Ariels were a time a Loooong time ago.
A spent force by Any Yardstick.
Few have heard them Others could simply care less 😉
Wanna Forum? Start one.
A spent force by Any Yardstick.
Few have heard them Others could simply care less 😉
Wanna Forum? Start one.
And yet this thread has a huge amount of real information developed by some very talented people and backed up by experimentation and measurement.
Certainly more valuable than your yardstick.
I think this thread is great in a way that it talks about lots of personal experience, some we may agree with, some not, some actually inspire us to think. I think tube amplifiers have their strengths, I have personally heard some qualities in tube devices that reveal the spacial and dynamic qualities of a recording in a way that I have not heard in solid state devices, but I also see some potential in getting close to these qualities. One thing that I probably do different is that I like to ask people to nit pick during listening rather than whether they like the sound or not. Generally you get lots of variety of feedback. Some people will give you too generic opinions like the highs sound too bright or dull, but the really useful feedback is like "the timbre of the cymbal is a bit on the harsh side", then when you get to the vocals, people whom are familiar with different performers can tell you what part of their vocal quality and skills are not adequately reproduced in a certain way. All these are valuable feedback in the process of finding where the exact problem is, and can usually also be identified through different measurement methods.
Sent from my iPad using Tapatalk
Sent from my iPad using Tapatalk
Bare if you don't like the thread just don't come here. Pretty simple! Continuing to crap in the thread will earn you bin time or worse.
Gary’s Horns + AMTs?
Apologies if I happen to misrepresent someone’s work or words in the course of proposing yet another of my world famous brainstorms: To extend the HF range beyond 10kHz as well to attenuate a possibly existing resonant bump or notch (?) of the Azurahorn AH425/Radian 745NeoBe driver, Gary and Pierre added a passive HF EQ filter (?).
In any case, Lynn mentioned (# 12468) that AMTs are the next best thing to horns for the HF range. So might an AMT somehow be a better choice to deal with one or both problems?
Could that have been why the RAAL Lazy ribbon that Gary originally tried proved unsatisfactory?
If yes, than please recommend an AMT from ESS Heil AMT | ESS Labs., LLC or Beyma PLEATED DIAPHRAGM TWEETERS | Beyma .
Also, in my newbie attempt to address all variables, might an AMT in this case not work as well as it might otherwise due to placement issues? That is, could one reason why the RAAL ribbons didn't work was because Gary may have had little choice but to sit it atop the Altecs and alongside the horns?
And on a possibly related point, if a Beyma model didn’t include a waveguide does that mean that the AMT wouldn’t need one for this particular application?
Otherwise, if a particular model AMT would be perfect to cross with Gary's horns (and simply placed right alongside the AH425), please specify the best kind of (passive) filter, slope and crossover frequency.
Apologies if I happen to misrepresent someone’s work or words in the course of proposing yet another of my world famous brainstorms: To extend the HF range beyond 10kHz as well to attenuate a possibly existing resonant bump or notch (?) of the Azurahorn AH425/Radian 745NeoBe driver, Gary and Pierre added a passive HF EQ filter (?).
In any case, Lynn mentioned (# 12468) that AMTs are the next best thing to horns for the HF range. So might an AMT somehow be a better choice to deal with one or both problems?
Could that have been why the RAAL Lazy ribbon that Gary originally tried proved unsatisfactory?
If yes, than please recommend an AMT from ESS Heil AMT | ESS Labs., LLC or Beyma PLEATED DIAPHRAGM TWEETERS | Beyma .
Also, in my newbie attempt to address all variables, might an AMT in this case not work as well as it might otherwise due to placement issues? That is, could one reason why the RAAL ribbons didn't work was because Gary may have had little choice but to sit it atop the Altecs and alongside the horns?
And on a possibly related point, if a Beyma model didn’t include a waveguide does that mean that the AMT wouldn’t need one for this particular application?
Otherwise, if a particular model AMT would be perfect to cross with Gary's horns (and simply placed right alongside the AH425), please specify the best kind of (passive) filter, slope and crossover frequency.
