I was trying to know if you tried them with larger horns and what you found to be good sounding and safe. For me for now it is ~700 on a 350Hz cutoff horn, buy wanting try a somewhat larger horn (around 250Hz )someday and check what the difference is.
Answering two questions, so here goes:
Both the Altec Valencia and the Model 19 (introduced about ten years apart) were modern successors to the grand-daddy of them all, the Lansing Iconic of the late Thirties. A medium format horn with a crossover around 800 Hz and a 15" direct radiator.
An all-horn system will invariably be larger, typically two or three times larger if performance down to 50 Hz is desired. This means a folded bass horn, which in turn gets into trouble around 300-400 Hz. So the mid horn has to be really big, or compromise by running the bass horn where it is getting rough and also running the mid horn right through cutoff.
It is possible to make 15" drivers as midwoofers, although the vast majority of 15" drivers made now are for theater or PA use, and are designed for power handling, not sound quality. Pro crossovers use 24 dB/octave slopes with brickwall cutoffs, so mid and upper-frequency breakups are of no concern in pro use.
But there are still a few 15" drivers with light cones, a smooth response in the 1-2 kHz region, and well-controlled rolloff above. Not many. Aside from the Altec/GPA 416, I can't think of any. Maybe Troy or one of the contributors in this forum might have some ideas.
Although a speaker this simple might be aiming too low, I just don't have the room for an all-horn rig. Besides, two of the best speakers I've ever heard were the big TAD monitors and the Blue Thunder in Switzerland. They were not demanding in terms of room size, and neither was auditioned in a treated room. They just played music, and played it superbly.
The commercial all-horn systems, like Avant-Garde, have been very disappointing. I would never own one, and the local DIY efforts, although physically imposing, are essentially tuned to just one living room, after months and years of twiddling and balancing. No thanks. The less to twiddle, the less to go wrong.
It might sound a little dumb, but I try to keep the number of variables small. The complexity of a speaker design goes up as the cube of the number of parallel pathways, so a 3-way takes far longer to tune than a 2-way. I have yet to hear one successful example of a 4-way, and truly excellent 3-ways are very thin on the ground. (I think the only 3-way I've ever liked was the Quad ESL57.)
Gary Dahl's speaker is a mature product. I took the path of least resistance and just twiddled with the bass enclosure a bit. If I have to spend a little time tuning the crossover, that's not so difficult if the horn is free of diffraction and nasty lumps in the response. That means avoiding horns with throat pinches and poor overall design (sorry, Altec).
Not surprisingly, my speaker is designed for my amplifier, which will be called the Blackbird and manufactured by Spatial Audio in early 2024. Class A push-pull 300B, with balanced triode- connected 6V6 drivers, and 6SN7 input. Custom interstage transformers with Monolith output and power transformers. Balanced throughout with separate, isolated B+ regulation for input+driver and output sections. 27 watts RMS with 220-watt regulators.
Frankly, I don't know of any good-sounding transistor amps. Maybe Rowland Research, but they are very very expensive. I did hear the Valhalla amps designed by Don Sachs and built by Spatial Audio at the Seattle show, and they were quite impressive, and only $5000 for a stereo amp, a superb value. Way better than nearly any other amp at the show. A classic Mullard circuit with 6L6 outputs and 6SN7 driver and input tubes. Don has excellent taste, and his amps are right up there. Also Quicksilver ... they make first-class tube gear right here in the USA, and super value too. Both are strong recommendations.
Both the Altec Valencia and the Model 19 (introduced about ten years apart) were modern successors to the grand-daddy of them all, the Lansing Iconic of the late Thirties. A medium format horn with a crossover around 800 Hz and a 15" direct radiator.
An all-horn system will invariably be larger, typically two or three times larger if performance down to 50 Hz is desired. This means a folded bass horn, which in turn gets into trouble around 300-400 Hz. So the mid horn has to be really big, or compromise by running the bass horn where it is getting rough and also running the mid horn right through cutoff.
It is possible to make 15" drivers as midwoofers, although the vast majority of 15" drivers made now are for theater or PA use, and are designed for power handling, not sound quality. Pro crossovers use 24 dB/octave slopes with brickwall cutoffs, so mid and upper-frequency breakups are of no concern in pro use.
But there are still a few 15" drivers with light cones, a smooth response in the 1-2 kHz region, and well-controlled rolloff above. Not many. Aside from the Altec/GPA 416, I can't think of any. Maybe Troy or one of the contributors in this forum might have some ideas.
