Definitely, Neil. I haven’t heard anything that truly rivals a SET amp with high sensitivity speakers.

SET amps are more of a niche. They are simple designs but the prices are ridiculously high for what it is.

Thumbs-up for "distortion when driven close to their rated voltage". Things have to be designed properly. As Paul said, push-pull requires each 'half' to be complementary. There's a 1980's p-p power amp on my bench which has NO capacitors in the signal path. It's a right pain because every pair of transistors was hand matched in the factory.

Thanks for that I didn't know it was single-ended . I assembled one in 1971 when in the USAF in California . Loved the sound until some junky needed everything while I was away .

The Dynaco 120 was actually class B or i guess you could say class AB. It did run a high bias to keep it in class A longer than modern amps but once you got over a watt or two it went into class B. Early transistor amps had severe crossover distortion problems and there were many tricks to try and alleviate that.

@@dnealukDirect coupled output stages have nothing to do with class of operation. My tweeters run on class A amps with no coupling capacitor*. It's just a matter of using symmetrical power supplies like +24v and -24v so the output signal is at ground with zero signal. There are a lot of other benefits from a DC coupled amp as well such as no capacitor distortion. Another benefit is cost reduction. A stereo amp with a single rail power supply requires three large capacitors, power supply and one for each amp as an output capacitor. A DC amp only needs two, one for each side of the power supply. This is a win-win with lower cost and higher audio quality. *I do have an external 33uf foil capacitors at the tweeters which means that don't ever see below 600hz. Reason is any pop or other low frequencies that get through the amp will quickly destroy a direct connected tweeter. In a normal speaker the crossover provides this protection. I learned this the hard way with two $300 Vifa tweeters destroyed while adjusting the crossover and accidentally letting bass through the tweeter amp.

@@andydelle4509 "Direct coupled output stages have nothing to do with class of operation." You are correct. "It's just a matter of using symmetrical power supplies like +24v and -24v so the output signal is at ground with zero signal." You are incorrect in saying 'JUST a matter'. The two halves must complement each other, as Paul said in the video, with facial expression that said that's not a trivial thing. When carefully designed (in my example, with hand-matched transistors) it's MORE than a matter of symmetrical power supplies.

Me too, I bet they could sound great in hifi circuits.

That is more about output device than amp class. Tubes and FETs become nonlinear and compress as they reach max levels. So they create even order harmonics which are smoother and more musically pleasant. While bipolar transistors just clip when they reach max out and create harsher odd order.

my 8w does just fine on my 96db boston lynfield speakers

Klipsch speakers tend to sound better when driven by tubes amps. You have the right combo

@@giannagiavelli5098 That sounds like fun

I'm not sure that "magical" and "realistic" belong in the same sentence. I'm sure you are hearing more of the former and less of the latter.

@@marianneoelund2940 Some time ago I was involved with the BAS and the MIT Media Lab to quantify the "tube" sound. What it was and how to create it using non-tube electronics. Once we had defined what it was we shared those results and some companies put the solution into their products. One of the easiest ways is to take a very good solid state amplifier and put a tube stage at the input. PS Audio does such a thing with some of their amps. Given a non-audiophile audience, and presented with the "tube" sound, there will be a 90% preference for the tube sound. Even with all the distortion components.

Wow. Years ago, when I was still a kid, someone gave me an old simple AM radio from a '62 Chevy. It had like 4 tubes and a DS503? output transistor. Many years later I learned that it may have been a single ended output circuit, since the radio consumed like 3 amps when on. A waste of fidelity for just AM!

Yeah, some times added musical distortion is what wins in the ears of listeners. Much of the tube hype is about this.

Lol distortion is the enemy of quality

@@Harald_Reindl it depends on your definition of quality. If it is lowest distortion with ruler flat frequency response you can get that in a $100 class d amp off Amazon.

@@Harald_Reindl People who are into adding tubes in the signal chain are actually into added distortion. They happily enjoy their favorite music that has gone through a ton of opamps in the studio to get an extra flavor of “musical tube distortion” in the home.