Oltos,
It just seems you are trying to design a system based on words and opinions you are finding about many different devices. You could be going down the rabbit hole so to speak as until you hear a complete system you won't know what you are really attempting to do. From what I know of the Radian drivers there are a few small areas that need some attending to, most likely this is with all the Radian drivers as all of them are built with some real similarities. Whether any of them have anything to do with the type of suspension used with all of their drivers I wouldn't say. The bottom line is none of the problems are so severe that they can't be corrected by a well designed network, whether that is passive or active analog or even a dsp type of network. That is something you can't predict without accurate FR, Impedance curves and waterfall plots. But with the number of these driver horns already in use out in the wild I wouldn't think that information isn't already in one of the threads on this site. What you are missing when you start looking at adding another device over the top of the round horns are that you are going to have some serious issues with trying to make this work. You will more than likely have major time alignment problems, you will have a large distance between device centers up in the frequencies where this can become real obvious instantly. Also I would imagine that the radiation pattern at the crossover point will be a big discontinuity. If you really want to do something that can work then my recommendation would be to use a second small horn above the other with one of the small !/2" or 3/4" exit plastic diaphragm drivers or if you have tons of money look for a pair of the TAD ET-703 drivers whether new or used. They are tiny. Perhaps you need to get on a plane and go hear these speakers that Gary and others have already built and see what you think when you personally hear them yourself. But as Lynn has so well pointed out it is not the speakers alone you are listening to but the speaker and electronics package that is driving them.
My 2 cents of advise before you dream up another unknown combination that may cause more issues than you seem to be chasing.
It just seems you are trying to design a system based on words and opinions you are finding about many different devices. You could be going down the rabbit hole so to speak as until you hear a complete system you won't know what you are really attempting to do. From what I know of the Radian drivers there are a few small areas that need some attending to, most likely this is with all the Radian drivers as all of them are built with some real similarities. Whether any of them have anything to do with the type of suspension used with all of their drivers I wouldn't say. The bottom line is none of the problems are so severe that they can't be corrected by a well designed network, whether that is passive or active analog or even a dsp type of network. That is something you can't predict without accurate FR, Impedance curves and waterfall plots. But with the number of these driver horns already in use out in the wild I wouldn't think that information isn't already in one of the threads on this site. What you are missing when you start looking at adding another device over the top of the round horns are that you are going to have some serious issues with trying to make this work. You will more than likely have major time alignment problems, you will have a large distance between device centers up in the frequencies where this can become real obvious instantly. Also I would imagine that the radiation pattern at the crossover point will be a big discontinuity. If you really want to do something that can work then my recommendation would be to use a second small horn above the other with one of the small !/2" or 3/4" exit plastic diaphragm drivers or if you have tons of money look for a pair of the TAD ET-703 drivers whether new or used. They are tiny. Perhaps you need to get on a plane and go hear these speakers that Gary and others have already built and see what you think when you personally hear them yourself. But as Lynn has so well pointed out it is not the speakers alone you are listening to but the speaker and electronics package that is driving them.
My 2 cents of advise before you dream up another unknown combination that may cause more issues than you seem to be chasing.
I considered the Beyma TPL-150H for my speaker project, but became suspicious that it wouldn't make as seamless a transition to a 15" woofer as does the AH425.
The RAAL ribbon did everything it was supposed to do...HF extension was superb. But overall coherency was better when I equalized the Radian and went without the ribbon. It's a simple tradeoff. For me the choice was easy because of the high priority I place on lifelike tone colors of orchestral instruments--the results went solidly in this direction.
It is likely that the sound could be improved further with additional optimization of the crossover and EQ, but I don't want to add another forward-facing HF driver. Lynn and I have discussed the possible use of adding an extra driver facing the rear (or some variation) but I haven't taken any steps in that direction yet.
Gary Dahl
The RAAL ribbon did everything it was supposed to do...HF extension was superb. But overall coherency was better when I equalized the Radian and went without the ribbon. It's a simple tradeoff. For me the choice was easy because of the high priority I place on lifelike tone colors of orchestral instruments--the results went solidly in this direction.
