What’s the world’s best 2 watt amplifier?

There has to be more to the “sound” of rectifier tubes than just the different output voltages. Let me give two illustrations. First, different brands of the same tube can sound very different, and consistently so—-a GE 5u4GB sounds different from a Sylvania 5u4GB and from an RCA 5u4GB despite all of them having the same output voltage. Second, I have a linestage with a regulated B+ voltage—-whether the rectifier outputs 350 or 390 volts the B+ is held tightly at 275 volts, yet different rectifiers sound different.
Presumably the physical (materials/assembly) differences between tubes might also have an audible impact, but I don't care why they sound different. I've done comparative listening among a half dozen different rectifier tubes and the blue Sophia 274B and 1950's U52 CV575 5U4G are in a league of their own at least in the Modwright PS 9.0.
 
There has to be more to the “sound” of rectifier tubes than just the different output voltages. Let me give two illustrations. First, different brands of the same tube can sound very different, and consistently so—-a GE 5u4GB sounds different from a Sylvania 5u4GB and from an RCA 5u4GB despite all of them having the same output voltage. Second, I have a linestage with a regulated B+ voltage—-whether the rectifier outputs 350 or 390 volts the B+ is held tightly at 275 volts, yet different rectifiers sound different.
Did you measure the actual output voltages?? In the linestage, is the filament supply regulated? I would assume so, but the filament supply for the rectifier is probably not. More on that later...

In order to know the voltage drop, its not enough to measure the output voltage- you have to measure the input voltage as well. The line voltage can change as well as the voltage of the transformer itself, based on how its loaded (due to internal resistance).

I would also suggest paying attention to the waveform that comes out of the tube, but if the power supply is doing its job, this should be of less consequence; the output of the power supply should be a straight line. Seeing how well a regulated supply can react to variable input voltages will tell you a lot about its design. Since AC line voltage can vary quite a lot, we spec our regulators to not drop out until the line voltage sags below 105VAC. Based on that you've mentioned above, it sounds like a tube regulator is involved in your line section. Were its filament supplies regulated?

One variable I didn't mention is how the tube reacts to its cathode/filament voltage. Obviously with a lower filament voltage the tube will have a higher voltage drop. So you'd need to keep track of that bit of information as well.

Have you quantified how the rectifiers 'sound different'? One way to do this would be to run a distortion check, using a microphone placed at the listening position, also to run a frequency sweep with the mic at the place. I have customers who have auditioned different parts and have used this technique to show what it was they heard ('before' and 'after'). So often the measurement people complain about 'anecdotal evidence' but the kicker is that its pretty easy to sort out what's going on if you figure out what to measure.

Finally, confirmation bias is a really tricky beast. We are programmed to think that the more expensive a thing is, the more value or performance it has (the Veblen Effect) which can alter our perception of a thing like a tube, if that tube happens to be rare and pricey. Being different from a wine though, tubes are knowable in that they can be measured and correlations exist between how well they behave and what they sound like. I think the biggest issue here with the use of a solid state rectifier system is the reticence of an SET owner to allow semiconductors in that might 'pollute' the setup, based entirely on that notion and nothing else. That is why I mentioned the variable voltage system in my prior post, since this is a topic guitar players have dealt with a lot; in this way the voltage drop issue can be eliminated, allowing for a more even playing field between the various rectifier types.

Put another way, finding that there is a simple method of getting the amp to 'sound better' is a thing that most audiophiles would find attractive.
 
@atmosphere
Ralph, I appreciate the effort spent in replying to my comment, but it appears we are approaching this matter from different perspectives. When I am building something, I use my DVM to confirm the circuit is working correctly. I don’t routinely measure power supply voltages after that and hardly ever after doing something like changing rectifier tubes. And I don’t routinely measure my house line voltage.

To answer your questions, the linestage I mentioned has a solid state regulator for the B+ and it has a simple unregulated DC supply for the 12b4a signal tubes. i do not measure its operating voltages while I am listening to it. I also do not have distortion measuring equipment (other than my ears).

