About Caps
I only brought up the cap-coloration issue because it became an annoying factor in the Ariel crossover, requiring a half-dB adjustment because of switching brands of caps (same value, but required a tweeter level adjustment).
Right now I use waxed-paper Jupiter caps for the HP of the Ariel, but I hope the JBL biasing technique works well, allowing the use of good-quality industrial film-and-foil types with the appropriate 0.01uF bypass. I would use more than 9V of bias, though, and plan to use at least 18 or 27V of polarizing voltage (so the zero-crossing is never experienced at any reasonable listening level, and to also get the cap a bit closer to the breakdown voltage).
The Gary Pimm methodology of breaking-in the cap with 30% of the breakdown voltage at a high frequency (100 kHz) for several hours is probably worth trying as well. Our hypothetical model of what's going on with the break-in process is that the manufacturing process impresses a static charge on the film, and swinging the voltage up and down many times at a high rate of change gradually removes this latent polarization. Or something like that.
As for audibility of small stuff, I flip a switch and listen for a while. If there is no difference, then no matter. If the difference is merely a difference, i.e., just perceptible, again, so what. If the difference seems to affect a subjective quality - tonality, vividness, dynamics, intimacy, spatiality, or whatever, then more listening is called for, and at length. Many times the results are inconclusive - better in some ways, worse in others. At that point all it tells me that more work is needed, and the optimum solution has not been attained.
At other times I expect a difference and hear none, or something so small I don't care that much. I'm not very sensitive to absolute phase, for example, while other people will jump out of the their seats and say it's either terrible or wonderful. Not me. I just hear a subtle difference in timbre with certain instruments. With most records, a So What reaction.
Other times I expect to hear no difference and am dismayed to hear something large. Making small adjustments (in the 1% range) to the filament voltage of the Karna amplifiers revealed much more difference than I wanted to accept, since it implied I was going to have to spend a lot of time working on the filament supplies. This really disgusted me, frankly, and made me lose interest in all audio for about three months.
Neither outcome would be anything I predicted. Absolute phase is obvious on the scope, and some golden ears really get all wound up about it. But I just don't notice it that much - certainly not enough I'd bother to go to the effort of flipping the phase switch with every CD or record.
I was expecting some sort of audible effect as the direct-heated triodes gradually dropped in emission, but certainly not the kind of extreme sensitivity I heard. There was nothing in the literature that indicated anything like this existed.
During the rough-in phase of the loudspeaker, I won't be listening to music at all, just measuring with MLS and listening to pink-noise. Very boring and tedious, getting rid of the really big problems, chasing out resonances, and getting the time-domain to behave.
Just getting pink-noise to sound like falling water takes a lot of time - and this simple test for coloration is something most high-end speakers cannot pass. I'm not too sure many of the high-efficiency drivers are going to pass this test, for example, and I draw the line at too much notch-equalization to "rescue" a driver from gross coloration.
I don't even want to listen to music unless the speaker is reasonably flat and free of coloration - this process pretty much takes all whizzer drivers out of contention, and I expect will take a lot of time chasing out horn coloration. Maybe none of the horns will pass. I dunno. There are a lot of open variables here - all I can do to minimize the complexity is try and avoid the known-bad regions for each of the drivers.
I only brought up the cap-coloration issue because it became an annoying factor in the Ariel crossover, requiring a half-dB adjustment because of switching brands of caps (same value, but required a tweeter level adjustment).
Right now I use waxed-paper Jupiter caps for the HP of the Ariel, but I hope the JBL biasing technique works well, allowing the use of good-quality industrial film-and-foil types with the appropriate 0.01uF bypass. I would use more than 9V of bias, though, and plan to use at least 18 or 27V of polarizing voltage (so the zero-crossing is never experienced at any reasonable listening level, and to also get the cap a bit closer to the breakdown voltage).
The Gary Pimm methodology of breaking-in the cap with 30% of the breakdown voltage at a high frequency (100 kHz) for several hours is probably worth trying as well. Our hypothetical model of what's going on with the break-in process is that the manufacturing process impresses a static charge on the film, and swinging the voltage up and down many times at a high rate of change gradually removes this latent polarization. Or something like that.
As for audibility of small stuff, I flip a switch and listen for a while. If there is no difference, then no matter. If the difference is merely a difference, i.e., just perceptible, again, so what. If the difference seems to affect a subjective quality - tonality, vividness, dynamics, intimacy, spatiality, or whatever, then more listening is called for, and at length. Many times the results are inconclusive - better in some ways, worse in others. At that point all it tells me that more work is needed, and the optimum solution has not been attained.
