If you take the GPA 515 a little bit lower, you could go for a tapped horn for the lowest two octaves. Check out the THAM15 design, which can be modified to go down to 30 Hz. I have a feeling the character of this TH will match the rest of your system very well, not to mention your taste; The bass is dynamic, crisp, tuneful, fast and punchy. Sensitivity is very high and the enclosure is relatively compact. Use with Beyma 15P80Nd or similar. Require steep LP and HP though (i.e. realistically active).
www.martinsson.cc/blog
That's what I thought and was hoping to here. I pointed out that when implementing the 1st order at 6000hz, the signal will be down 18db at fs which is lower than if a second was used at 1800hz. However,,,I'm sure there is more to look out for in application if one was looking for perfection.
Unless your waveguide is CD, you cannot get away with a first order HP. That's because in a constant directivity waveguide the axial response falls at 6 dB / octave and a first order filter at something like 10 kHz flattens it out. If you use a "standard" horn then the axial response will be flat and first order LP set very high will create a very bright system.
Designing crossovers for horns is not simple. You just can't go using "textbook" filters and expect things to work out right.
Designing crossovers for horns is not simple. You just can't go using "textbook" filters and expect things to work out right.
In a line level application, proper iron isolates the stages and can help prevent ground loops, as well as a reduction in common mode noise. In a speaker application, for meaningful attenuations, it's more energy efficient and maintains a higher damping factor. Whether those are important or not speaker level, they are legitimate advantages. Energy efficiency for the SET/horn crowd is particularly nice.
Not for you, I get that, but that doesn't make it invalid.
I do similar though I cascade in a second filter pole below resonance and usually do some impedance shaping, to keep the cap filter acting right.
I was surprised by the measurements too! But I listened to it first. After hearing the sound, almost didn't care what the measurements were going to show.
I agree, it would be very interesting to compare the AL and BE diaphragms in the NEO. But I don't have the AL diaphragm to fit the NEO body. If anyone has a 745 NEO/AL to lend, I would be happy to test it in the same setup and make an overlay graph.
I was not talking about electronic situation. That's entirely different.
Agreed on the energy savings - it's the "Green" solution! But that's about the only advantage that I can see. But as to that being of benefit to the SET crowd, it doesn't improve the sensitivity, just the efficiency. It would have to save a lot of power to offset its cost.
Yes they certainly are pieces of work here.
No pride of workmanship at all.
The Radian 465 has a similar graph.
Extra headroom without adding more output devices etc is very appealing. Most amps lower distortion into higher impedances, too, though one would want to be careful. I use class D and so need to keep my high frequencies at a moderate load, to prevent the LC filter from killing my HF extension. This could also be an issue in some designs with OPTs, as the parasitics could make difficulties. Another case might be OTLs, where some double their output power into double the load, and can achieve much higher power into a higher impedance speaker, particularly if designed around the specific load. These cases are not ones you are particularly sensitized to as you're a devoted SS user, but for people who are fond of esoteric amps, there are absolutely cases where the higher efficiency is less expensive than a bigger amp. Most of the power may be consumed below the horn's passband, but extra watts can be quite dear in tube designs.
The damping may (or may not, depending on system) be a significant benefit over resistor solutions. There are cases where a resistive attenuator might have advantages, with the parallel (to the device) resistor offering significant impedance compensation in cases of high attenuation (low parallel R value). As with everything, the details matter.
Direct-heated triode (DHT) watts are very, very expensive. A single-ended 2A3 (at around $50 or more each) amplifier puts out 3 watts at 3% distortion (fortunately almost entirely benign-sounding 2nd harmonic). A single-ended 300B (between $75 and $400 each) amplifier puts out 8 watts at 3% distortion. (The prices are for a single output tube, not the entire amplifier.)
My 20-watt amps are quite a lot more expensive, with PP 45 drivers, and PP 300B output tubes. For a stereo pair, that's four 45's, and four 300B's (in matched pairs). Even the Chinese tubes are not cheap, and the premium Czech versions are double to triple the price. (Russian tubes are about midway in price.)
