60W Ultra-Simple Audio Power Amplifier
Fig.1 Complete and compact 4-channel version built into an aluminum casted box.
This audio power amplifier is simple and robust. The used components give a good compromise between performance and cost. The overall performance is suitable for serious home use. All used components are simple garden variety types that can probably be found in your component junkyard. The amp is dimensioned for 8-ohm loads, in which it can feed 40W continuous power with a distortion level of less than 0.05%. 6-ohm loads are also OK. This amp is not suitable for loads less than 6 ohms. Original instructions are in Finnish. I have translated here the most important parts.
This design has many similarities with the old and popular Texan amplifier from 1972. The output stage consists of a clever Darlington stage with voltage gain. It uses complementary, feedback or "compound" - or "Sziklai" pair - transistors. The output power transistors' Q6 and Q7 Vbe voltage (and its variation due to temperature changes) have a minuscule effect on the idling current. The control of the output stage idle current is fairly simple since temperature feedback from the power transistor heatsink is not needed. It is enough that Q1 and Q4 (could have been also Q1 and Q2) are thermally connected to each other.
Electronic short term overload protection is also added, and the maximum output current is reduced by limiting the maximum base current to the power transistors.
The Texan design had a fuse at the output. This caused additional distortion at lower frequencies. Electronic protection circuits also add some distortion but have some advantages over fuses. The overload protection in Texan was not fast enough to protect the output transistors during brief overloads. Let's look at the schematic (TULO means input):
Fig.2 Complete schematic of one channel.
The output stage voltage gain is determined by the ratio of R17 to R13 (and R16 to R12). The overall resistance at the driver transistor (Q2, Q4) emitters must be fairly low since transistors Q2 and Q4 supplies the base current to the output transistors (Q6, Q7). Excess resistance would increase distortion, too small resistance would increase the power dissipation in the gain setting resistors. The shown values are a good compromise. R14 and R15 limit the maximum base current to the output transistors during clipping, and also keeps the current and power dissipation in Q2/Q4 at a safe level when the overcurrent protection transistors (Q5, Q3) goes into action.
Since the output stage has a voltage gain of about 11, the op-amp output only needs to swing a few volts for a full output swing of ±35V on the output. The design can thus tolerate a relatively slow op-amp. The current gain in the output stage is also relatively high (about 2000) due to the compound Darlington transistors. The op-amp "sees" a relatively easy load to drive.
The output RC (Zobel) network (C7, R20) is essential for stability, guaranteeing proper loading of the output at high frequencies. The traditional output choke is not required since this design is stable with moderate capacitive loads. However, do NOT drive loads that are highly capacitive - like electrostatic speakers with this amp - this will certainly cause oscillations and fast destruction. Also, the power dissipation would drastically increase when the current is leading in phase the output voltage.
Biasing and quiescent current
The idle current in the power transistors is determined by the dc-voltage between the Q2 and Q4 bases. This voltage controls how much current is fed at idle to the power transistor bases. This bias voltage is controlled by Q1 that is mounted in thermal contact with Q4. The voltage drop across Q1 is set by R9. R5 and R6 generate a 5mA current to properly bias Q1 to work as a shunt regulator. The opamp receives also well-regulated and stable voltage from the two Zener-diode shunt regulators. Since the voltage to the bias circuitry is taken from these same regulated and smoothed op-amp power supply lines, the output stage is inherently very quiet and totally hum-free. Due to the symmetry and lack of bigger (electrolytic) capacitances, the dc-operation point is immediately correct when the power supply lines are still rising. Therefore there is no output thump (voltage spike) during turn-on, and there is no direct need for a speaker disconnecting relay.
The feedback loop is closed by the (in the year 1978) high quality bipolar low noise externally compensated uA748 op-amp. The op-amp's task is to fix the overall DC-balance - and close the feedback loop by suitable AC-feedback. The open-loop linearity of the output stage is not very good (especially with reactive loads), neither is the open-loop output impedance something to write home about in this current driven design. The performance relies very much on the op-amp and available loop gain. TL071 can be used as a replacement. The ±18V supply lines are too high for TL071 so the Zeners must in that case be replaced with 15V types (1N4744A).
R4 isolated the op-amp output from parasitic capacitances. The input and feedback chains are more or less traditional stuff. For best performance, it is good to use tantalum capacitors in locations C1 and C2. However, they can be substituted by similar valued and rated electrolytic caps. Tantalum capacitors work well with zero or even a small negative voltage. The correct (prudent) capacitor polarity is determined by the op-amps input leakage current direction if the op-amp is a bipolar input stage type, as the original part is.