@@flargosa quality for audio reproduction is simple: don't touch the signal, just amplify it! it's not a instrument and so it must not have a sound

👍

Cool building your own amp. Those Chinese kits are quite cheap nowadays e.g. via Aliexpress. How much did you pay?

Power supply

They generally use single-rail supplies, but you could design one to use bipolar supplies (that would probably not allow removal of the output capacitor, however). The supply choice does not determine whether or not there will be crossover distortion; it is a "given" that there will not be, because there is only a single output device. You are correct that the output becomes asymmetrical as the amplitude increases. This gives rise to even-order harmonics, which many listeners find adds a desirable character to the sound. It is a "warm" type of distortion. The added harmonics may be a stronger contributor to the character of the sound, than the avoidance of crossover distortion.

@@marianneoelund2940 You misunderstood what I said, when you said "... The supply choice does not determine whether or not there will be crossover distortion;... " Look at what I wrote, quoting myself "... no crossover distortion... " my comment relates to the fact that only one type of device that would be used (not a complementary pair like a NPN and a PNP together), and it would be biased on, continuously, therefore no flat spots between the negative and positive parts of a sine wave. Other than that, your comment is interesting. I like that part about the harmonics and pleasing distortion. That's a really interesting subject.

@@marianneoelund2940 Marianne, you got my curiosity about using a split supply for this type of amplifier. Normally with a single ended supply the Q point would be set at half of the rail supply voltage. Which is you had a single +15 volt supply, would mean that the Q point would be set at 7.5 at the collector (assuming and NPN transistor.) Now, if you had a split supply with say +17 volts and -17 volts, you'd have a 34 volt range between rail supplies. I'm trying to imagine where you'd set the collector voltage with DC biasing (Q point.) If you started from the negative supply rail (and that kind of considered that zero volts) If you set the Q point at 17 volts positive, relative to the negative rail, you'd be sitting right in the middle of the load line for the transistor. The output of the transistor could swing negative and positive, with a greater linear range (avoiding asymetrical distortion) at a higher voltage level (peak to peak) on the collector, than you could get with a single ended supply with a lesser voltage on the rail. Because the collector's Q point would be at zero volts (relative to ground), the base would have to be biased (DC) more negative than zero volts, and the emitter voltage would have to be .7 volts more negative than the base voltage (.7 volt drop between emitter and base to keep the device turned on)? You got the gears turning, here. Interesting stuff.

@@KenTeel The quiescent point would largely be chosen based on the voltage swing capability of the supply-side devices, which typically differs from the low-side transistor.

@@marianneoelund2940 Isn't the Q point chosen to be half way between the rail voltage(s), so that a sine wave can swing the same distance from the Q point, going in either the positive direction or negative direction? I've seen examples of the Q point being centered between rail(s) voltage(s)? Perhaps you know different? I'd be curious if they do put the Q point in a place other than centered, and whether that has to do with maintaining the symetry of a sine wave?

Yes, unfortunately this is amplifier theory, and I can only offer a very brief explanation here. At zero audio output level, complimentary (push-pull) output stages have zero current flowing in the transistors, other than the bias current. As output power rises due to increased input amplitude, the transistors carry more current. Single ended amplifiers normally operate in class A. What does this mean as far as power goes? With minor exceptions, a 10W class A amplifier will draw a certain amount of power from the wall, no matter what power is flowing to the speakers. The power load will be the same at zero volume or full volume. They are very inefficient, but they make good room heaters. (The "musicality" people talk about in class A amps is mostly due to harmonics)

Yes, you understand the basic concept. Class AB only maintains a small current through the output transistors when the current to the load is at or near zero. At higher load currents, most of the current flowing through the transistors goes directly to the load. With single output devices, they must be carrying an amount of current at "rest" (i.e., when little power is going to the load) that is at least the magnitude of the maximum peak current the amplifier can deliver to the load.

There are all kinds of monster tubes that you could use to build larger SE amplifiers, but at some point, the electric bill and the heat become intolerable.