At other times I expect a difference and hear none, or something so small I don't care that much. I'm not very sensitive to absolute phase, for example, while other people will jump out of the their seats and say it's either terrible or wonderful. Not me. I just hear a subtle difference in timbre with certain instruments. With most records, a So What reaction.
Other times I expect to hear no difference and am dismayed to hear something large. Making small adjustments (in the 1% range) to the filament voltage of the Karna amplifiers revealed much more difference than I wanted to accept, since it implied I was going to have to spend a lot of time working on the filament supplies. This really disgusted me, frankly, and made me lose interest in all audio for about three months.
Neither outcome would be anything I predicted. Absolute phase is obvious on the scope, and some golden ears really get all wound up about it. But I just don't notice it that much - certainly not enough I'd bother to go to the effort of flipping the phase switch with every CD or record.
I was expecting some sort of audible effect as the direct-heated triodes gradually dropped in emission, but certainly not the kind of extreme sensitivity I heard. There was nothing in the literature that indicated anything like this existed.
During the rough-in phase of the loudspeaker, I won't be listening to music at all, just measuring with MLS and listening to pink-noise. Very boring and tedious, getting rid of the really big problems, chasing out resonances, and getting the time-domain to behave.
Just getting pink-noise to sound like falling water takes a lot of time - and this simple test for coloration is something most high-end speakers cannot pass. I'm not too sure many of the high-efficiency drivers are going to pass this test, for example, and I draw the line at too much notch-equalization to "rescue" a driver from gross coloration.
I don't even want to listen to music unless the speaker is reasonably flat and free of coloration - this process pretty much takes all whizzer drivers out of contention, and I expect will take a lot of time chasing out horn coloration. Maybe none of the horns will pass. I dunno. There are a lot of open variables here - all I can do to minimize the complexity is try and avoid the known-bad regions for each of the drivers.
Ai (Audio Intelligence) website up; fairly in-depth information on Geddes-designed constant-directivity waveguide speakers including some imho quite impressive off-axis response curves.
http://www.ai-audio.com/home.html
http://www.ai-audio.com/home.html
Misc
Hi Lynn
I'm with you on the filament voltage sensitivity . Getting some amazing improvements in transparency from running the KR 300B XLS filaments at a lower voltage . More experiements on that this weekend .
On bass drivers , not sure if these have been mentioned earlier in the thread, but the BD15 custom drivers Bert Doppenberg developed for his open-baffle speaker options are certainly well worth considering.
http://www.bd-design.nl/contents/en-us/d9_DIY_Drive_Units.html
James Doddington put me onto these the other evening when we were discussing what to use with a pair of Feastrex D5nf's .
I also noticed a link to the rather tasty Duelund capacitors, resistors etc from Bert's site. No idea what these cost of sound like, but the blurb seems interesting .
http://www.bd-design.nl/contents/media/duelund_coherence.jpg
Mark
Hi Lynn
I'm with you on the filament voltage sensitivity . Getting some amazing improvements in transparency from running the KR 300B XLS filaments at a lower voltage . More experiements on that this weekend .
On bass drivers , not sure if these have been mentioned earlier in the thread, but the BD15 custom drivers Bert Doppenberg developed for his open-baffle speaker options are certainly well worth considering.
http://www.bd-design.nl/contents/en-us/d9_DIY_Drive_Units.html
James Doddington put me onto these the other evening when we were discussing what to use with a pair of Feastrex D5nf's .
I also noticed a link to the rather tasty Duelund capacitors, resistors etc from Bert's site. No idea what these cost of sound like, but the blurb seems interesting .
http://www.bd-design.nl/contents/media/duelund_coherence.jpg
Mark
Duelund
They are expensive. Very. Priced in USD...
Duelund VSF
1.0 uF in aluminum = $86.70
1.0 uF in copper = $124.95
1.0 uF in silver = $1026.40
Duelund Cast PIO
1.0 uF in copper = $248.65
1.0 uF in silver = $1128.40
From what I have read they are excellent....but I would hate to think what the average pair of crossovers might cost with 30-40 components total.
C
They are expensive. Very. Priced in USD...