Coincidentally I did install meter tip jacks on amps I built recently (Western Electric 124 clones) so I am able to measure plate, screen and cathode voltages on the output tubes while the amps are operating. These voltages do change slightly as the amps warm up and they change very slightly from day to day, but I have never noticed any change in voltages after swapping 5u4GB rectifiers. Yet the different brands of 5u4GB sound different and the differences are consistent. By consistent I mean the RCAs sound the same, the GEs sound the same, Tung Sols sound the same; yet the different brands sound quite different from each other.

You may say this is all due to confirmation bias. For all I know maybe that is a factor but I am not going to lose sleep wondering about it. Unlike you I am just a hobbyist. My hifi only needs to please one person—me!
 
@atmosphere
Ralph, I appreciate the effort spent in replying to my comment, but it appears we are approaching this matter from different perspectives. When I am building something, I use my DVM to confirm the circuit is working correctly. I don’t routinely measure power supply voltages after that and hardly ever after doing something like changing rectifier tubes. And I don’t routinely measure my house line voltage.

To answer your questions, the linestage I mentioned has a solid state regulator for the B+ and it has a simple unregulated DC supply for the 12b4a signal tubes. i do not measure its operating voltages while I am listening to it. I also do not have distortion measuring equipment (other than my ears).

Coincidentally I did install meter tip jacks on amps I built recently (Western Electric 124 clones) so I am able to measure plate, screen and cathode voltages on the output tubes while the amps are operating. These voltages do change slightly as the amps warm up and they change very slightly from day to day, but I have never noticed any change in voltages after swapping 5u4GB rectifiers. Yet the different brands of 5u4GB sound different and the differences are consistent. By consistent I mean the RCAs sound the same, the GEs sound the same, Tung Sols sound the same; yet the different brands sound quite different from each other.

You may say this is all due to confirmation bias. For all I know maybe that is a factor but I am not going to lose sleep wondering about it. Unlike you I am just a hobbyist. My hifi only needs to please one person—me!
The type of regulation also matter. If this is a series voltage regulator, it can cope with over-voltage, but not under-voltage. You need a shunt regulator if you want to have tight regulation when the supply voltage drops. This could account for the difference you detect in rectifiers.
 
The 421A [ snip ] Might be fun to build a PP amp using that tube.
Interesting idea. What kind of power would such an amp reasonably produce?
 
@atmosphere
Ralph, I appreciate the effort spent in replying to my comment, but it appears we are approaching this matter from different perspectives. When I am building something, I use my DVM to confirm the circuit is working correctly. I don’t routinely measure power supply voltages after that and hardly ever after doing something like changing rectifier tubes. And I don’t routinely measure my house line voltage.

To answer your questions, the linestage I mentioned has a solid state regulator for the B+ and it has a simple unregulated DC supply for the 12b4a signal tubes. i do not measure its operating voltages while I am listening to it. I also do not have distortion measuring equipment (other than my ears).

Coincidentally I did install meter tip jacks on amps I built recently (Western Electric 124 clones) so I am able to measure plate, screen and cathode voltages on the output tubes while the amps are operating. These voltages do change slightly as the amps warm up and they change very slightly from day to day, but I have never noticed any change in voltages after swapping 5u4GB rectifiers. Yet the different brands of 5u4GB sound different and the differences are consistent. By consistent I mean the RCAs sound the same, the GEs sound the same, Tung Sols sound the same; yet the different brands sound quite different from each other.

You may say this is all due to confirmation bias. For all I know maybe that is a factor but I am not going to lose sleep wondering about it. Unlike you I am just a hobbyist. My hifi only needs to please one person—me!
An oscilloscope and sine/square wave generator can tell you a lot about how a circuit is behaving. If you are building stuff this equipment is not expensive and can help you know if the circuit is really working right- sometimes voltages are not enough!