The alternatives for more power are AM-radio transmitter tubes that run at B+ voltages between 1kV and 3kV. Construction techniques are quite different, power-supply caps get pretty exotic, and the output transformers are even more exotic, since they have to withstand peak voltages in excess of 5kV, which tends to substantially increase capacitance and degrade HF response.
The more realistic method of more power in the tube world are paralleled arrays of KT88, KT120, etc. power pentodes. The drawback is that they are pentodes, not as linear as triodes, with many more high-order distortion terms. Global or local feedback reduces distortion overall, but does not change the spectral distribution of harmonics, so pentodes still sound like pentodes.
The so-called "ultra-linear" connection (a 30~40% fractional tap on the transformer primary) is not more linear than a triode; it is simply an intermediate mode of operation between triode and pentode. The UL circuit has an additional drawback of frequently exceeding the maximum voltage rating of the screen grid. This shortens the life of a power pentode that may cost as much as $35 to $100 each (and should be used in matched pairs).
The cheapest vacuum-tube watts are paralleled arrays of Class AB pentodes, and this throws away most of the linearity that tubes can offer. 20 dB or more of feedback is usually required to make the circuit listenable, and this starts to beg the question of simply using a transistor amp, since by this time we have Class AB arrays of nonlinear power devices that require feedback correction.
Selecting devices for linearity instead of power is a very different choice. Class A instead of Class AB, direct-heated triodes in preference to triode-connected pentodes, and using devices in single complementary pairs instead of arrays. Power drops, and prices rise.
If you're in the "all competently designed amps sound the same" school, save yourself the trouble, and don't bother with any of my designs. Seriously, your best loudspeaker is elsewhere. Your choices are much wider, since you have the option of using as many watts as you want, and the price of the watts in the Class AB or Class D world is very reasonable. Even kilowatt amplifiers are not that expensive.
Everything I've worked on since the Ariel is aimed at people who own or have built direct-heated triode amplifiers with outputs in the 3 to 20-watt range. The rest of you have far more choices ... electrostats, planar-magnetics, conventional direct-radiators in the 83 to 88 dB/meter range, and exotics like the MBL. If I didn't have the power monkey looking over my shoulder, I'd probably be working with electrostats.
My 20-watt amps are quite a lot more expensive, with PP 45 drivers, and PP 300B output tubes. For a stereo pair, that's four 45's, and four 300B's (in matched pairs). Even the Chinese tubes are not cheap, and the premium Czech versions are double to triple the price. (Russian tubes are about midway in price.)
The alternatives for more power are AM-radio transmitter tubes that run at B+ voltages between 1kV and 3kV. Construction techniques are quite different, power-supply caps get pretty exotic, and the output transformers are even more exotic, since they have to withstand peak voltages in excess of 5kV, which tends to substantially increase capacitance and degrade HF response.
The more realistic method of more power in the tube world are paralleled arrays of KT88, KT120, etc. power pentodes. The drawback is that they are pentodes, not as linear as triodes, with many more high-order distortion terms. Global or local feedback reduces distortion overall, but does not change the spectral distribution of harmonics, so pentodes still sound like pentodes.
The so-called "ultra-linear" connection (a 30~40% fractional tap on the transformer primary) is not more linear than a triode; it is simply an intermediate mode of operation between triode and pentode. The UL circuit has an additional drawback of frequently exceeding the maximum voltage rating of the screen grid. This shortens the life of a power pentode that may cost as much as $35 to $100 each (and should be used in matched pairs).
The cheapest vacuum-tube watts are paralleled arrays of Class AB pentodes, and this throws away most of the linearity that tubes can offer. 20 dB or more of feedback is usually required to make the circuit listenable, and this starts to beg the question of simply using a transistor amp, since by this time we have Class AB arrays of nonlinear power devices that require feedback correction.