Although a speaker this simple might be aiming too low, I just don't have the room for an all-horn rig. Besides, two of the best speakers I've ever heard were the big TAD monitors and the Blue Thunder in Switzerland. They were not demanding in terms of room size, and neither was auditioned in a treated room. They just played music, and played it superbly.
The commercial all-horn systems, like Avant-Garde, have been very disappointing. I would never own one, and the local DIY efforts, although physically imposing, are essentially tuned to just one living room, after months and years of twiddling and balancing. No thanks. The less to twiddle, the less to go wrong.
It might sound a little dumb, but I try to keep the number of variables small. The complexity of a speaker design goes up as the cube of the number of parallel pathways, so a 3-way takes far longer to tune than a 2-way. I have yet to hear one successful example of a 4-way, and truly excellent 3-ways are very thin on the ground. (I think the only 3-way I've ever liked was the Quad ESL57.)
Gary Dahl's speaker is a mature product. I took the path of least resistance and just twiddled with the bass enclosure a bit. If I have to spend a little time tuning the crossover, that's not so difficult if the horn is free of diffraction and nasty lumps in the response. That means avoiding horns with throat pinches and poor overall design (sorry, Altec).
Not surprisingly, my speaker is designed for my amplifier, which will be called the Blackbird and manufactured by Spatial Audio in early 2024. Class A push-pull 300B, with balanced triode- connected 6V6 drivers, and 6SN7 input. Custom interstage transformers with Monolith output and power transformers. Balanced throughout with separate, isolated B+ regulation for input+driver and output sections. 27 watts RMS with 220-watt regulators.
Frankly, I don't know of any good-sounding transistor amps. Maybe Rowland Research, but they are very very expensive. I did hear the Valhalla amps designed by Don Sachs and built by Spatial Audio at the Seattle show, and they were quite impressive, and only $5000 for a stereo amp, a superb value. Way better than nearly any other amp at the show. A classic Mullard circuit with 6L6 outputs and 6SN7 driver and input tubes. Don has excellent taste, and his amps are right up there. Also Quicksilver ... they make first-class tube gear right here in the USA, and super value too. Both are strong recommendations.
Last edited:
Going into more detail on the Altec Valencia and the Altec Model 19, the Valencia was a pre-Theile/Small vented speaker with a very basic 12 dB/octave crossover at a nominal 800 Hz. No horn EQ.
The Model 19 came out ten years later as the successor to the Valencia. The bass cabinet is a T/S vented bass alignment in a 9 cubic foot enclosure, and the crossover was moved up to 1200 Hz, with adjustable mid and HF equalization. The bass driver and HF horn were essentially the same as before, while the compression driver was improved with the Tangerine phase plug.
Another difference between the Valencia and the Model 19 was the Valencia was designed at the end of the tube era, and transistors had completely replaced tube amps when the Model 19 was designed.
The new speaker is intended for use with triode amplifiers with damping factors of about 4 (Zout = 2 ohms), a closed box is a better choice, giving an overall system Qts between 0.9 and 1.0, not that different than the classic acoustic suspension speakers of the Fifties ... a moderate bass lift. Crossovers are similar, and the horn is a much flatter modern design.
The Model 19 came out ten years later as the successor to the Valencia. The bass cabinet is a T/S vented bass alignment in a 9 cubic foot enclosure, and the crossover was moved up to 1200 Hz, with adjustable mid and HF equalization. The bass driver and HF horn were essentially the same as before, while the compression driver was improved with the Tangerine phase plug.
Another difference between the Valencia and the Model 19 was the Valencia was designed at the end of the tube era, and transistors had completely replaced tube amps when the Model 19 was designed.
The new speaker is intended for use with triode amplifiers with damping factors of about 4 (Zout = 2 ohms), a closed box is a better choice, giving an overall system Qts between 0.9 and 1.0, not that different than the classic acoustic suspension speakers of the Fifties ... a moderate bass lift. Crossovers are similar, and the horn is a much flatter modern design.
Last edited:
Thanks very much for these recommendations! Though the Blackbird mono blocks are probably too costly (unless available with more affordable options), these are wonderful alternatives. http://www.dsachsconsulting.com/valhalla integrated.html quicksilveraudio.com.