It is likely that the sound could be improved further with additional optimization of the crossover and EQ, but I don't want to add another forward-facing HF driver. Lynn and I have discussed the possible use of adding an extra driver facing the rear (or some variation) but I haven't taken any steps in that direction yet.
Gary Dahl
Have you tried using active crossovers? Solutions like miniDSP and Ultimate Equalizer software are pretty flexible.
Not my cup of tea, except perhaps for bass frequencies. I bought a QSC DSP-30 last year with the intent of using it with my Parasound HCA-1500A to drive the low bass cabinets, but haven't gotten that far (still using plate amps). With the bass performance I'm getting now, I'm not especially motivated to experiment with it.
If I were using only digital program material, it could be an interesting possibility. Last time I visited Gary Pimm, his crossovers and room correction were all done in DSP before conversion to analog. The results were quite impressive. But the idea of an extra A/D - D/A step doesn't appeal to me.
Gary Dahl
If I were using only digital program material, it could be an interesting possibility. Last time I visited Gary Pimm, his crossovers and room correction were all done in DSP before conversion to analog. The results were quite impressive. But the idea of an extra A/D - D/A step doesn't appeal to me.
Gary Dahl
Hi,
Great reading and thank you.
Might you try reducing plate current of 5687 to 8-10mA around 100V plate, and again even to 4-5mA and see if that doesn't provide more satisfactory result before changing out for another tube.
Best to you.
LH
I did not realize that. It makes things more complicated if you are only playing the traditional records. Lot's more variables. The most perfect tone arm I had seen was Vertere Reference, but anything along the signal path can be the source of problem that prevents you from getting better fidelity with less sacrifice. Identifying the most critical point is an issue. When you get the speakers better technically and it is not giving you the right sound, then it would be hard to judge if it is something overlooked in the speaker or whether the true problem is upstream unless you have a systematic way of isolating these issues.
Gary Pimm tested a zillion tubes with various resistive and current-source loads (using his current source of course). I was there for some of the tests and was surprised what we found.
When you look at the entire spectra out to the 10th harmonic, each tube types has a specific "fingerprint". It's quite distinctive, and is pretty consistent within a vendor over the years.
Surprisingly, varying the current 2:1, or even 3:1 (while keeping the plate voltage constant) only moves the whole spectra up and down about 3 dB, sometimes a little more. But the shape stubbornly stays the same.
Since the shape of the spectra is mostly the result of the tube type, followed by the manufacturer, our best guess the spectra is created by the interaction of the space-charge electron cloud, the proximity of the grid to the cloud, and the construction of the grid (uneven winding pitch, crooked structure, electron leakage around the grid, etc.)
There's no way the user can rebuild the tube. A brand-new tube will wander around for the first 10 hours as traces of gas are removed by the "getter", but once it settles in, that's what it's going to do for the next several thousand hours. (If it's defective, it never settles in, and is unstable.)
So the measured spectra pretty much stays as-is. An ideal current-source load almost always gives the fastest rolloff of upper harmonics, but a slanted load-line from a resistive load then raises the upper harmonics to some degree.
The extent to which the upper harmonics start to increase very much depends on tube type and manufacturer. Slanting the load-line (asking the tube to deliver audio-frequency current into a load) is a way of testing the current linearity of the tube ... some are happy delivering power, and some aren't. The 6SN7 family (including the single-triode predecessors) is exceptionally good in this respect compared to the 9-pin miniatures that replaced them in the early Fifties.
By then, feedback was in universal use for audio circuits, so linearity became less important. A lot of 9-pin miniatures were designed for use in televisions, where heat emission, compact size, and low cost were design priorities. Very few TVs at the time had linear grayscale tracking, so high distortion was of little concern.
The 5687/7044/7119 MILSPEC and commercial-grade tubes were intended for long-distance telephone repeaters, analog computers, and aviation applications. Analog computers needed linearity, but it's not as important as it is for audio applications. The prime goal was reliability; these were not sold into consumer applications.
The 12AX7, 12AU7, and 12AT7 were consumer-grade tubes used in B&W and color televisions and audio products of all kinds. They replaced the hotter and bulkier octal types that preceded them, but weren't any better in performance ... if anything, they were a step down from the predecessors.