Selecting devices for linearity instead of power is a very different choice. Class A instead of Class AB, direct-heated triodes in preference to triode-connected pentodes, and using devices in single complementary pairs instead of arrays. Power drops, and prices rise.
If you're in the "all competently designed amps sound the same" school, save yourself the trouble, and don't bother with any of my designs. Seriously, your best loudspeaker is elsewhere. Your choices are much wider, since you have the option of using as many watts as you want, and the price of the watts in the Class AB or Class D world is very reasonable. Even kilowatt amplifiers are not that expensive.
Everything I've worked on since the Ariel is aimed at people who own or have built direct-heated triode amplifiers with outputs in the 3 to 20-watt range. The rest of you have far more choices ... electrostats, planar-magnetics, conventional direct-radiators in the 83 to 88 dB/meter range, and exotics like the MBL. If I didn't have the power monkey looking over my shoulder, I'd probably be working with electrostats.
Last edited:
Any news on the comparison between the AE TD15M's and the Altecs?
Hi Gary,
Have you had a chance to compare the TD15M's with the Altec drivers?
I am really curious about your listening impressions - thanks!
Best regards
Peter
Hi Gary,
Have you had a chance to compare the TD15M's with the Altec drivers?
I am really curious about your listening impressions - thanks!
Best regards
Peter
Thanks Lynn, a nice summary of tube amps for a tube newbie like me. Once my speakers are finished to a stage where they can be used in various formats I will build a few low powered tube amps.
Well that's definitely me, but my point was a little different. The transformer saves power, but not sensitivity. If the amp is a voltage source then the power savings doesn't matter because its voltage swing that gives you the SPL (this is my pet-peeve about rating amps in power when speakers are rated in voltage sensitivity). Tube amps are closer to a voltage source than a current source, although they are neither (which IMHO is why they "sound" different). Hence the transformer does not really save anything as regards SPL capability, it just saves on your electric bill.
Output transformers have the convenient feature of 4, 8, and 16-ohm taps, while having (push-pull) primary impedances in the 3.3K to 8K region, depending on which tubes they are designed to match. Since pentodes have high output impedances, if they are used without feedback, the source impedance on each of those (4, 8, 16) taps would be higher than the stated number, probably by several times. This would make the amplifier useless with most loudspeakers, and would result in intolerably boomy bass. (The frequency response of the loudspeaker would track the impedance curve.)
It's standard practice for push-pull pentode amplifiers to have 20 to 26 dB of global feedback, sensed at the 16-ohm tap, and applied at the cathode of the input tube. This reduces the output impedance of each tap to about 1/10 to 1/20th of the rated value (a damping factor of 10 to 20). If power pentodes are connected in ultralinear mode, the damping factor may increase. Similarly, if amplifier uses an additional local loop (cathode feedback from a tertiary winding on the OPT), the damping factor can also be increased. Any time you see transistor-like damping factors on a tube amp, you can be confident it uses multiple feedback loops (with implications for amplifier stability and increased settling time from clipping).
Zero-feedback triode amplifiers typically have damping factors anywhere from 3 to 8. An example: a single-ended 300B with a plate impedance of 700 ohms is connected to an output transformer with a primary impedance of 3500 ohms. 3500/700 = 5, or put another way, the Z out of the 8-ohm tap is 1.6 ohms. There are additional resistive losses in the transformer, but it is industry convention to hold losses (primary+secondary) to less than 5% of the rated impedances.
This also implies that modern loudspeakers with high-order crossovers and vented enclosures with classical T/S alignments may not be a good match for zero-feedback amplifiers, entirely aside from the issue of marginal power supplies and inability to deliver the peak currents the loudspeaker may require. Back-EMF currents from the woofer can saturate the power supply of small amplifiers, and the recovery time may extend to several hundred milliseconds. If this happens, the bass quality will be very poor.