@@marianneoelund2940 lol

@@marianneoelund2940 And transmitter tubes require deadly voltages and currents. Not "you may be electrocuted" but rather "WILL be electrocuted" at these power levels. Plus you will need a BIG custom wound output transformer as they are just not available OEM. Power transformers - tons on the used market but OPT s for these amps are very special. Still, there are few built by some wealthy enthusiasts. Personally I don't see the point.

@@andydelle4509 My engineering manager from back in the 1990's used to refer to audio transformer design as a "black art."

I thought I heard EL-34. ( he didn't say he dash tho )

@@bruss529 same same

Yes, the EL34 is a pentode. They can run as triodes if connected up appropriately. A typical modification of one Chinese EL34 kit amp (possibly the one the questioner refers to) is to add a switch so you can go between pentode and triode. Pentode for more power, triode for the 'clean magical' sound being referred to.

@@dnealuk If he's only getting 4 watts from an EL34, it must be connected as a triode.

@@dnealuk I used to own a pair of Audio Research mono amps with 6550's wired as triodes. The amps were not single ended and were class AB. After building my first single ended triode amp, I sold the Audio Research amps. Half the cost of building that first amp was tubes (2) and output transformers (2). EL-34s can indeed be wired in a triode operation, but they are used because they are available and inexpensive, as were my 6550s. I doubt, and I use that word because I don't know for sure, that they would pale against a good power triode like the 2A3, 300b, or my favorite, the 45. I will say that any Class A amplifier will sound better that a similar quality Class AB or D amplifier.

Well that's exactly what I thought. When he mentioned Class A as being something else I was like what!? Isn't a single transistor or tube amplifier class A? It's the same thing! I was yelling at the TV! Lol!

No, that's not true. A single ended amplifier with no output signal draws only its operating bias current. Only when it begins to have a signal does the output current go up. So, a 50 watt SE sitting with nothing on its input is not consuming 50 watts. That happens only when there's a big signal. In contrast, a class A amplifier always draws the same current (wattage) regardless of the input signal. So, a 50 watt class A amplifier draws 100 watts from the AC wall socket even when there is no input signal. That power is then converted to heat. At full input signal the amplifier still draws the same 100 watts but now half of it goes into the loudspeaker load.

@@Paulmcgowanpsaudio Nope. If anything a single ended amplifier draws less current when there is sound due to the diode effect of the grid and cathode.

When it's broken and needs repair. On a more serious note, it's just semantics. When one speaks of Class A, it usually implies class A complementary, rather than single ended. But both use similar biasing schemes.

@@marianneoelund2940 class A complementary is just class B with heavy bias (class "AB"). At least that's how I always looked at it. Class AB sound fine but zero feedback class A is the best, in my opinion :)

@@cheapygeek That's a fair basic description. The key for class A is that the output transistors are never turned completely off. But there are variations in how much bias current is used, and how the opposite-side current is tapered, so class A designs may sound different, and may not handle reactive loads the same.

Single ended can be both class A and Class B or AB, it just depends on biasing. AB is always a bit confusing because it is often referred to a push pull configuration, but in reality those are typically biassed AB, though you can also bias them A (if you require a space heater) or B. And then there is even class C which is only useful for railway station PA systems.

@@hugobloemers4425 I'd like to see a schematic for one of those "single-ended Class B" designs.

True, but 300B tubes are pretty expensive…

The first watt is pure class A but beyond that A/B. The point was they use a unique output stage with only one sex output device, an N channel, where normally the complimentary output would be N and P.

@@octaverecordsanddsdstudios1285 Isn't the definition of a "class AB amplifier" one that operates in class A up to a certain current output, then switches to class B at a point determined by the biasing when more current is needed?

@@gotham61 Yes, that's correct. Class A for a small portion of the signal (where the amplifier bias current is greater than the signal current) and then class B above that.

Are you saying different types of amplifier topologies all sound the same? Class A, A/B, D, etc?

@@cha7260 when good done and with proper room-EQ they should - a neutral/linear frequency curve is a very simple and strict thing which can be measured and adjusted - unless that is done you have a problem when it comes to the "Fi" of "HiFi"