If you really are hearing differences with rectifiers (having eliminated other variables) then you are faced with the situation of something not being right in your power supply. FWIW, one issue with tube rectifiers is that they cannot support a large amount of filter capacitance at their output. As they warm up and begin charging that capacitance, if the capacitance is too large it can cause the rectifier to fail. So traditionally the value is chosen to not exceed the limit of the rectifier, rather than set up a timing constant that is lower than the LF cutoff frequency of the amplifier! When the power supply has timing constants higher than that of the amplifier, this means that low frequencies can modulate the power supply, thus introducing IM distortion to the amplifier.

In an SET, the defense against this is the class A operation of the amp, which in theory presents a constant load. The problem with this is that the circuit really isn't perfect. This simple reason is why it can be worth it to overbuild a power supply so as to minimize its effect on an imperfect circuit.

IM is usually pretty audible. To prevent this problem in our tube amps, we use a different power transformer for the output section as opposed the the driver section- and made sure both had timing constants lower than the cutoff frequency of the amplifier, which in our case is defined by the coupling capacitors.

Placing an oscilloscope on the B+ rails and then putting the circuit through its paces can tell you if the power supply is up to snuff. At all signal levels (including total overload) and regardless of the type of signal, you should see a straight line on the 'scope on its highest gain setting. If you see a signal related to the audio in any way, you know you have a problem.
 
An oscilloscope and sine/square wave generator can tell you a lot about how a circuit is behaving. If you are building stuff this equipment is not expensive and can help you know if the circuit is really working right- sometimes voltages are not enough!

If you really are hearing differences with rectifiers (having eliminated other variables) then you are faced with the situation of something not being right in your power supply. FWIW, one issue with tube rectifiers is that they cannot support a large amount of filter capacitance at their output. As they warm up and begin charging that capacitance, if the capacitance is too large it can cause the rectifier to fail. So traditionally the value is chosen to not exceed the limit of the rectifier, rather than set up a timing constant that is lower than the LF cutoff frequency of the amplifier! When the power supply has timing constants higher than that of the amplifier, this means that low frequencies can modulate the power supply, thus introducing IM distortion to the amplifier.

In an SET, the defense against this is the class A operation of the amp, which in theory presents a constant load. The problem with this is that the circuit really isn't perfect. This simple reason is why it can be worth it to overbuild a power supply so as to minimize its effect on an imperfect circuit.

IM is usually pretty audible. To prevent this problem in our tube amps, we use a different power transformer for the output section as opposed the the driver section- and made sure both had timing constants lower than the cutoff frequency of the amplifier, which in our case is defined by the coupling capacitors.

Placing an oscilloscope on the B+ rails and then putting the circuit through its paces can tell you if the power supply is up to snuff. At all signal levels (including total overload) and regardless of the type of signal, you should see a straight line on the 'scope on its highest gain setting. If you see a signal related to the audio in any way, you know you have a problem.
Would a choke input power supply solve this problem ? There is of course a penalty in voltage drop using choke input.
 
Would a choke input power supply solve this problem ? There is of course a penalty in voltage drop using choke input.
Its all about timing constants. Where is the plate B+ for the power tube taken off? If the timing constant there is higher than that of the amplifier, audible effects are unavoidable. Usually chokes are designed to filter the 50 or 60Hz sawtooth waveform coming out of the rectifier, but this is in tandem with the capacitive bypass at its output. That needs to be large enough that the resulting time constant is low enough that the amp can't modulate that node in the power supply.

Again, an oscilloscope can help by allowing you to observe what happens to the B+ when the circuit is put through its paces.

If you've wondered why battery power can sound so nice, this is why. Batteries are true DC with no real time constant. So the amplifier circuit can't modulate them- they present a very low impedance; it would be like taking a sip out of Lake Superior instead of a sip out of a glass.
 