My loudspeakers are designed for zero-feedback amplifiers with Z outs in the 1~2 ohm range; the bass alignments and crossovers are optimized for these output impedances, not the 10~100 milliohm range that characterize transistor amplifiers.
Last edited:
Lynne - I was strictly referring to the proposed transformer for the compression driver, not the amplifier.
And yes, of course, it is possible to design a crossover to any amp output impedance, but not all at once. You have to pick one.
On a slightly related topic, it has occurred to me that there are at least two groups of horn enthusiasts, and we've been talking past each other.
One group, that I do not know personally, appear to come out of the prosound and sound-reinforcement worlds. They're comfortable with modern professional practice of multi-amping and digital equalization, and expect professional levels of dynamic range. There are others who are old-time "hornies" who stayed with the high-efficiency protocol (Klipsch, JBL, Altec, et al) that was standard practice in the Fifties, and never went along with the low-efficiency bookshelf speaker paradigm that arrived in the Sixties.
I'm a member of the johnnie-come-latelies who are part of the DHT revival in the Nineties. We're not from the prosound world. For many of us, horns are a solution to the problem of inherently low-powered amplifiers, not an end in themselves. The folks I know personally are all from this group: without Jean Hiraga, there would have been no triode revival in Japan, Europe, and North America, and following that, no enthusiasm for horns, either.
My goals are different than the first group. I have no interest in using a rack of pro-grade amplifiers and digital equalizers to get the sound of a rock band on stage, or the full THX theater experience at home. Not going there. What I want is the best sound from the type of amplifier, and type of electronics, that I'm interested in. This is probably at cross-purposes with many horn enthusiasts, who have different goals, and different aims for their system.
We're using many of the same tools, but for different purposes. This can lead to misunderstanding.
One group, that I do not know personally, appear to come out of the prosound and sound-reinforcement worlds. They're comfortable with modern professional practice of multi-amping and digital equalization, and expect professional levels of dynamic range. There are others who are old-time "hornies" who stayed with the high-efficiency protocol (Klipsch, JBL, Altec, et al) that was standard practice in the Fifties, and never went along with the low-efficiency bookshelf speaker paradigm that arrived in the Sixties.
I'm a member of the johnnie-come-latelies who are part of the DHT revival in the Nineties. We're not from the prosound world. For many of us, horns are a solution to the problem of inherently low-powered amplifiers, not an end in themselves. The folks I know personally are all from this group: without Jean Hiraga, there would have been no triode revival in Japan, Europe, and North America, and following that, no enthusiasm for horns, either.
My goals are different than the first group. I have no interest in using a rack of pro-grade amplifiers and digital equalizers to get the sound of a rock band on stage, or the full THX theater experience at home. Not going there. What I want is the best sound from the type of amplifier, and type of electronics, that I'm interested in. This is probably at cross-purposes with many horn enthusiasts, who have different goals, and different aims for their system.
We're using many of the same tools, but for different purposes. This can lead to misunderstanding.
Last edited:
Hi Lynne
I have always understood what you are saying, which is why I don't comment on your specific design choices (like Leclerc, tubes, etc.) - we are simply in different worlds.
I do comment when I see things said that are simply incorrect in all worlds - (like damping in a compression driver).
I have little interest in "revival" audio and nostalgic sound quality. I like much more modern music and modern sound quality. And I like HT. I have actually always been as much into film as audio (I made a film in college.) And for film the quality needs to be very much "neutral" without any "nostalgia". We are in different worlds, that is for sure.
I have always understood what you are saying, which is why I don't comment on your specific design choices (like Leclerc, tubes, etc.) - we are simply in different worlds.
I do comment when I see things said that are simply incorrect in all worlds - (like damping in a compression driver).
I have little interest in "revival" audio and nostalgic sound quality. I like much more modern music and modern sound quality. And I like HT. I have actually always been as much into film as audio (I made a film in college.) And for film the quality needs to be very much "neutral" without any "nostalgia". We are in different worlds, that is for sure.