Start with the assembly of the resistors, then capacitors, and last the semiconductors. Fig. 7 shows the component placement. Be very careful with the polarity of the polarized components. Q1 and Q4 should be mounted back to back and in contact with each other. A small drop of thermal compound is recommended, alternatively, a drop of Cyanoacrylatic glue works also well. Glue the transistors back together before soldering them.
Remember to assemble the one and only jumper wire on this single-sided PCB, if you use this layout as a basis for your design. The jumper is close to where the power supply ground wire enters the PCB. Leaving out the jumper will kill the transistors. Make a final check for shorts or bad soldering and proper component placement.
The output power transistors can be assembled directly to the PCB if a suitable and at least 6mm thick aluminum profile is mounted between them and the PCB. This aluminum profile needs to be mechanically and thermally connected to a bigger heatsink.
Either TO-3 (metal) or TO-247 (epoxy) encapsulated transistors can be used with this PCB design. The power transistors collectors must be isolated from the heatsink by mica washers or low thermal coefficient insulators (SilPad). See Fig3333 as an example. Don't forget to use a thermal compound.
Alternatively, you can assembly the transistors on a separate heatsink and using flexible wires (AWG16 or thicker). The wires should not be longer than 30cm (one foot). The insulation requirement is the same.
Independent of the mounting, make sure that the power transistor collectors are in contact with the PCB's speaker output copper traces. Insulating washers on the mounting screws cannot be used with TO-3 metal cased transistors, the mounting screw needs to contact both the transistor case - AND the PCB. But not the heatsink. It sounds complicated but is not. Just be careful.
The schematic shows that the collectors of the transistors are connected together. Mounting the transistors directly to the heat sink - without insulators - is a tempting idea because they are allowed to be in electrical contact. This would technically be possible - assuming you then electrically float the heatsink from the rest of the system. However, since this amplifier is non-inverting (as amps usually are) the risk of inducing capacitively positive feedback into the input is overwhelming. A floating heatsink could not use the chassis of the amp for additional cooling since it must be isolated. I strongly advise always isolating the power transistors from the heatsink, and then properly grounding the heatsink to the power supply ground.
Last but not least, make sure by using an ohm-meter that the isolation and the power transistor attachment are properly done.
Power supply and wiring
The amplifier needs a power supply of a maximum ±35VDC. Fig.5 is an example of a simple and working PSU. An unregulated voltage is OK as long as there is enough filtering capacitance in the supply lines. At least 4700uF for a stereo amplifier. 100VA transformer power is recommended per channel. Fuses in the power lines after the filter capacitors and before the amplifier are required for safety. You can use a rectifying bride instead of the four MR501 diodes. At least a bridge with 100V and 6A rating is needed.
The wiring is shown in Fig 6. Use flexible and shortest possible wires. Keep the input shielded coax away from the speaker wires. The heatsink and metallic enclosure must be grounded to the power supply ground.
Q1, Q2, Q3
MJ2955 or 2N2955
TO-3 / TO-247
MJ3055 or 2N3055
TO-3 / TO-247
1/4W or 1/2W
1/4W or 1/2W
R4, R12, R13
1/4W or 1/2W
1/4W or 1/2W
R14, R15, R20
4W (low ind.)
C5, C6, C7
Table.1 Parts list for one channel.
Fig.3 The original construction article, as published in 1978 in a Finnish electronics magazine. The article targeted absolute beginners in the art of electronics. Click the upper right corner to open.
Fig.4 Single-sided PCB layout seen from below (solder side). Note that the PCB accepts both TO-3 and TO-247 encapsulated power transistors. This picture in not in scale and is shown here for reference only, if you plan to design your own PCB.
Fig.5 Suggestion for the power supply. This example is good for one channel. For stereo, you need to double the transformer power rating,, double the filer capacitors capacitance (of double their count) and use separate fuse pairs for each channel.
Fig.6 Wiring of the power transistors (if located remotedly), speaker output, power supply connections, and the input connection.
10Hz...30KHz @ 40w
Signal to Noise Ratio
0.09% @ 100mW
0.06% @ 10W
0.05% @ 40W
0.12% @ 60W
1kHz into 6 ohm resistive load.
Table.2 Some basic but real-world measured parameters.
I made some crude measurements on the amplifier in 1978, they are gathered in Table 2. If you plan to design your own PCB or build this amplifier from scratch, I strongly recommend the following modifications:
Fig.7 Component placement seen from the component side. Apologizes for the poor image quality.
Fig.8 Stereo amplifier with the transistors mounted straight to the PCB using an aluminum bracket heatsink. Never cross the input signals (must be shielded cable) with the speaker output wiring. Keep these a separate as possible.
Fig.2 Stereo amplifier with the transistors mounted straight to the PCB using an aluminum bracket heatsink. Here the PCB's are slightly shortened to make them fit.
Copyright @ DEXTREL / 2019 / All Rights Reserved