  • Like
Reactions: aLLeARS and carolus
If that’s your standard for “adequate “ sound levels, good for you. Not for me. I attend live music venues quite often and love the experience. I realized a long time ago that I do not n/want to replicate that volume at home for my extended and frequent listening sessions.

I enjoy my audio system listening sessions immensely at what I find to be very comfortable and engaging levels. But clearly that’s me. Obviously you do what works best snd satisfies you.
Charles
I agree completely. I do not need or want volume levels at home that are as loud as a live performance. In fact, as much as I love going to live performances, the volume levels are often too loud for my 70-year old ears, especially if there is any amplification.
 
….
If you've wondered why battery power can sound so nice, this is why. Batteries are true DC with no real time constant. So the amplifier circuit can't modulate them- they present a very low impedance; it would be like taking a sip out of Lake Superior instead of a sip out of a glass.
I thought Duluth had that fiasco 30 years with the asbestos in the water?
;)
 
  • Like
Reactions: Atmasphere
Try the Decware SE84UFO25 25th anniversary amp 2.3 watts per channel or 6 watts mono, or you could try the Decware Mini Torrii 4 watts per channel. Lifetime Warranty
 
  • Like
Reactions: acousticsguru
What speakers would mate well with a 2.3W amp?
I have 105dB speakers and even in a small (12’x13’) room my Kondo Japan Souga 2x2A3(7W) was not sufficient.
I love the Souga and will keep it for another room; but no idea what speakers to buy.
Even for a desktop system that never draws more than ~0.5W, the Souga didn’t cut it. And the Souga is one of the best 2A3 amps available.
I guess I’m just not sold (at all) on low watt amps; but I don’t want to sell the Souga…
 
What speakers would mate well with a 2.3W amp?
I have 105dB speakers and even in a small (12’x13’) room my Kondo Japan Souga 2x2A3(7W) was not sufficient.
I love the Souga and will keep it for another room; but no idea what speakers to buy.
Even for a desktop system that never draws more than ~0.5W, the Souga didn’t cut it. And the Souga is one of the best 2A3 amps available.
I guess I’m just not sold (at all) on low watt amps; but I don’t want to sell the Souga…
I don’t know what going on with your audio system. Your outcome doesn’t make sense to me. However, I accept what you report for your situation.

A good friend uses 2A3 SET mono blocks (3 watts) to drive his Cube Audio Nenuphar speakers (92 db /8 ohm/single driver) with outstanding sound quality and music presentation. YMMV I suppose.
Charles
 
105 db speakers can be powered by “milli” watts at very reasonable listening volumes.
Charles
 
  • Like
Reactions: carolus
I think it’s about what the expectations are of “good sound”. I used to think the Souga was fantastic with my 105dB horns. Until I heard what the 150W DarTZeel can do with the same horns. No high frequency roll off, lighting fast, prodigious bass…
I guess it Is what is — the Zeel is just in its own (or very small) class…
I think you have to be willing to forego recreating dynamics that mirror a live performance (e.g., 105dB peaks) with a low power amp. It’s just not going to happen. And that’s OK If that’s not your intent.
 
Try the Decware SE84UFO25 25th anniversary amp 2.3 watts per channel or 6 watts mono, or you could try the Decware Mini Torrii 4 watts per channel. Lifetime Warranty
I have a Decware 25th Anniversary Zen Triode, the pimped version with V-Cap CuTF and stepped attenuators. I absolutely love it. Having said that, I may not have heard it at its full potential.

I'm saying this because so far, I've heard it with Acapella Fidelio, Boenicke W22, EJ Jordan exponential horn, hORNS Universum, Odeon La Bohème, and a pair of my own speakers, and have come to the conclusion that a) nominal sensitivity ratings don't tell the whole story, and/or b) (almost) everybody lies.

I've always loved low-powered amplifiers, single-ended triodes in particular, since I heard a number of Altec, RCA and WE setups in my now seemingly distant youth, and since then in a number of other systems spanning the whole range from pretty nice to spectacular.