It has to be remembered that a bargain-basement B&W television with truly abominable performance (visibly distorted picture, poor focus, little or no interlace, no black-level compensation, and poor sync stability) had about 16 to 18 tubes in a very tight and hot chassis, and a good-quality color TV based on the RCA patents had about 30 tubes! Vacuum tubes are not the easiest way to video signal processing ... sync extraction, sweep circuits, color-burst detection and decoding, chroma demodulation, luminance and chroma signal matrixing, and final amplification into the CRT grids.
Most tubes went into TV sets and the replacement market. Existing sets needed replacement tubes about every three years or so, mostly due to extreme abuse of recommended operating conditions. The main criteria for performance was simply whether they worked or not, which is about all a TV-type tester will tell you.
The total hifi market (including consoles from RCA, Zenith, and Magnavox) was much smaller than TV sets, I'd guess less than 1/10th the size, so the set manufacturers mostly drove the market. On the other hand, tubes like the KT66 and KT88 were very much designed for high-quality audio amplifiers.
(Direct-heated triodes like 45's and 2A3's were obsolete at the time, and equally obsolete 300B's were never sold to consumers or manufacturers since Western Electric leased their equipment to theaters. Telephones were not for sale; you rented them from Bell Telephone, and the cord went into a non-removable wall fixture. Each additional telephone resulted in a higher monthly charge from Ma Bell; take it or leave it.)
Last edited:
On integrating a supertweeter
All very true and measurable, of course.
But: at such high frequencies (> 7kHz), won't all these problems (sub-ERB comb filtering, lack of time alignment) be essentially inaudible?
My own experiments have led me to favour the use of a 'constant power' 2nd order Butterworth crossover between the horn (850Hz-7.2kHz bandwith) and the supertweeter. Not a 'perfect' integration technically, of course, but the least of all evils, as it were.
I have discussed this in this other thread:
http://www.diyaudio.com/forums/multi-way/186540-my-island-speakers-4.html#post4353104
I would appreciate your thoughts on this!
Marco
All very true and measurable, of course.
But: at such high frequencies (> 7kHz), won't all these problems (sub-ERB comb filtering, lack of time alignment) be essentially inaudible?
My own experiments have led me to favour the use of a 'constant power' 2nd order Butterworth crossover between the horn (850Hz-7.2kHz bandwith) and the supertweeter. Not a 'perfect' integration technically, of course, but the least of all evils, as it were.
I have discussed this in this other thread:
http://www.diyaudio.com/forums/multi-way/186540-my-island-speakers-4.html#post4353104
I would appreciate your thoughts on this!
Marco
My experience too.
I originally wanted a 3 way and experimented at length with two tweeters, the RAAL 140-15d and the Electro-Voice T350. Tried them crossed at 7 kHz and at 9 kHz in different locations around the horn. To my surprise, the equalized Radian Be sounded more natural, more coherent, and more dynamic. The dispersion is more limited, but I can live with this. Among other qualities, it reproduces trumpet without "kazooness".
Oltos - My experience has been that forum members have stepped in and helped me when I needed help. This is a wonderful community. Once you decide what you want to build, you will not be alone. This holds for equalizing the Radian. There is no recipe but it is simple to do. The circuit flattens the response. Then "small" adjustments become a matter of personal taste, and tuning the speaker to your upstream electronics and your room.
Last edited:
Jumping on the band wagon. It is very good to read and see how the project is opening up again. It seems to be firming up to my mind as a two way mid/treble horn with an IB or reflex 15" and not the open baffle. Allowing for Radian B&C Beyma options with the round LeCleac'h horn (revered) and GPA416B or TD15M for US or one of the European equivalents in the mix, it is then just down to the horn equalisation and cross over.
People complained about th slowness of the project, but if you have a good but dated system, you can spend like Lynn the considerable time evolving with the eager ones, to a stable platform. If only some body could do a Gary Dahl here so we can share the experience.
For Europe, Troels approach has allowed followers to gradually migrate to a favoured choice and be very pleased with it. So is anyone in the UK migrating to do a Gary Dahl Arial replacement look alike. I hope so.
