Homemade 25Wx2 EL34 tube amplifier circuit

Due to the limitation of the characteristics of the electronic tube device itself, the output power of amateur tube amplifiers is generally not too large, especially when using a single-ended output amplifier, the output power is even smaller, usually only a few watts, generally not more than 10W. For a 10W power amplifier, it can often only provide an average working power of 1 to 2W. Because when performing high-fidelity playback, due to the large dynamics of the program source, the peak power when the average power is 1 to 2W has exceeded 10W, thereby increasing the playback distortion, especially when using speakers with lower sensitivity, the distortion is more serious. The International Electrotechnical Commission (IEC) has made provisions in the minimum performance requirements for high-fidelity home audio power amplifiers, requiring the rated output power of the amplifier to be ≥10W.

It can be seen that if you want to better enjoy various music programs including large-scale symphonies, a 10W×2 amplifier is the minimum requirement. Generally speaking, a rated power of 20~30W×2 can achieve the basic goal of high-fidelity playback. This article introduces the production of a 25W×2 amplifier in detail, which can be imitated by tube amplifier enthusiasts who need a larger output power.

1. Basic structure

The output power of the tube amplifier mainly depends on the output stage tube model and the circuit program used. In terms of the circuit program, if you want to double its output power, there are usually two ways. For single-ended Class A amplification, two output power tubes can be used in parallel, as shown in Figure 1 (a). Compared with single-tube operation, under the premise that the working voltage remains unchanged, the screen current increases by 1 times, the output power also increases by 1 times, and the input voltage and distortion remain unchanged. In addition, due to the parallel connection of two tubes, the internal resistance is reduced by half, and the load resistance is also reduced by half. Therefore, as long as the current capacity of the power supply is increased and the primary side impedance of the output transformer is reduced by half, it is easy to change the single-tube Class A output to a dual-tube parallel Class A operation mode. In fact, the primary side impedance of the output transformer does not have to be strictly reduced by half, and much greater output power can be obtained than when a single tube is working. For example, when 300B is working as a single-ended Class A, the primary side impedance of the output transformer is 3~3.5kΩ. When the double-tube parallel connection is changed to Class A operation, the primary impedance of the output transformer can be 1.5-1.6kΩ, and the output power is roughly twice that of a single tube. However, if you cannot find a ready-made output transformer with an impedance of 1.5-1.6kΩ, you can also use an output transformer with a primary impedance of 2-3kΩ as a substitute. When designing a 350V circuit, the actual voltage between the screen and the ground is about 350+30=380(V).

According to the measured data in Figure 3, the cathode voltage (to ground) of the output stage is 29V, so the static current of the two tubes is 29V/300Ω=96mA, the current of each tube is 48mA, and the actual screen loss of EL34 is 350V×48mA=16.8W. This is only 67% of the maximum screen loss of EL34 (25w). Since there is a large margin, it is very beneficial to extend the working life of the tube.

The inverter stage (V2) uses a cathode-coupled inverter circuit. As shown in Figure 3, both tubes use fixed resistors, but the resistance values ​​are slightly different. R7 is 30kΩ, while R8 is 33kΩ, so the balance adjustment can be omitted. If the maker has a signal generator and an oscilloscope, the adjustment can be made as follows. First, change R8 to 30kΩ, and then connect a variable resistor (5kΩ) in series to the screen circuit of V2 (the lower tube); then input a 1kHz sine signal, and use an oscilloscope to observe the waveforms of the upper and lower output signals, and at the same time fine-tune the variable resistor to make the amplitudes of the two waveforms equal; finally, use a fixed resistor with a similar resistance value. Of course, the output power at this time is reduced, not reaching twice that of a single tube, but it is still significantly higher than the output of a single tube. Similarly, the output impedance is also lower.

Another way to increase the output power is to use push-pull output, as shown in Figure 1(b). Since the working state of push-pull output can be set to Class A, Class AB, and Class B, the output power can reach 1 to 2 times that of single-ended output. For Class A push-pull with less distortion, the output power is generally 1 times greater than that of single-ended output.

