TESTSCH.TXT This file accompanies TESTSCH.GIF in presenting a comparison in distortion performance for 8 different circuits. Fig 1 is a standard 6SN7 amplifier stage, with bypassed cathode. It presents a stage gain of about 15. Fig 2 Uses 2 6SN7 sections in an SRPP configuration. The lower cathode is unbypassed. The system produces a gain of about 15. Fig 3 is similar to figure 1, except it uses a 6SL7. Stage gain is about 35. Fig 4 is similar to figure 2, except it uses a 6SL7. Stage gain is about 35. Fig 5 is a standard 6SL7 with unbypassed cathode, coupled to a 6SN7 acting as a cathode follower. It is unique in that the grid of the input stage is raised above ground, and proper bias is maintained by using a larger cathode resistor. This has the advantage of providing better operating point (DC) stability with variation in devices. If you do not bypass the cathode, it also allows you to set the stage gain independent of the operating current. A combined gain of 20 and 10 data are presented. Fig 6 is similar to figure 1, except it uses a 417A/5842, biased to 10 mA. Stage gain is about 25. Fig 7 is an SRPP configuration of 417A/5842 tubes. Like figure 5, the grid of the input stage is raised DC wise to allow the gain to be manipulated. This SRPP is set for a gain of about 25. As the SRPP configuration is an active load, the current can be increased, providing better linearity than the stage operating at 10 mA. The current in this stage is set to 21 mA. Fig 8 is an SRPP configuration with a twist. The twist is to use a large valued inductor in series with the "upper" cathode resistor. This effectively makes the AC current flow relatively constant in the lower tube, lowering distortion. It also raises the maximum available output swing relative to the supply voltage. In this example a 6SN7 (6CG7 performs essentially identically) is used, and a very small ammount of negative feedback is incorporated into the configuration. The advantage, as mentioned above is very low distortion and large swing. This can be seen in the data below, even though the power supply value (+175V) is much lower than the other circuits. The disadvantage is a slew limiting condition at high levels and high frequencies. This is separately shown after the data in the first table. This concept was originally described in Glass Audio by Ignacio Vila, who used a transistor to achieve results similar to the inductor shown here. In the following table, THD represents total harmonic distortion from a distortion meter. HD2, HD3, HD4 etc represent individual distortion products as measured on a wave analyzer. The input frequency for all tests was 1 kHz from a source with negligable distortion. Note that THD is not necessarily the linear sum of the individual distortion components. Output level is RMS volts, distortion measurements are percentage distortion. Figure Output THD HD2 HD3 HD4 HD5 HD6 HD7 Level ------ ------ ------ ------ ------ ------ ------ ------ ------ 1 1 0.08% 0.08% 2 0.18 0.17 0.02 5 0.35 0.46 0.032 0.01 10 0.68 0.86 0.04 0.02 0.03 15 1.0 1.3 0.052 0.03 0.05 20 1.5 1.72 0.08 0.04 0.06 25 1.7 2.2 0.12 0.04 0.07 30 2.1 2.7 0.20 0.03 0.08 40 2.8 3.6 0.34 0.06 0.10 50 3.6 4.6 0.50 0.08 0.15 60 5.0 6.0 0.78 0.12 0.11 0.04 ------ ------ ------ ------ ------ ------ ------ ------ ------ 2 1 0.03% -not measured in this setup- 2 0.16 5 0.28 10 0.55 20 1.2 30 1.8 40 2.4 50 3.0 60 3.9 ------ ------ ------ ------ ------ ------ ------ ------ ------ 3 20 0.50% 0.60% 0.064% 30 0.80 40 1.50 2.1 0.80 0.42 50 4.0 ------ ------ ------ ------ ------ ------ ------ ------ ------ 4 10 0.15% 20 0.40 30 0.60 40 0.88 50 1.40 1.7 0.64 0.042 ------ ------ ------ ------ ------ ------ ------ ------ ------ Gain=10 5 10 0.04% 0.032% 0.04% 20 0.06 0.04 0.06 30 0.06 0.052 0.08 40 0.07 0.06 0.12 0.01 50 0.07 0.068 0.22 0.03 0.01 55 0.12 0.08 0.30 0.05 0.04 60 0.5 0.14 0.64 0.08 0.26 Gain=20 5 10 0.07 0.07 0.034 20 0.1 0.14 0.06 30 0.14 0.2 0.1 40 0.15 0.25 0.16 0.015 50 0.2 0.28 0.3 0.04 0.01 60 0.37 0.2 0.54 0.17 0.1 0.01 0.05 ------ ------ ------ ------ ------ ------ ------ ------ ------ 6 10 0.42% 0.48% 0.028% 20 0.8 0.92 0.064 30 1.1 1.4 0.12 40 1.4 1.82 0.195 50 1.8 2.4 0.6 60 2.5 3.0 0.72 0.028 0.032 70 3.1 4.0 0.9 0.1 0.1 0.02 0.034 ------ ------ ------ ------ ------ ------ ------ ------ ------ 7 10 0.04 0.052 0.04 0.016 0.012 20 0.06 0.06 0.06 0.016 0.012 30 0.08 0.064 0.096 0.016 0.014 40 0.09 0.072 0.16 0.022 0.012 0.022 0.022 50 0.18 0.14 0.30 0.06 0.042 0.038 0.030 60 0.39 0.32 0.48 0.06 0.092 0.040 0.048 70 1.40 1.40 1.20 0.30 0.34 0.08 0.11 ------ ------ ------ ------ ------ ------ ------ ------ ------ 8 5 0.04 0.04 0.01 10 0.08 0.08 0.02 0.01 20 0.16 0.16 0.04 0.01 0.01 30 0.32 0.33 0.13 0.02 0.04 40 1.1 1.0 0.9 0.32 0.14 0.09 0.05 50 2.8 2.5 2.9 0.51 0.33 0.20 0.10 It should be noticed that sometimes, the cathode follower stage following a standard gain stage produces a composite distortion lower than the individual stage by itself. Also, depending on the devices, either a standard stage + cathode follower -or- the SRPP configuration works better. Notice the extremely low distortion even at very high output levels of the 6SL7+6SN7, and notice that the harmonic structure of the 417A/5842 SRPP stage is almost entirely second and third order, with second order predominating until clip point is reached. The distortion of the modified SRPP is impressive when you realize the supply voltage was about half the other circuits. The modified SRPP does not provide a very effective output when you consider load capacitance, as it is running at a "constant" current. In the following table, the distortion vs level vs frequency show this difference. For the passive inductor case, a "cheap" audio output transformer was used. The low frequency distortion can be improved using a larger value inductor. A "transistor gyrator" can be substituted. As this simulates a larger valued inductor, the LF distortion is better, but the capacitance of the transistor makes the high frequency performance slightly worse. Here's the data: Inductor Transistor Gyrator Vout 20Hz 1kHz 20kHz 20Hz 1kHz 20kHz 5 0.3% 0.04% 0.03% 0.05% 0.04% 0.03% 10 0.6 0.08 0.09 0.09 0.08 0.1 20 0.95 0.16 0.32 0.18 0.16 0.34 30 1.35 0.32 2.1 0.33 0.32 2.3 40 1.9 1.1 8.4 1.2 1.1 8.9 50 3.7 2.8 --- 3.7 3.7 ---