People complained about th slowness of the project, but if you have a good but dated system, you can spend like Lynn the considerable time evolving with the eager ones, to a stable platform. If only some body could do a Gary Dahl here so we can share the experience.
For Europe, Troels approach has allowed followers to gradually migrate to a favoured choice and be very pleased with it. So is anyone in the UK migrating to do a Gary Dahl Arial replacement look alike. I hope so.
Marco,
I have had the advantage in the past that I was designing and making my own horns. While I did use multiple horns I made each with identical patterns, they were just larger in scale as I went down in frequency. The only horn that wasn't a match was the bass horns as they just get to big to do that. In my later designs as you see in my avatar I used the shortest horns I could with 1/4 wavelength mouth sizes that had very wide dispersion patterns. It was much easier to integrate each section and because I wasn't using a round mouth the distances from centers was less that way. The mouths never had any parallel surfaces, I think I was one of the first to make this happen and I used a round over at every junction so diffraction was minimized. If it wasn't for the strange enclosure shape and lack of marketing money I think you would have seen a lot of this done in the industry.
I just have a problem with trying to integrate round horns, it becomes impossible to have a limited spacing between centers and you have these crazy surfaces that each stacked horn has to reflect off of, you have to consider that the outside of these exposed horns are now the reflective surfaces that the next horn up in size is reflecting off of, not some smooth surface or controlled shape. It becomes a scattered mess of reflections.
Even the TAD ET-703 which has such incredible response bothered me in the sense that it was truly a diffraction lens, I considered at the time just removing that lens and making another horn that matched all the others I was using at the time but never took it that far. I was only using that little device from 10Khz up and I was using fast 4th order LR slopes where the integration wasn't as critical. It is always a choice of where any discontinuities were least audible.
I never use 2nd order filters myself, I just can never get past the impulse response you get when combining those devices, do you or don't you invert the phase, I could just never come to a satisfactory answer when trying to make that work.
The picture was of something of a more traditional shape, not as nice as my molded shapes but really something different. There were two bass driver, the one you see and another down firing. Each was in a separate enclosure build into the cabinet, totally isolated from each other with the larger device down firing. They both were ported enclosures with the lower unit tuned into the 25hz range. The top horn was the TAD ET 703 and the next horn down used a TAD 2001 driver. Below that was 6 1/2" cone driver on the horn, I just like the sound of a cone in the lower range over any compression driver. There was no time alignment so that would be the only thing I would change today, a dsp solution wasn't available at the time, passive time alignment was out and active time alignment was only used in PA at the time and was rather crude. The bottom driver shown was a custom made Radian 8" driver which was never made commercially for anyone else that I know of. It was a paper cone instead of the plastic cone they were selling at the time.
I'll be out the rest of the day. As you can see there are as many ways to do things as we can think up.
I have had the advantage in the past that I was designing and making my own horns. While I did use multiple horns I made each with identical patterns, they were just larger in scale as I went down in frequency. The only horn that wasn't a match was the bass horns as they just get to big to do that. In my later designs as you see in my avatar I used the shortest horns I could with 1/4 wavelength mouth sizes that had very wide dispersion patterns. It was much easier to integrate each section and because I wasn't using a round mouth the distances from centers was less that way. The mouths never had any parallel surfaces, I think I was one of the first to make this happen and I used a round over at every junction so diffraction was minimized. If it wasn't for the strange enclosure shape and lack of marketing money I think you would have seen a lot of this done in the industry.
I just have a problem with trying to integrate round horns, it becomes impossible to have a limited spacing between centers and you have these crazy surfaces that each stacked horn has to reflect off of, you have to consider that the outside of these exposed horns are now the reflective surfaces that the next horn up in size is reflecting off of, not some smooth surface or controlled shape. It becomes a scattered mess of reflections.
Even the TAD ET-703 which has such incredible response bothered me in the sense that it was truly a diffraction lens, I considered at the time just removing that lens and making another horn that matched all the others I was using at the time but never took it that far. I was only using that little device from 10Khz up and I was using fast 4th order LR slopes where the integration wasn't as critical. It is always a choice of where any discontinuities were least audible.