In fact, when working in push-pull mode, for AC signals, the two electron tubes work in series, so the primary side impedance of the output transformer is 1 times that of single-ended output, and the output power is also roughly increased by 1 times.

Compared with the single-tube single-ended output, the output power of the dual-tube single-ended output and push-pull output is about twice as large. However, the push-pull output is not only efficient, but also because the DC directions flowing through the primary side of the output transformer of the two tubes are opposite, the DC magnetic flux cancels each other, and the iron core of the transformer does not need to leave an air gap. Under the same conditions. The inductance is much higher than that of a single tube, which is conducive to simplifying the structure. In addition, the push-pull output transformer has the function of offsetting the power supply ripple, thus reducing the requirements for power supply filtering. Finally, the push-pull circuit can also offset even harmonics, and the distortion is significantly lower than the single-ended output. The disadvantage of the push-pull circuit is that the tubes need to be paired, and two driving signals with the same amplitude and opposite phases are required, that is, an inverter circuit is required. In addition, the two primary windings of the output transformer should be symmetrical and balanced as much as possible in terms of the number of turns, DC resistance, leakage inductance and distributed capacitance. The better this is done, the more obvious the advantages of the push-pull circuit will be.

There are three common inverting circuits required for push-pull output, as shown in Figure 2. Figure 2 (a) is a transformer inverting circuit, which does not require the use of active devices, has a simple circuit, a good inverting signal waveform, low distortion, low output impedance, is not easily affected by the output stage grid current, and has a strong driving capability, but this requires an input transformer with excellent performance. Figure 2 (b) is a screen-cathode split load inverting circuit. Since the signals taken from the screen load resistor and the cathode resistor are exactly inverted, as long as the screen resistor is equal to the cathode resistor and 100% of the current negative feedback is added through the cathode resistor, an inverted driving signal with equal amplitude can be obtained from the screen and cathode. This circuit has low distortion and a gain of ≤1, so it requires a sufficiently high gain in the previous stage, or a push-pull voltage amplifier stage is added after it, otherwise it is difficult to obtain a sufficiently large power output.

Neither of the above two phase inversion circuits needs to be adjusted, and the gain is basically 1. Figure 2(b) is the most common in amateur production in my country. Figure 2(c) is a cathode coupled phase inversion, which is characterized by obtaining a higher gain while inverting the input signal. However, the amplitude of the phase inversion signal output by this circuit is not strictly equal, that is, it needs to be balanced and adjusted, which is an inconvenience for amateur production. Therefore, this phase inversion circuit is more common in commercial machines produced by manufacturers. In the past, this type of phase inversion circuit was not widely used in amateur production in my country, so we should have a general understanding of its working principle, please refer to Figure 2(c).

The input signal is first amplified by a voltage amplifier (v1), and then enters Vup and Vdown for inversion and amplification. The output signal of V1 directly enters the gate of Vup and takes out the amplified signal from its screen as a driving signal for the subsequent push-pull amplifier stage. Obviously, this signal is inverted with its gate input signal. In other words, the working conditions on Vup are exactly the same as those of a general voltage amplifier stage.

The other push-pull drive signal is obtained from the lower screen of V. However, the working state of V is different from that of V. Its cathode is connected to the cathode of V. That is, the two share a cathode resistor. Its gate is connected to the gate of V through a high-value resistor R (1MΩ), and the gate of V is grounded through a large-capacity capacitor C (0.47μF). In other words, the gate of V is grounded for AC, that is, V works in the state of gate grounding, and the input signal is taken from the cathode resistor of V and injected by the cathode. In this way, V works in the cathode grounding amplification state, the screen output signal is in phase with the gate input signal, and the signal on the cathode resistor is in phase with the gate input signal, that is, in phase with the screen output signal. V works in the state of gate grounding amplification, and its screen output signal is in phase with the cathode input signal, that is, in phase with the screen output signal of V., so that the screens of V. and V can take out inverted signals with opposite phases. As for the amplitude between the inverted signals, because the amplification amount in the two working states is different, when the two tubes take the same screen resistor, the output signal amplitude is not the same. Usually the amplification is greater when the cathode is grounded. Therefore, as long as the screen resistance under V is appropriately increased, that is, an adjustable resistor VR is connected in series with R2 in Figure 2(c) for fine-tuning, an inverted signal with the same amplitude can be obtained.