Duelund VSF
1.0 uF in aluminum = $86.70
1.0 uF in copper = $124.95
1.0 uF in silver = $1026.40
Duelund Cast PIO
1.0 uF in copper = $248.65
1.0 uF in silver = $1128.40
From what I have read they are excellent....but I would hate to think what the average pair of crossovers might cost with 30-40 components total.
C
Caps
Thanks Chris for the info .
So , somewhat more expensive than Jensen coppers then . Sort-of AudioNote - style prices .
Luckily I won't be needing 15-20 components per side in any crossover I build . Been there done that . The most impressive speakers I heard recently ( OB ) used two components per side ( one L , one C ) in a series arrangement .
MJ
Thanks Chris for the info .
So , somewhat more expensive than Jensen coppers then . Sort-of AudioNote - style prices .
Luckily I won't be needing 15-20 components per side in any crossover I build . Been there done that . The most impressive speakers I heard recently ( OB ) used two components per side ( one L , one C ) in a series arrangement .
MJ
Duelund
Over at the HT Guide forum some of the more accomplished folks came up with an LCR design utilizing Duelund style filters. These use many components.....but not Duelund as cost was a consideration in the build.
http://htguide.com/forum/showthread.php4?t=27218
Over at the HT Guide forum some of the more accomplished folks came up with an LCR design utilizing Duelund style filters. These use many components.....but not Duelund as cost was a consideration in the build.
http://htguide.com/forum/showthread.php4?t=27218
"Which Roger Waters CDs use Ambisonics?"
Amused to Death
Peter Gabriel
Up
The Waters disk is really good sounding with several wrap around effects. Kind of freaky when you hear stuff behind you and you know it's 2 channel. Great set-up disk. If your phasing is screwed up it all falls apart.
Rob
Amused to Death
Peter Gabriel
Up
The Waters disk is really good sounding with several wrap around effects. Kind of freaky when you hear stuff behind you and you know it's 2 channel. Great set-up disk. If your phasing is screwed up it all falls apart.
Rob
Hi Rob,
great, I will listen to both CDs. Do you know if they are UHJ encoded?
great, I will listen to both CDs. Do you know if they are UHJ encoded?
Robh3606 said:"Which Roger Waters CDs use Ambisonics?"
Amused to Death
Peter Gabriel
Up
The Waters disk is really good sounding with several wrap around effects. Kind of freaky when you hear stuff behind you and you know it's 2 channel. Great set-up disk. If your phasing is screwed up it all falls apart.
Rob
Looking at this I would say no but I don't know if it's up to date. Off hand I don't know. I would have to see what's in the liner notes with the disks.
http://members.cox.net/surround/uhjdisc/ambivari.htm
Rob
http://members.cox.net/surround/uhjdisc/ambivari.htm
Rob
Bjorn Kolbrek's Horn Articles in AudioXpress
Just received my latest copy of Ed Dell's AudioXpress magazine (to think I started my subscription to Audio Amateur in 1970 - I have the Summer 1970 issue in my hands now, complete with the address of my college dorm!)
Bjorn Kolbrek has just written Part Two of the articles, and you can find Part One here. Here's the link to the Horn Section of Bjorn's website, along with a very interesting discussion (and measurements) of a 200 Hz Midrange horn.
My prejudice in favor of impulse response over PA-style controlled-directivity is pretty well known, so I welcome articles with hard data (no smoothing, thank you very much). This data only confirms the decision to NOT use horns in the frequency regions where resistive radiation no longer applies - it's the extreme HF (above 7~10 kHz) where diaphragm breakup, phase-plug diffraction, and problems with horn directivity all work together to degrade the impulse response.
What I want to avoid most of all is the contemporary theater and PA practice of 10+ dB of boost equalization in the region where the drivers, phase-plugs, and horns are coming apart - adding more power in the same region where distortion is rapidly rising, and impulse response is storing a lot of energy, seems a sure prescription for harsh and fatiguing sound.
Just received my latest copy of Ed Dell's AudioXpress magazine (to think I started my subscription to Audio Amateur in 1970 - I have the Summer 1970 issue in my hands now, complete with the address of my college dorm!)
Bjorn Kolbrek has just written Part Two of the articles, and you can find Part One here. Here's the link to the Horn Section of Bjorn's website, along with a very interesting discussion (and measurements) of a 200 Hz Midrange horn.