I never use 2nd order filters myself, I just can never get past the impulse response you get when combining those devices, do you or don't you invert the phase, I could just never come to a satisfactory answer when trying to make that work.
The picture was of something of a more traditional shape, not as nice as my molded shapes but really something different. There were two bass driver, the one you see and another down firing. Each was in a separate enclosure build into the cabinet, totally isolated from each other with the larger device down firing. They both were ported enclosures with the lower unit tuned into the 25hz range. The top horn was the TAD ET 703 and the next horn down used a TAD 2001 driver. Below that was 6 1/2" cone driver on the horn, I just like the sound of a cone in the lower range over any compression driver. There was no time alignment so that would be the only thing I would change today, a dsp solution wasn't available at the time, passive time alignment was out and active time alignment was only used in PA at the time and was rather crude. The bottom driver shown was a custom made Radian 8" driver which was never made commercially for anyone else that I know of. It was a paper cone instead of the plastic cone they were selling at the time.
I'll be out the rest of the day. As you can see there are as many ways to do things as we can think up.
Attachments
On the topic of the Karna amp. I too have collected parts to build this amp, but apart from the power supply blocks haven't finished the amps (one day soon I may). I'm using a pair of 211 SETs for now thanks to Thomas Mayer.
Anyway, I think the one of the challenges is coming up with a quad of nice 45s or 46s. I collected some of both over the years but still haven't gotten great matches. Then I started reading about people trying out the (newly popularized) russian 4P1L. This is a DHP that has great performance as in triode mode. They are still plentiful and cheap. I bought a box of 30 or so a while back. I was hoping to explore the possibility to use the 4P1L in lieu of the 45/46. It has a bit higher gain. I haven't worked out all the characteristic for the operating point to know if it would be a viable solution, but at first blush looks to be a good choice.
Kevin of K&K had a version of the PP 300B amp that used 6SN7 into 6BX7 driving the IT with the first two stages DC coupled with a choke on the SN7 plates and a CCS feeding the whole 2 stage front end. He doesn't show it on his site at the moment but mentioned he still has the schematic (I do as well somewhere). I was thinking of doing 6SN7 into 45(4P1L) ala Kevin's topology, but now you guys have me thinking.
Anyway, I think the one of the challenges is coming up with a quad of nice 45s or 46s. I collected some of both over the years but still haven't gotten great matches. Then I started reading about people trying out the (newly popularized) russian 4P1L. This is a DHP that has great performance as in triode mode. They are still plentiful and cheap. I bought a box of 30 or so a while back. I was hoping to explore the possibility to use the 4P1L in lieu of the 45/46. It has a bit higher gain. I haven't worked out all the characteristic for the operating point to know if it would be a viable solution, but at first blush looks to be a good choice.
Kevin of K&K had a version of the PP 300B amp that used 6SN7 into 6BX7 driving the IT with the first two stages DC coupled with a choke on the SN7 plates and a CCS feeding the whole 2 stage front end. He doesn't show it on his site at the moment but mentioned he still has the schematic (I do as well somewhere). I was thinking of doing 6SN7 into 45(4P1L) ala Kevin's topology, but now you guys have me thinking.
I agree. I use a radial horn myself.
OK, so that ET-703 bolted to the front baffle was clearly out of alignment with the TD-2001 below by several wavelengths at the 10kHz crossover frequency. Also, while a 4th order high-pass would minimize the comb filtering, the latter would still be clearly visible when measured.
Now, did you find those issues bothersome or not, when actually listening to the system? I'd guess not really. That was my point. And the fact that if one uses a CPC, rather than an APC, the overall 1/3 oct smoothed summation remains unaffected within a reasonably wide listening window.
Thank you for the reply anyway! I bet that 'speaker of yours there sounded really good!
Marco
> 7 kHz we do have very limited hearing resolution and as such problems in this area will tend to be less audible. But then doesn't that just beg the question of why bother with another driver up that high. Most drivers can do a reasonable job at these frequencies - so why mess with another driver and another crossover and all the unavoidable problems?