The SE84UFO25 uses triode-wired pentodes as output tubes, EL84 or, as per Steve Deckert, the Russian 6P15P-EV. The reason I was curious about getting one was that this is the only amp I've ever seen that uses a voltage regulator tube each for every audio tube, plus tube rectifier (there are other amps that use VR tubes for input and/or driver tubes, from Decware, Lamm etc., but none that I know of that does for all including the output tubes).

What this does is it sounds as if running on battery power, with a sound that "carries" at low level unlike any amp I've owned or tried. I'll admit I primarily bought it hoping it would be fun for late-night listening, and indeed it has, in spades, what I love most about low-powered amps: that intimate, emotional connection to the music.

It does this in a balanced and harmonious, transparent, not particularly "tubey", let alone colored fashion, depending on the tubes used it also sounds convincingly large-scaled and impactful. It does walk that fine line of never sounding bright or harsh which I love about tube gear, though.

The problem is one might wrongly conclude it's all that or near useless - depending on the speaker one pairs it with. The nominal sensitivity rating has so far proven to be an inadequate indicator of whether a combination is going to produce that intimate, emotional "magic" it is capable of.

Depending on the speaker used, it'll sound dynamically flat, not to mention start compressing macro dynamics well before it runs out of juice (it'll do so visibly, i.e. once the 0A3 voltage regulator tubes start sagging, they start flickering to e.g. the bass line in the music - something I've only seen them do once, with Boenicke's W22 speakers in an admittedly large room, but then, because only one VR tube would do this, Steve Deckert recommended I replace the 0A3s, which I subsequently did with even older NOS RCA that make the amp sound even more powerful).

I'm saying "even more powerful" because that's what "little Ms D", as Christoph fondly calls it, sounds like at its best. With my speakers, the sheer volume and density of the sound, soundstage and scale, not to mention powerful and textured bass, are nothing short of amazing (the bass is only fractionally less tight than with Spectral solid-stage amplification on the same speakers).

Now, the speakers it synergizes with best are currently placed in a very small room, plus they're extremely easy to drive, by which I mean, no anomalies such as steep electrical phase angles, no sudden impedance dips, none of those problems speaker designers wisely forget to mention…

As a matter of fact, Steve Deckert might not have recommended using this amp with speakers that only offer 91.5dB sensitivity, but when I told him I built and measured them myself, that the drivers are time-aligned, using my own phase-coherent filters, in short, that this isn't about some "nominal" sensitivity rating spotted in some manufacturer's sales ad blurb, but an actual calibrated 2.83V/1m measurement, he immediately gave his thumbs up.

It's by far the least expensive component in this system, literally, the interconnects used with it cost more, but doesn't at all seem out of place. It performs beautifully, a technically minimalistic, wonderfully direct communicator of music, atmosphere and emotion. Maybe my favorite tube compliment includes the Reflektor 6N23P Single Wire Silver Shield from 1975 as input/driver with a Miniwatt GZ34 Metal Base rectifier from 1957, alternatively Shuguang GZ480. Mind you, I'm not just playing chamber music, but symphonic, church organ, and it's occasionally freaked me out flooding the room with bass, seemingly shaking the foundations with music such as Max Richter's Sleep (non-fatiguing listened to complete, the whole eight and a half hours, a positively uplifting experience). All this without ever showing signs of output transformer saturation of e.g. my 211 amp driving the Jordan horns.

Christoph liked the little "Deckie" quite well with his Universums, a combination I'd not recommend, especially not in a room that size, even though it really only hinted at macro-dynamic "flattening" with some music and not all (but then, what's the point?). It didn't do particularly well with the Boenicke W22, nor the Odeon La Bohème, both of which are touted "high-sensitivity" speakers. The combination with the EJ Jordan resulted in way too much bass, and the most unlikely combination, with the Acapella Fidelio, another remarkably time- and phase-coherent speaker, works well up to a point, that is beautifully, if only at low to medium volume level.