2. Practical Circuit

Figure 3 is the electrical schematic diagram of a 25W×2 power amplifier designed based on the above circuit structure, in which the power supply is shared by the left and right channels. As can be seen from the figure, the entire amplifier consists of three parts: input voltage amplification (V1), inversion (V2) and output (V3, V4).

The output stage uses EL34, which has stable performance and good sound quality. It is also easy to buy in the market, without the worry of out-of-stock. The attached table shows the main application characteristics of EL34. As can be seen from the table, when EL34 is used for single-ended Class A amplification, the maximum output power is 11W (distortion rate 10%) under the screen load impedance of 2kΩ. When it is used for push-pull amplification, the maximum output power can reach 36W (distortion rate 5%) under the screen load impedance of 3.8kΩ.

Since the maximum output power requirement of this machine is about 22 to 25W, the working parameters in the AB class state in the table can basically be selected, that is, the screen voltage is 350V, and the screen current is slightly larger (about 48mA), which is conducive to reducing distortion. It should be pointed out here that the screen or screen grid voltage in the table is relative to the cathode potential. Since this output stage adopts a cathode self-bias circuit, that is, the voltage drop (about 30V) on the cathode common resistor R13 is used as the bias voltage of the output stage, the screen voltage in the table can be used as a resistance to replace the variable resistor.

Since the inverter stage of this machine has a large gain and the output power of this machine is not too large, the gain requirement for the input voltage amplifier stage is not high, and a single-stage triode voltage amplification is sufficient. Originally, it is also possible to use a dual triode to share the voltage amplification of the left and right channels, but in order to isolate the left and right channels, a dual triode is connected in parallel to perform voltage amplification of one channel.

Power amplifiers that use multi-pole tubes as output stages usually apply a certain amount of overall negative feedback to reduce output impedance and increase damping factor (DF) due to their high internal resistance. Generally speaking, the larger the negative feedback amount, the lower the damping factor, and the stronger the power amplifier's ability to brake the speaker. However, for a specific speaker, a damping factor that is too high or too low is not the best. For this reason, the negative feedback amount of this machine can be changed through the band switch. The negative feedback amount can be selected as 8dB, 10dB and 12dB in three levels, which are determined by the user according to actual usage.

Since the push-pull amplifier has relatively low requirements for power supply, the power supply part of this machine is also quite simple. In order to improve the rectification efficiency, a crystal diode is used instead of the traditional electron tube rectification.

3. Production and debugging

After purchasing the power supply and output transformer and other major components, the chassis can be designed and processed. The dimensions in Figure 4 are for reference. It should be noted that this dimension only gives the dimensions of the upper and rear side panels of the chassis. The side panels and the front side panels are not drawn. The reader can decide the processing dimensions by himself. The front and rear of the center of the upper chassis are the power filter choke and the power transformer, and the left and right of the latter are the output transformer. The left and right sides of the rear side panel are the input socket and the speaker terminal; the middle is the power socket, and the two sides are the fuse holder and the band switch that controls the amount of negative feedback. The power switch is installed in the front center of the chassis, and the left and right channels are volume controls.

The output transformer is U-405 (ISO) from Hirata Electric Co., Ltd., Japan. If it is not available, it can be commissioned to be processed according to the following requirements: rated output power 60W, primary side (screen to screen) impedance 5kΩ. Primary side impedance 4Ω, 8Ω and 16Ω. Primary side maximum DC current (balanced) 270mA, unbalanced current 7mA. Primary side maximum inductance 380H, minimum 160H. Frequency characteristics 4Hz~80KHz. Attached screen grid tap.