Don and I directly compared the Blackbirds to the Valhalla at the show. The tuning is quite similar, reflecting Don's excellent parts choices (bypass caps, etc.), but the resolution of the Blackbird is much higher, as you would expect from an extreme-category DHT amplifier. No surprise there.
But the 6L6 Valhalla (and the KT88 Kootenai) have nothing to apologize for. I thought they were the best amps at the show, surpassing the latest and greatest Class D amps, with the sole exception of the Blackbird. The Valhalla and Kootenai are very, very good, way better than most high-end tube amps, and are very sharply priced.
With a classical large-format studio monitor ... Altec 604 Duplex. Valencia, Model 19, and this speaker, a clean 25 watts will blast you out of the room. They're 97 dB efficient, after all, and that takes you to 111 dB undistorted, which is insanely loud.
It's the stupid 85 dB audiophile speakers that need 500-watt amplifiers ... and even then, they severely compress at those levels, thanks to voice-coil heating. A speaker with a 6" midrange and a 1" dome ain't gonna sound like an orchestra, I don't care what the reviews say.
This cartoon dates from 1956, when little 25-watt 6L6 amps and big 15" 97 dB speakers were the norm. Those things were LOUD.
But the 6L6 Valhalla (and the KT88 Kootenai) have nothing to apologize for. I thought they were the best amps at the show, surpassing the latest and greatest Class D amps, with the sole exception of the Blackbird. The Valhalla and Kootenai are very, very good, way better than most high-end tube amps, and are very sharply priced.
With a classical large-format studio monitor ... Altec 604 Duplex. Valencia, Model 19, and this speaker, a clean 25 watts will blast you out of the room. They're 97 dB efficient, after all, and that takes you to 111 dB undistorted, which is insanely loud.
It's the stupid 85 dB audiophile speakers that need 500-watt amplifiers ... and even then, they severely compress at those levels, thanks to voice-coil heating. A speaker with a 6" midrange and a 1" dome ain't gonna sound like an orchestra, I don't care what the reviews say.
This cartoon dates from 1956, when little 25-watt 6L6 amps and big 15" 97 dB speakers were the norm. Those things were LOUD.
Last edited:
August 1956 issue of Audio magazine
Go to page 28 to see the schematic of the Marantz power amplifier, which uses a Mullard circuit, just like the Valhalla and Kootenai. Unlike the Marantz, the Valhalla and Kootenai are fully regulated, and arguably use better tubes.
Multiply prices by ten to fifteen to get a better idea what things really cost back then. Tape decks, by modern standards, were shockingly expensive, at $500 each, and pre-recorded tapes were $15. It wasn't a cheap hobby unless you built your own.
Go to page 28 to see the schematic of the Marantz power amplifier, which uses a Mullard circuit, just like the Valhalla and Kootenai. Unlike the Marantz, the Valhalla and Kootenai are fully regulated, and arguably use better tubes.
Multiply prices by ten to fifteen to get a better idea what things really cost back then. Tape decks, by modern standards, were shockingly expensive, at $500 each, and pre-recorded tapes were $15. It wasn't a cheap hobby unless you built your own.
Last edited:
Hi Lynn Olson,
Can you be more specific, please?
Kindest regards,
M
I have read your replies, but still do not understand what is the designed feature that make it "designed for my amplifier"?
Can you be more specific, please?
Kindest regards,
M
No doubt their performance are stunning for the price. But perhaps for somewhat beyond the Valhalla's price, might there be an equally good 300B amp (s)?
Hi Lynn Olson,
Kindest regards,
M
Yes, I understand that. But, what feature makes it so suitable for a modest damping factor and unsuitable to a 200-watt amplifier?
Kindest regards,
M
The damping factor affects both the Q of the closed-box alignment and decisions made about the crossover. More complex higher-order crossovers are more sensitive to source impedance, while 2nd-order networks are relatively insensitive. They're a good starting point for a large format direct-radiator combined with large-format horn with a crossover in the 640 to 800 Hz region. If the horn is smooth enough, not much EQ will be required, which is a desirable goal. I have no interest in multi-amping or digital equalization.
The vast majority of commercial speakers on the market are optimized for high-powered transistor amps of 100 watts or more, and don't match zero-feedback triode amps all that well. This speaker is one of a small group that are optimized for moderate to medium power vacuum tube amps.
If somebody already owns a conventional or Class D transistor amp I have no speaker recommendations for them. If it helps any, my current Center speaker is a Dynaudio with the Esotar dome tweeter, which is a good tonal match with my Ariel speakers. The Center speaker is used with my Marantz 5.1 home theater setup.
