The headphone amplifier design goals are:Very beautiful sound, low power consumption, small size (30x40 cm), the power section can have a variety of options without using poorly shaped power adjustment tubes, such as 6080 and 6C33-B. Low output impedance (about 33 ohms) The output stage bias current is sufficient for music Peak requirements (300 ohms up to 20Vpp) This circuit is optimized by SPICE design software and uses the tube model designed by Mithat F. Konar. This model is very close to the real situation, and an interactive program has been developed to determine the correct parameters. This is the result of E182CC (red is the simulated curve, white is the real curve):
Figure 1 E182CC SPICE simulation diagram
This is 5814 (Military model of ECC82):
Figure 2 5814 SPICE simulation diagram
The curve simulated by SPICE is very close to the real curve and is almost indistinguishable. The real curve is made with AudiomaTIca's Sofia vacuum tube curve tracer. SPICE circuit simulation software is free. There are Unix (Berkeley) platform and Windows (WinSpice) platform. There are also many commercial software under Windows 95 / NT. [Original Editor: A beta version of the graphical SPICE simulation software under the Windows platform can be downloaded from Interface Technologies / MicroSim. ]
In the circuit of the headphone amplifier (Figure 3a, Figure 3b and Figure 3c), the gain stage uses a single-ended Class A triode, and the output stage uses a double triode in parallel as a cathode follower. The first version uses E82CC / 5814. It sounds pretty, but it is distorted because of poor line shape. I use JAN Philips 5814, maybe other tubes are better, like Mullard or Jan Philips E82CC / ECC82. The bias current of each 5814 is adjusted to 4.42mA, and the power dissipation is 3.76W. In order to reduce distortion, no bypass capacitor is added to the cathode resistance to form local negative feedback. You can also try to add a 220uF, 16V ELNA electrolysis.
The second version uses (mine is RCA 3A5) direct heating transistor, the distortion is less, the smooth sound can only be achieved by direct heating transistor. 3A5 must use DC power supply, as shown in Figure 7. The third edition uses Jan Philips 6SN7 (my favorite now). Although the sound is not as smooth as the 3A5, the distortion is very low, and the low frequency response is eye-opening! Sometimes, when I listen to 6SN7, I wonder if high frequencies are lost, but this is not true.
Note: The 3A5 is very sensitive and has a microphonic effect (if you touch it, the headset will make a 'gong' sound). I mounted the 3A5 on a floating tube base.
In the above three versions, the bias current of the output tube V2 is 26mA each (total 52mA), and the total power consumption is 6W (2x3W). Only refers to the use of Mullard or JAN Philips tubes.
All resistors use Allen Bradley carbon film for better sound. Seven 2W22K resistors can be replaced with a 3K20W TO220 packaged Caddock non-inductive resistor. Figure 3d shows how to install seven 22K resistors.
The oil-impregnated paper dielectric capacitors of Jensen or Audio Note are used for the cross-connect and output capacitors. Of course, MKP capacitors like Solen can also be used. For electrolysis, I like Black Gate or ELNA Cerafine (manufactured by Audio Note), ROE, PRAGUE, MALLORY are also good choices.
In recent years, there have been two trends in the use of capacitors in Hi-end amplifiers: one is the use of MKP polypropylene capacitors (on the output stage and cross-connect), the other is electrolytic capacitors (on the power supply section) and oil-impregnated paper dielectric capacitors ( Used as a cross-connect). MKP capacitors are tested in a dynamic environment, and their low distortion, low internal resistance and fast speed are well received. Electrolytic capacitors generally have large distortion and low speed. However, there are also excellent audio-specific varieties such as Black Gate and ELNA Cerafine.
This time, I used ROE electrolysis and SOLEN MKP capacitors. Actual listening test, MKP capacitor makes the sound smooth. You can choose according to your hobby! In this paper, the power supply filter capacitor and the output capacitor of the amplifier can be replaced by MUF capacitors of 220UF or above.
About the power supply, the filaments can be powered by AC or DC except that the 3A5 (version 2 of the amplifier) â€‹â€‹requires a DC filament supply. Each supply has a TImed relay that mutes the audio output unTIl the tubes stabilize. I have designed 3 different power supplies-regulated, passive and cheap passive:
As far as the power supply is concerned, the filament can be powered by AC or DC, except for the 3A5 in version 2, which must be powered by DC. Each power supply has a delay relay, which mute the output before the tube enters a stable working state. I designed three kinds of power supplies-regulated power supply, passive power supply and cheap passive power supply:
There is no noise when the filament of this machine is not powered by DC. However, in order to make the filament voltage more stable, I suggest to use the slow start circuit of Figure 7 instead of the circuit of Figure 4 and Figure 5. Also, version 2 of 3A5 requires DC filament power supply, and this circuit is also most suitable. Slowly starting the power supply can extend the life of the tube! MJ15004 needs to install a heat sink, and its power consumption is 12W.
Here is a full-size printed circuit board diagram of the slow start power supply. The following is the installation diagram and recommended chassis layout structure.
Part of the indicators of the amplifier (using a regulated power supply):Gain 21.5db output resistance 33 ohm THD = 0.88% output voltage = 9.5v load 300 ohm => 200 mW 0.40% output voltage = 5.9v load 300 ohm => 80 mW "Max power of Sennheiser 580"
0.17% output voltage = 2.9v load 300 ohm => 15 mw
0.07% output voltage = 1.2v load 300 ohm => 2.5mw frequency response 0.7Hz to 1GHz
Figure 9 shows the attenuation of THD (distortion) at 9.5V, 2.9V, and 1.2V. so perfect! They show linear harmonic attenuation (the lower the power, the lower the distortion).Note
9/17/98: Improve Figure 4 regulated power supply: replace single MOSFET with MOSFET / BJT Darlington tube to better control the inrush current of large-capacity electrolysis. Improve Figure 4 and Figure 5: The filament is grounded, disconnecting the "E" point in Figure 3. Increase the "cheap" version of the power supply (Figure 6).
9/22/98: Add the second version of the amplifier (Figure 3b) and modify the power supply circuit (Figures 4, 5, 6 and 7).
9/24/98: Add Rin to the amplifier circuit (Figure 3a and Figure 3b), and remove the 6VDC output from the slow-start power supply (Figure 7). Correct the resistance of R7, and correct the spelling error DCC90 of V1 in the circuit diagram of version two.
9/29/98: Improved version 2 circuit (Figure 3b)-Add R8, R9. Improve the slow start filament power supply (Figure 7). Install the schematic diagram of the 7x22K resistor (Figure 3c), add the schematic diagram of the chassis. Add a section on capacitor selection.
9/30/98: Increase the printed circuit board map.
10/15/98: Add 6SN7 headphone amplifier version. Remind BUX48 and MJ15004 to install radiator.
Download Audiophile APP
Follow the audiophile class