Figure 5 shows the layout of the main components of the chassis and the schematic diagram of the physical wiring. There is a "grounding point" on the left side of the power transformer. The chassis here must be polished and cleaned, and it must be securely fixed with screws using a toothed welding piece to ensure that it is electrically grounded. All wiring that passes through the chassis (output transformer leads, B power line, screen grid leads, etc.) must be protected with a sleeve. The input leads must use shielded wires (see Figure 5), and other leads can use ordinary plastic wires.

Figure 6 is a schematic diagram of the AC power supply and filament wiring. Figure 7 is a schematic diagram of the negative feedback control switch wiring, and the six negative feedback resistors are directly installed on the band switch.

It should be added that the cathode resistors and bypass capacitors (300Ω/20W enamel resistor and 100μF/100V) of the left and right channel output stages in Figure 5 are shown to be installed close to EL34, but in fact they are installed close to the output terminals, especially the 300Ω/100μF of the R channel is installed between 100μF×2 and 100μF/500V in front of the output terminals.

After the installation is completed and checked, you can insert the electron tube for power-on test, and pay attention to whether there is any abnormal situation such as smoke or odor. If there is no abnormal situation, you can use a multimeter to measure the voltage between the cathode of EL34 and the ground, which should be 28-30V, indicating that the output stage is working normally, and then measure the voltage of each pole of 12AU7 and 6CG7, which should be close to the voltage value shown in Figure 3. This shows that the circuit is installed correctly and works basically normally.

Next, disconnect the negative feedback wiring, connect the speakers and CD player to listen. Since negative feedback has not been added, the sound may not be ideal. After shutting down, restore the negative feedback wiring, and then power on to listen. If self-excitation occurs as soon as the power is turned on, it means that the negative feedback is mistakenly connected to positive feedback. To do this, disconnect the power supply and swap the wiring of the primary side of the output transformer Pl, SG1 and P2, SG2. Power on again to listen and change the negative feedback amount. You can hear that the sound quality will change. At this point, the installation is complete.

4. Performance indicators

First, the residual noise of this unit is 0.3mV when the negative feedback is 12dB. It should be said that the data is not very good. However, even when playing with large speakers, there is no hum.

Figure 8 shows the input characteristics of this machine. The clipping power at different negative feedback amounts is basically 22W. When the negative feedback amount NFB=8dB, the maximum output voltage measured when clipping on the 8Ω terminal on the secondary side of the output transformer is 14V (1kHz), which is equivalent to U2/RL=14V×14V/8Ω=24.5W in output power. At this time, the input sensitivity is about 0.74V. It can be said that the expected design requirements have been met.

Figure 9 shows the frequency response, which is 30Hz~15kHz (-1dB) without feedback, and expands to 10Hz~70kHz (-1dB) when 10dB negative feedback is applied, and only drops 2dB at 100kHz. Such a good frequency response is mainly due to the wide frequency response of the output transformer.

Figure 10 shows the total harmonic distortion (THD) characteristics of this machine when the negative feedback amount is 8dB and 12dB respectively. When NFB=12dB, the output power is only 0.02% at 1W (1kHz), 0.3% at 5W, and 2% at 22W. The output impedance of this machine is 1.44Q (1kHz), and the damping coefficient when the load is 8Ω is 8Ω/1.44Ω=5.5, which is obviously lower than the damping coefficient of general single-ended output.

Figure 11 shows the square wave response of 200Hz, 1kHz and 10kHz (the upper part is the input square wave, the lower part is the output square wave) and the waveform of 22.5W sine wave (1kHz, NFB=12dB), all of which are very good. This unit is tested together with the JBL4312Mll speaker. The sound is transparent, wide and very attractive. You will not feel tired even if you listen for a long time. It is very durable.