As for detailed subjective comparisons between Troy's ES290 and the Athos Audio Yuichi A290, I honestly have no idea. Troy is a careful worker who does thorough measurements on all his designs, and also measures an interesting variety of compression drivers and direct radiators. He just recently measured a GPA 416 Alnico and it came out surprisingly well, with very low IM distortion and very smooth response, particularly for a 15" driver. His measurements of the inexpensive SB Audience Rosso CDN-65-T were also very impressive, which is why I will try them first. The new Textreme carbon fiber compression drivers also look promising, although at three times the cost.
The vast majority of commercial speakers on the market are optimized for high-powered transistor amps of 100 watts or more, and don't match zero-feedback triode amps all that well. This speaker is one of a small group that are optimized for moderate to medium power vacuum tube amps.
If somebody already owns a conventional or Class D transistor amp I have no speaker recommendations for them. If it helps any, my current Center speaker is a Dynaudio with the Esotar dome tweeter, which is a good tonal match with my Ariel speakers. The Center speaker is used with my Marantz 5.1 home theater setup.
As for detailed subjective comparisons between Troy's ES290 and the Athos Audio Yuichi A290, I honestly have no idea. Troy is a careful worker who does thorough measurements on all his designs, and also measures an interesting variety of compression drivers and direct radiators. He just recently measured a GPA 416 Alnico and it came out surprisingly well, with very low IM distortion and very smooth response, particularly for a 15" driver. His measurements of the inexpensive SB Audience Rosso CDN-65-T were also very impressive, which is why I will try them first. The new Textreme carbon fiber compression drivers also look promising, although at three times the cost.
Last edited:
Hi Lynn Olson,
is it correct?
Kindest regards,
M
If I understand you correctly, you have accounted for the resistance part of the impedance of the tube amplifier (which I understand is in tens of Ohms), when designing the enclosure. The enclosure has a Q that is dependent on the T/S parameters of the driver and the enclosure parameters (leakage, volume, stuffing), which will be modified by the resistance in series with the driver.
is it correct?
Kindest regards,
M
Incorrect. On the 8-ohm tap, the Blackbird/Karna amplifier, along with nearly all SET amplifiers, have a Zout of 2 ohms, which equals a damping factor of 4. If you go back to traditional push-pull pentode amps with 12 to 20 dB of feedback, the Zout is typically 0.5 ohms, or a damping factor of 16. To get a Zout that is significantly lower, you have to use transistor amps with much higher feedback ratios.
To this Zout, you add perhaps 0.5 ohm from cabling and the inductor in the crossover, so the woofer sees a source impedance of 2.5 ohms. This source impedance is in series with the voice coil resistance, which is typically 6.5 ohms in an "8-ohm" driver.
The Qes and Qts figures of the driver assume a zero source impedance, but the voice coil resistance is not removable and is always there. What the added resistance does, since it is series with VC resistance, is shift it upwards from 6.5 ohms to 9 ohms. This shifts Qes and Qts upwards (see math on the Internet for the equations).
In both vented and closed box enclosures, the correct procedure is to increase box volume, and if vented, retune the box. However, in vented systems, there is a practical upper limit for driver Qts when the required box size goes to infinity (or the size of a room). This upper limit for driver Qts is quite a bit higher for closed boxes, going all the way to 1.0 for a true "infinite baffle" that opens to another room. The practical upper limit for vented boxes is closer to 0.5, so it's not a small difference.
The same warnings apply to passive crossover design. As the order of the crossover increases, the sensitivity to detuning goes up rapidly. The least sensitive to detuning is 2nd-order, which is also the traditional big-monitor crossover going back to the Lansing Iconic of the Thirties. Not by accident, the Iconic had a 800 Hz crossover, which is set by woofer size and horn cutoff frequency. The driver physics have not changed since the Thirties, so that's still the best choice now.
The system has a lot of Lansing Iconic in it. Pretty much the same woofer, similar box size, same low or zero feedback amplifiers, same crossover, and a modern horn and compression driver.
Not surprisingly, it will sound similar to a Lansing Iconic, Altec Valencia, or Model 19, just with a much smoother modern horn and compression driver ... and a lower diffraction bass enclosure, because that results in smoother response and better phase transfer around the 800 Hz crossover frequencies. Things we know now that were not known back then.
