load at higher frequencies (potential slewing problems).
This a creates a requirement for a driver section with less distortion than the output tube at grid-drive levels around 200V rms - a small, ultralow distortion Class A power amplifier in its own right. No known RC-coupled driver can accomplish this due to the slope of the load-line and power losses in the plate-load resistor. Choke, transformer, or dynamic loads can do this, but only DHT's achieve the requisite low distortion levels - and it seems pointless to use a low-distortion output tube if the driver distortion swamps it out.
So-called distortion-cancellation schemes using driver tubes with complementary distortion to the output do not account for the upper harmonics, which are the ones that really matter. (2nd harmonic dominates THD measurements but is nearly inaudible. The magnitudes of the upper harmonics - 3rd, 4th, 5th, etc. - are much smaller but are also much more noticeable.)
If you like the sound of the 845 (SE or PP), seek out a design that uses a DHT driver and choke, transformer, or active loads for the driver. No RC-coupling, sorry. It's easy and cheap to do but has 3~5 times the driver-tube distortion, not a small difference. That probably eliminates most commercial 845 amplifiers, and many DIY designs.
While keeping things off-topic:
Do you mean something like this ?
If memory serves well there are quite a few topologies using 300B as a driver and 845 (or even an 805 !!) as output tupe, in either SET or PP topologies. Alas not many of those have are freely available circuit diagrams ...
Lynn Olson said:
Do you mean something like this ?
If memory serves well there are quite a few topologies using 300B as a driver and 845 (or even an 805 !!) as output tupe, in either SET or PP topologies. Alas not many of those have are freely available circuit diagrams ...
augerpro said:
Hi Augerpro,
Guess I must be the biggest dummy. No problem accessing your photo album but there no description to tell which measurement is for what.
Can see the photo of 6G40Nd but without desc, can't tell what driver is next to it. Is it bec I don't have a login ID to the site?
All the driver tests are listed along the left side of the screen. The rest of the of the stuff on the main page is just my junk.
Sometimes I forget to give the pic a title, but if you hover the mouse cursor over the pic it will show the filename, which describes the measurement. Generic measurment info for all the tests is in the READ THIS FIRST folder.
The driver next to the Beyma in the picture is teh B&C 6NDL38.
Sometimes I forget to give the pic a title, but if you hover the mouse cursor over the pic it will show the filename, which describes the measurement. Generic measurment info for all the tests is in the READ THIS FIRST folder.
The driver next to the Beyma in the picture is teh B&C 6NDL38.