My prejudice in favor of impulse response over PA-style controlled-directivity is pretty well known, so I welcome articles with hard data (no smoothing, thank you very much). This data only confirms the decision to NOT use horns in the frequency regions where resistive radiation no longer applies - it's the extreme HF (above 7~10 kHz) where diaphragm breakup, phase-plug diffraction, and problems with horn directivity all work together to degrade the impulse response.
What I want to avoid most of all is the contemporary theater and PA practice of 10+ dB of boost equalization in the region where the drivers, phase-plugs, and horns are coming apart - adding more power in the same region where distortion is rapidly rising, and impulse response is storing a lot of energy, seems a sure prescription for harsh and fatiguing sound.
The PDF is the same as the magazine article - I've read both. Very curious to hear your response on Part One and Two of the Kolbrek articles. (These PDF's load very slowly, by the way.) As far as I can see, Part Two concludes the Kolbrek series.
Dr. Geddes, thanks for your insightful comments on the many problems of conventional phase-plugs that you made over in the Geddes thread - much appreciated. Still puzzled why compression driver phase-plugs almost intentionally seem to be diffraction generators, whether they use traditional circumferential or Tangerine/radial slots.
There are many articles analyzing 2nd-harmonic air distortion based on calculations of throat diameter, yet the spaces inside a phase-plug are much, much smaller, not to mention right-angle turns where the sound from the "covered" part of the diaphragm enters the phase-plug slits. Surely these right-angle turns generate velocity noise.
Dr. Geddes, thanks for your insightful comments on the many problems of conventional phase-plugs that you made over in the Geddes thread - much appreciated. Still puzzled why compression driver phase-plugs almost intentionally seem to be diffraction generators, whether they use traditional circumferential or Tangerine/radial slots.
There are many articles analyzing 2nd-harmonic air distortion based on calculations of throat diameter, yet the spaces inside a phase-plug are much, much smaller, not to mention right-angle turns where the sound from the "covered" part of the diaphragm enters the phase-plug slits. Surely these right-angle turns generate velocity noise.
I just read part two and unfortunately he only added more errors to the batch.
There is, unfortunately, almost nothing written about horns that is correct. It is virtually all based on Websters Equation, which is completely wrong. When one realizes just how wrong Webster's equation is, and that it can't be used for any real device, you come away with the realization that only in the last few years have we come to really understand how horns work. Anything prior to that has to be viewed with extreme skepticism.
Lynne - thanks for that link to part two. I can write my rebuttal now.
There is, unfortunately, almost nothing written about horns that is correct. It is virtually all based on Websters Equation, which is completely wrong. When one realizes just how wrong Webster's equation is, and that it can't be used for any real device, you come away with the realization that only in the last few years have we come to really understand how horns work. Anything prior to that has to be viewed with extreme skepticism.
Lynne - thanks for that link to part two. I can write my rebuttal now.
Re: Bjorn Kolbrek's Horn Articles in AudioXpress
Lynne - I agree with your concerns, they are mine too, but if you look carefully at my data you will see that I have acheieved a very admirable impulse response WITH my waveguides. Your concerns are real, but your uneasyness about their solution is unfounded.
Once one understands the problem, and that prior work on honrs was mostly wrong, it becomes possible to design out the issue that you highlight. Bt you can't do it if you hold on to the older incorrect ideas as propagated in the works stated above.
Lynn Olson said:
Lynne - I agree with your concerns, they are mine too, but if you look carefully at my data you will see that I have acheieved a very admirable impulse response WITH my waveguides. Your concerns are real, but your uneasyness about their solution is unfounded.
Once one understands the problem, and that prior work on honrs was mostly wrong, it becomes possible to design out the issue that you highlight. Bt you can't do it if you hold on to the older incorrect ideas as propagated in the works stated above.
Lynn Olson said:
The reason for the "bad" phase plugs goes back to the Bob Smith work on them. He solved a non-existant problem and offered up as proof data that is completly ambiguous - proving nothing. But yet this "idea" persists.
As ALL of mine and others work on distortion perception has shown 2nd order nonlinearities are virtually inaudible. Hence there really isn't a concern over this factor.
In our AES paper we shown how nonlinear distortion in a compression driver was completely inaudible to some 30 subjects all the way up to the drivers thermal limit. In a 2" driver this is a whole lot of SPL. Lets face it nonlinear distortion in these devices is just not a factor. Diffraction, which exists in about 95% or the designs, is a very big factor.
And yes, I would never build a horn if I had to bend it. The problems out number the benefits.