Finally I had the opportunity to get my hands on one of these “sacred monsters” so praised by audiophiles. Therefore I kindly ask the admirers of this brand not to take it badly if I point out some technical differences. The first thing I want to highlight is the schematic and the variations compared to the 2A3 version. Below you can see both schematics: SUN Audio 2A3 and 300B…

2A3	300B

What changes between the two schematics? In reality, nothing. Or better, the 2A3 has been replaced with a 300B, the value of the cathode bias resistor has been modified and the 300B has been connected to the 3500 ohm tap of the transformer instead of the 2500 ohm one. Everything else remains the same. However there is a problem: the 300B version delivers 4 watt at clipping (3 watt undistorted). In the photo below you can see on two of my oscilloscopes (2 volt per division scale) the amplifier on an 8 ohm resistive load. The amplitude visible is 16 volt peak to peak, corresponding to about 4 watt RMS…

For those who may still have doubts, I simulated the circuit on Spice obtaining the same results. If you want to test the schematic it is available above and you can find it on many other websites. You can download LTspice for free here…

The manufacturer generically states 8 watt, that is 4+4, counting on many people assuming it means 8 watt per channel, which would be a good power for a 300B. In reality it is 4 watt per channel (less than 4 if considering the power before clipping begins). The 300B has been run at the same dissipation as a 2A3. Personally I think they could have designed a circuit tailored for the 300B instead of recycling the 2A3 design, especially considering that these devices are not sold for 300 euro. They relied on the usual 300B gimmick to sell, but in fact the valve is not used to its full potential. The amplifier was almost original, built in 1995, but it showed several age related issues. The first concerned the worn out sockets of the 300B and the rectifier. I was told the valves are still the original ones, or at least that they have never been replaced by the current owner. However I do not believe that the sockets were this worn from the factory, especially since the 6SN7 sockets were still tight. I think they loosened due to continuous valve swapping in search of a sound that could not change by swapping valves alone. Many audiophiles fall into this compulsive tube rolling, but unfortunately the sound is determined by the whole circuit and by the transformers, where the valves are not even the most important element. Changing valves may slightly alter the tone because of manufacturing tolerances between one valve and another, but these are only marginal changes.

The second problem concerned the two stepped attenuators (fake Chinese ALPS, not original) which were damaged, extremely stiff and had two ugly plastic knobs unpleasant to touch.

The third problem concerned the cathode bypass electrolytic capacitors, which were not high quality and over the years their ESR had increased above 1ohm. The restoration and upgrade involved replacing the two stepped attenuators with original ALPS potentiometers (personally I would have preferred to keep the stepped attenuators as with two separate volumes per channel you can match them more easily, but the customer wanted standard potentiometers). The worn sockets and the various cathode bypass capacitors were replaced with higher quality ones with higher capacitance (the reasons for this choice will be explained later in the instrumental section). The only small circuit change was a slight adjustment to the cathode resistor value of the power stage, adding another resistor in parallel to slightly increase the BIAS current since on LTSpice it seemed to be the only possible improvement.

Now let’s move on to the measurements:

Power: 4 watt RMS per channel at full clipping
Damping factor DF: 2.66 equivalent to an output resistance of 3ohm

THD distortion

THD at 1 watt	THD at 3 watt

As you can see, all harmonics are present up to the sixth (after that they become so small that they are irrelevant). They are not perfectly stair stepped and already at 1 watt you see a third harmonic almost as large as the second. Increasing the level up to 3 watt (before clipping begins) the third harmonic stands out over all the others. But how is this possible? A single ended triode amplifier with an original SUN Audio 300B and Tamura transformers generating strong third harmonic. How can this happen?

I want to underline that this is not a criticism of SUN Audio, this distortion behaviour is completely normal for a zero feedback circuit. At 1 watt the total THD is about 1 percent and at 3 watt about 3 percent. Out of curiosity, what does Spice simulate?

Very similar to what I measured at 1 watt. In reality everything can change if the valves are replaced with others of different type or brand. What really matters is that the total THD remains low. Now you have a clear idea about the issue of even and odd harmonics.

Bandwidth: 35Hz – 25kHz minus 1dB at 1 watt. The factory declared frequency response is 15Hz – 30kHz minus 3dB and here it is even better because the minus 3dB in the high range is at 50kHz and not at 30.

Graph on resistive load

And on reactive load

Now let’s look at my modification which consists of increasing the cathode bypass capacitance and slightly increasing the BIAS current. The resulting bandwidth is 20Hz – 30kHz with a reduction of 1dB and a notable improvement in the phase shift at 20Hz, which goes from 60 degrees to 40 degrees. This allows you to perceive a bit more low end when the amplifier is paired with single driver loudspeakers, as it is meant to be. If you want to carry out this modification you can contact me. It is important to underline that an indiscriminate increase in bypass capacitance can cause an area of instability where the amplifier may start to oscillate at low frequencies or stress the rectifier valve.

Below you can see the analysis of the spectrum after the bias current modification (the 300B is still well below its maximum dissipation, 16 watt out of 36).

Below are the square waves at 100Hz, 1kHz and 10kHz

The transformers installed in this device are adequate. I had the chance to listen to this amplifier in my listening room equipped with Tannoy reflex speakers. In this environment a zero feedback amplifier is not in ideal conditions. The mid high frequencies are acceptable although I find female vocals a bit piercing. However the poorly defined bass ruins the listening experience. An amplifier like this is better matched with single driver speakers. One modification that could be done, considering the excess gain of the circuit, is adding a bit of negative feedback. Imagine the faces of some people when hearing about feedback in a 300B amplifier. But trust me, it would lead to a significant improvement. It would preserve the same nice and clean mid high characteristics (thanks to the quality of the output transformers) and would allow pairing with reflex speakers for a more balanced listening experience with more present low frequencies. Anyone who believes that such an amplifier could be superior to other circuits, even with less expensive and less fashionable valves, is mistaken.

A user told me that some other Italian technicians completely modify the driver stage and asked me whether this could increase the power. The answer is NO. To increase the output power of a 300B you need to change the transformer impedance, increase the plate voltage and bias current and modify the driver to better drive the final stage. In practice the whole circuit must be changed. Modifying the driver alone provides no benefit because the 300B will still clip at 4 watt regardless of the driver valve used. Moreover the biggest limitation is the lack of damping (lack of negative feedback) so changing the driver circuit does not give real improvements. Be cautious with those who promise miraculous improvements with pointless modifications. To significantly change this amplifier, you would need to completely redesign it, but such invasive interventions are usually not worth it.