P.S. It is a simple exercise to calculate the Zout of a zero-feedback triode amplifier. Look up the plate impedance of the power tube. This is about 700 ohms for a 300B. The output transformer will have a stated primary impedance and an 8-ohm tap on the secondary. 3300 ohms is a very common primary impedance for a SET transformer, which gives a ratio of 412.5 if using the 8-ohm tap. Divide 700 ohms by 412.5, which gives 1.7 ohms output impedance. Transformers are designed to have no more than 5% in copper losses, so that adds a negligible amount to the Zout.
Pentodes have much higher plate impedances, but are always used in feedback amplifiers. The ultralinear connection is intermediate between pentode and triode. Those calculations are more complex.
Last edited:
Since there seems to be some confusion, let me discuss voltage-out, current-out, and Unity-coupled amplifiers.
Let's look at the simplest loudspeaker, a direct-radiator single cone driver in a closed box (with or without filling). This will have a sharp Zmax peak somewhere between 30 and 100 Hz, a very broad Zmin between 200 and 500 Hz, and a smooth rise in impedance above that. The Zmin is typically only 10% above the measured DC resistance of the voice coil.
If this driver is perfect (without midrange cone breakup), it will have flat response throughout the working range if connected to a voltage-out amplifier. (Ignoring diffraction loss, HF rolloff, etc. etc.)
If connected to a current-out amplifier, with a Zout many times higher than 8 ohms (say 1000 ohms), then the frequency response will look exactly like the impedance curve. It will require massive equalization to have response the same flatness as the voltage-out example.
So-called "Unity-coupled" amplifiers have the same Zout as the driver itself, say 8 ohms. This is nearly as useless as the current-out amplifier, since the response curve will look very nearly the same as the current-out example, just not as bad. In practice, I am aware of no commercial "Unity-coupled" amplifiers, except for the novelty "Z-matic" amplifiers of the Fifties with user-adjustable feedback.
In practice, both current-mode and Unity coupling are useless unless very deep response shaping is used. There are some exotic multi-amp systems where this approach used, along with low-level crossovers to shape the response back to flat.
Zero-feedback triode amplifiers are neither current-out nor Unity coupled. They are mostly voltage-out, so are compatible with most, but not all, loudspeakers.
Let's look at the simplest loudspeaker, a direct-radiator single cone driver in a closed box (with or without filling). This will have a sharp Zmax peak somewhere between 30 and 100 Hz, a very broad Zmin between 200 and 500 Hz, and a smooth rise in impedance above that. The Zmin is typically only 10% above the measured DC resistance of the voice coil.
If this driver is perfect (without midrange cone breakup), it will have flat response throughout the working range if connected to a voltage-out amplifier. (Ignoring diffraction loss, HF rolloff, etc. etc.)
If connected to a current-out amplifier, with a Zout many times higher than 8 ohms (say 1000 ohms), then the frequency response will look exactly like the impedance curve. It will require massive equalization to have response the same flatness as the voltage-out example.
So-called "Unity-coupled" amplifiers have the same Zout as the driver itself, say 8 ohms. This is nearly as useless as the current-out amplifier, since the response curve will look very nearly the same as the current-out example, just not as bad. In practice, I am aware of no commercial "Unity-coupled" amplifiers, except for the novelty "Z-matic" amplifiers of the Fifties with user-adjustable feedback.
In practice, both current-mode and Unity coupling are useless unless very deep response shaping is used. There are some exotic multi-amp systems where this approach used, along with low-level crossovers to shape the response back to flat.
Zero-feedback triode amplifiers are neither current-out nor Unity coupled. They are mostly voltage-out, so are compatible with most, but not all, loudspeakers.
Hi Lynn Olson,
thank you for the explanation in your post.
If I understand correctly your other posts, you plan to use a passive cross-over. Should its resistance not be added to that of the output resistance of the amplifier and the conductors?
I think that apart from that, I have all the information to make a model to simulate your enclosure to compare it with a similar design that I made, albeit with a different driver - TAD1601A.
Kindest regards,
M
thank you for the explanation in your post.
If I understand correctly your other posts, you plan to use a passive cross-over. Should its resistance not be added to that of the output resistance of the amplifier and the conductors?
I think that apart from that, I have all the information to make a model to simulate your enclosure to compare it with a similar design that I made, albeit with a different driver - TAD1601A.
Kindest regards,
M
