Air Tight ATM-211 tube monoblock power amplifier Measurements
With the bright glow of its directly heated output tube lighting up the room, the Air Tight ATM-211 amplifier was very different from the products that usually pass through my measurement lab. I also had problems getting it to behave at first. When its AC supply lead was grounded, the amplifier was unstable at output levels of more than a few tens of milliwatts, going into bursts of ultrasonic oscillation. Keeping the Audio Precision System One grounded and floating the Air Tight's supply ground with a cheater plug solved the instability problem, though this did nothing to resolve the low levels of supply hum in the amp's output.
Before performing any measurements, I set the ATM-211's output tube bias using its built-in meter. Voltage gain into 8 ohms was a respectable 27.6dB, and the amplifier inverted signal polarity. Although its input impedance was high, there was some dependence on the actual measured impedance with both frequency and the setting of the level control. Worst case was 41k ohms at 20kHz with the control set to its maximum. With the control set to its midpoint, the input impedance was 71k ohms at both 1kHz and 20kHz.
There was no indication of the output transformer tap, but RD noted that the review samples had been set to 8 ohms. Nevertheless, the output impedance measured a high 3.7 ohms at 1kHz. Unusually, this decreased slightly to 3.5 ohms at 20Hz and 20kHz. Nevertheless, there was a large (±2dB) variation in frequency response when the amplifier was used to drive our standard simulated loudspeaker (fig.1, top trace), and a significant drop in level each time the load impedance was halved (fig.1, lower traces). The high frequencies rolled off by 0.9dB at 20kHz into 8 ohms, somewhat less into lower impedances.
Fig.1 Air Tight ATM-211, frequency response at (from top to bottom at 2kHz): 2.83V into dummy loudspeaker load, 1W into 8 ohms, 2W into 4 ohms, 4W into 2 ohms (0.5dB/vertical div.).
Although it can't be seen in this graph, the ultrasonic rolloff of the ATM-211's response is broken by a slight peak just under 80kHz, which leads to a slight overshoot and a couple of cycles of ringing in the amplifier's reproduction of a 10kHz squarewave (fig.2).
Fig.2 Air Tight ATM-211, small-signal 10kHz squarewave into 8 ohms.
Despite it not having a negative feedback loop, the Air Tight is quite linear at midrange frequencies, provided it drives high impedances. Fig.3 shows that its THD level hovers just above the 0.1% level into 16 ohms between 100Hz and 1kHz. However, this graph also reveals that the distortion rises significantly with decreasing load and at frequency extremes. Regardless of load, the THD is more than 1% above 10kHz, with a maximum at half the frequency of the ultrasonic peak noted in figs.1 and 2.
Fig.3 Air Tight ATM-211, THD+N (%) vs frequency at 2.83V into (from bottom to top at 1kHz): simulated loudspeaker load, 16 ohms, 8 ohms, 4 ohms, 2 ohms.
The waveform of the residual distortion in the midrange (not shown) is heavily second- and third-harmonic in nature. At low frequencies (fig.4), the second harmonic is joined by even higher harmonics, these decreasing in level with increasing order. This is something that, a quarter-century ago, the French writer Jean Hiraga showed was associated with a subjectively pleasing nature, despite what might appear to be a high level of distortion in absolute terms. The picture is a little different in the midrange (fig.5), due to the even-order harmonics dropping in level compared with the odd-order ones. You can see in fig.4, by the way, the 120Hz hum that I couldn't eliminate with the amplifier grounded through the input signal lead. At -62dB relative to 1W into 8 ohms, this will be audible, as Bob Deutsch found in his auditioning.
Fig.4 Air Tight ATM-211, spectrum of 50Hz sinewave, DC-1kHz, at 5.8W into 8 ohms (linear frequency scale).
Fig.5 Air Tight ATM-211, spectrum of 50Hz sinewave, DC-22kHz, at 4W into 16 ohms (linear frequency scale).
The consequences of the Air Tight's poor linearity at high frequencies can be seen in fig.6. Even at the low power levels I used to get these data—just below visible waveform clipping on the oscilloscope screen—there are very high levels of intermodulation products in the amplifier's output. That RD noted nothing amiss I can put down only to the fact that music program is generally not as demanding as the twin-tone test signal, and that his horn-loaded Avantgarde speakers are so sensitive that he didn't ask the amplifier to deliver power levels close to the 2W featured in this graph. Nevertheless, I don't like to see this kind of behavior.
Fig.6 Air Tight ATM-211, HF intermodulation spectrum, DC-24kHz, 19+20kHz at 2W into 8 ohms (linear frequency scale).
With its brightly glowing output tube, the Air Tight amplifier is specified at 22W into 8 ohms (13.4dBW)—quite a high power for a single-ended design. Yet plotting its 1kHz output power vs THD+noise into resistive loads (fig.7) reveals that the usual 1% THD clipping point has to be relaxed to 3% for the amplifier to meet its specified power into 8 ohms. This graph confirms that the amplifier is most linear into 16 ohms, but the fact that slightly more power is available into 8 ohms confirms that this is the load best matched to the output transformer. Only a few watts are available into 4 and 2 ohms, however.
Fig.7 Air Tight ATM-211, distortion (%) vs continuous output power into (from bottom to top at 1W): 16 ohms, 8 ohms, 4 ohms, 2 ohms.
Its poor high-frequency linearity, incipient ultrasonic instability, and the presence of supply hum in its output are all factors that would rule the ATM-211 out for me, even if I had speakers that were sensitive enough and had a high enough modulus of impedance to minimize its measured problems. But, as RD found, if you can meet those conditions, the Air Tight amplifier appears to be able to produce a sound much better than its measured performance implies.—John Atkinson