Should perhaps mention it's really another power amp with input attenuators, that depending on the source component may or may not profit from using a preamp. As usual, V-Cap Elite Reference coupling caps as an option are not for the impatient, in dire need of 400-600 hours burn-in to sound at their best (having said that, they don't sound nowhere as "vile" during this time as Steve Deckert repeatedly warned me, on the contrary, the same caps seemed to take rather longer, starting out sounding analytically dry, elsewhere, such as in the Amplifon SET42, where they're also an available and, ultimately after a seemingly endless wait, highly recommended option).

If there's anything I've learnt from this journey, then to refrain from jumping to conclusions, as in trying to make so-called educated guesses as to which combinations may or may not work synergistically using such a flea-watt amp. If there's anything that I know doesn't work, that's it.

Greetings from Switzerland, David.
 
Last edited:
What speakers would mate well with a 2.3W amp?
I have 105dB speakers and even in a small (12’x13’) room my Kondo Japan Souga 2x2A3(7W) was not sufficient.
I love the Souga and will keep it for another room; but no idea what speakers to buy.
Even for a desktop system that never draws more than ~0.5W, the Souga didn’t cut it. And the Souga is one of the best 2A3 amps available.
I guess I’m just not sold (at all) on low watt amps; but I don’t want to sell the Souga…
We have used the Souga to drive the Living Voice VOX Olympian speakers (including the Elysian bass horns separately driven by class A transistor bass amps). The result was far less satisfactory than using the Gakuoh 300B push-pull or especially the Lamm ML-3, both generating about 30W. The VOX is rated at 105dB. The Souga could drive the system to a very loud level, but the scale of the sound (soundstage, depth) and the impact (dynamics, bass extension) were lacking. It did well on simple acoustic music such as a folk singer with a guitar etc., but a large scale orchestral work would show its deficiencies. It is not so much the max. SPL it can generate with the speakers, but how well it controls the drivers and how much distortion is generated in the course of generating those decibels. I think it would work if it is only used to drive the mid-range horns, for example, in a multi-amp setup with active crossover. In that case, it would avoid the passive crossover and the compression drivers have much lighter diaphragms than the bass drivers and hence much easier to drive. This is where I would use such an amp. I have used my Brook 12A (2A3 push-pull) monoblocks to drive my midrange compression horns (111dB sensitive field coil with 4" beryllium diaphragm) directly, and they worked beautifully.
 
We have used the Souga to drive the Living Voice VOX Olympian speakers (including the Elysian bass horns separately driven by class A transistor bass amps). The result was far less satisfactory than using the Gakuoh 300B push-pull or especially the Lamm ML-3, both generating about 30W. The VOX is rated at 105dB. The Souga could drive the system to a very loud level, but the scale of the sound (soundstage, depth) and the impact (dynamics, bass extension) were lacking. It did well on simple acoustic music such as a folk singer with a guitar etc., but a large scale orchestral work would show its deficiencies. It is not so much the max. SPL it can generate with the speakers, but how well it controls the drivers and how much distortion is generated in the course of generating those decibels. I think it would work if it is only used to drive the mid-range horns, for example, in a multi-amp setup with active crossover. In that case, it would avoid the passive crossover and the compression drivers have much lighter diaphragms than the bass drivers and hence much easier to drive. This is where I would use such an amp. I have used my Brook 12A (2A3 push-pull) monoblocks to drive my midrange compression horns (111dB sensitive field coil with 4" beryllium diaphragm) directly, and they worked beautifully.
Thanks, this is similar to what I found. Certainly it seems due to power, as the Gakuoh 300B has a lot more power (30W) than the Souga's (7W). I'm not trying to knock low power amps, they are just quite limited in their applications.
I do want to find a suitable pair of speakers for the Souga - perhaps with an active/powered sub like Jeff Jackson's latest?
What speakers are you using (111dB field coils)...?
 

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu