Power Amplifier


This audio power amplifier in principle is a "Current Feedback" design (opposite to most amplifiers that are Voltage Feedback designs), see figure 1 below. The input signal is fed to an input buffer with a voltage gain of one and output impedance R0.

Figure cfb The resistors RF and RI attenuate the output signal and make the feedback network. The difference between the input signal and the feedback signal results in the error current I. This is led to a current mirror and transformed to an equivalent voltage across the resistor RT, which normally has a high value. This voltage is fed to the output via an output buffer, which is a current amplifier (with a voltage gain of one). CT is a compensation capacitor, which provide for the stability of the amplifier (and the necessary phase margin). Amplifiers of the Current Feedback design are taken out to get amplifiers where the bandwidth is nearly independent of the voltage gain. Amplifiers of this type have a very large bandwidth and are very fast.


Schematic description

The amplifier, view the circuit diagram, is fully symmetric built up as a voltage amplifier followed by a current amplifier. The relation R14/R15 approximately gives the closed loop gain.

The input stage consists of two JFETs in a symmetrical coupling instead of the usual (minimum four) bipolar transistors. If one extract the difference between the drain current for these complementary field effect transistors, in theory a non-distortion amplifier can be realised. In practice the two transistors will be different, causing the difference current to superimpose even harmonic components. As it is used bipolar transistors in a symmetric coupling to make the difference current, these will add additional odd harmonic components. The final result is an amplifier reminding of a valve amplifier: The distortion consists of mostly second harmonic components followed by third harmonic etc, in other words symmetric falling harmonic components.

The offset adjustment is made up by the potentiometer P13. The resistors R1 + R2 set the input impedance. The resistor R1 will together with the preamplifiers output resistance and the input capacitance of the amplifier set the upper cut-off frequency. If the preamplifier has e.g. 600 ohms output impedance, the cut-off frequency is about 2 MHz, giving an equivalent input capacitance of about 45 pF.

The field effect transistors J11 (N-channel) and J12 (P-channel) withstand a maximum of respectively 40 V and 20 V to operate properly, and their internal capacitances are high. Using the cascode coupling as shown solves these problems. The common base transistors Q9/Q10 are bipolar types. It is also used 'base degeneration' by means of the resistors R3-6 instead of using a constant voltage on the bases of the transistors. The distortion from these transistors is very low compared to the relative high (even order) distortion from the field effect transistors.

Instead of using an ordinary current mirror for the current from Q9/Q10, here is used an amplifying current mirror. The voltage on the input of the amplifier is transformed to a proportional current in the JFETs. This current is compared to the feedback current (via R14). The resulting error current is transformed to a voltage across R7 (and R8). This error voltage is found across R19 (and R20) with little degradation, since Q16 (and Q17) operates with constant current (and with that constant base-emitter voltage). The error current is amplified as the relation given by R7/R19 (and R8/R20), and is transformed to a proportional voltage in the summing point before the current amplifier.

The open loop gain is given by the JFET's transconductance reduced with local feedback, multiplied with the relation R7/R19 and finally multiplied with the double load on the base of Q31/Q32. Without the resistors R23 and R24 the load at this point is both load dependent and parameter dependent. Thus these resistors determine the open loop gain to about 46 dB. With a closed loop gain at about 26 dB, the feedback factor is relatively low: about 20 dB.

Lead compensation is used instead of the normal lag compensation. This gives a high speed stable amplifier with large bandwidth, and the amplifier sound is not negative affected by the compensation capacitors C25 and C26. The low value of the compensation capacitors in turn means higher Slew Rate value. In this case the Slew Rate limitation do not set in before over 200 V/µs (with the input stage driven into saturation).

The amplifier has the same open loop gain and bandwidth for all audio frequencies, since the open loop bandwidth is very high, about 100 kHz. This is equivalent with nearly equal amount of distortion and the same output impedance over the whole audible range. The output impedance is nearly resistive througout the whole audio range, only barely inductive at the highest frequencies (a few degrees at 20 kHz).

The phase margin is about 75 degrees at the given closed loop gain. If higher phase margin is desired, the value of C25 and C26 may be doubled without any negative influence on the amplifier performance. The linearity is good, thanks to the use of an “inverted” Compound-coupling (Q16/Q21 and Q17/Q22). The distortion is less than it would be if an ordinary current mirror were used (without an additional buffer).

The current amplifier used is a usual Darlington emitter follower, chosen from stability and linearity requirements, and at the same time having good thermal stability. The bias generator consists of the components 27-30. The quiescent current is set by means of the potentiometer P29.

The unregulated power supply to the voltage amplifier is low pass filtered by means of R38 and R39 plus C40 and C41. Together with the decoupling capacitors C42 and C43 and the feedback, the low pass filtering ensures that ripple and noise on the supply voltage not reach the amplifier output.

The used power supply for the current amplifier may be common for the two channels, view the schematic (the transformer is not shown).

Since this is a class A amplifier with global feedback, a common supply is sufficient when the filtering capacitors are large enough. It is however possible to use separate supply for the current and voltage amplifier, if wanted. In this case, remove R38 and R39 and apply the voltage over C40 and C41. The power supply voltage for the voltage amplifier in this case may then be about 5 V higher than the power supply voltage for the current amplifier. Higher power output is thus achieved without higher power dissipation worth mentioning (for an equal value of the power supply voltage to the current amplifier).

Using the values in the parts list, the amplifier will perform 25 W RMS into 8 ohms. This demands a quiescent current of 1.25 A. The quiescent current may be increased above this value depending on the heat sinks used. Heatsinks with a minimum of 0.35K/W is recommended. A quiescent current of 1.6 A corresponds to class A working for full output voltage down to a load impedance of 6 ohms. With an increased voltage for the voltage amplifier (e.g. 5 V higher than for the current amplifier, i.e. 30 V), the output power is increased to 30 W RMS into 8 ohms. This requires a quiescent current of minimum 1.4 A. If the current is increased above this value, do remember the cooling demand.

Some measuring results

Output power:
Output impedance:
Frequency range:
Slew Rate:
THD at 10 V RMS:
Rise/Fall time:
Input impedance:

2x25 W RMS
< 35 mohm, 7 degrees (1Hz-20kHz)
DC-1.2 MHz
< ± 200 V/µs
0.02 % (200 Hz, 1 kHz and 5 kHz)
< 1 µs
0 dBu (0.775 V)
100 kohm//45 pF

The layout and component placement are shown as images. The printed circuit board measures 160x80 mm. Two boards are needed for a stereo version.

The parts list applies for both the circuit board and the power supply.

If you have the possibility to make your own boards, I have prepared a zipped file where you can find the necessary files. The file extensions are the following:

bot - Gerber plot file (copper tracks bottom)
ctr - Gerber plot file (board contour)
lis - Text file (Aperture list)
smb - Gerber plot file (soldermask bottom)
sst - Gerber plot file (silkscreen top)
tap - Exellon drill file

Mounting description

With the exception of the power resistors, 1/2 W metal film oxide resistors with 1 % tolerance may be used. The power resistors should be non-inductive, and alternatively they may be realized by connecting in parallel power metal film oxide resistors (e.g. 2 W). The JFETs 2SK170BL/2SJ74BL should be used for J11/J12. The couple 2SK147GR/2SJ72GR may be used instead. These have higher internal capacitances and transconductance and are slightly more linear. In return the price is higher. The value of R7 and R8 is chosen to match a current of 7 mA from the JFETs. If the current is higher, the resistor value should be reduced. A replacement for the couple 2SC1775/2SA872 used for Q9/Q10 is e.g. 2SC1815/2SA1015 or 2SC2240/2SA970. The transistors Q16/Q17 may also be replaced by 2SA970/2SC2240. 2SC3421/2SA1358 is used for (Q30)Q31/Q32. These may be replaced by 2SC4793/2SA1837. For all replacements, be sure that the pinning is correct when mounting the transistors on the circuit board.

The transistors Q21 and Q22 are running at a current of nominally 27 mA and should be mounted with a small heat sink. The transistors Q16 and Q17 are running at a current of nominally 4 mA, and no heatsinking is normally required. The output transistors used for Q34/Q35, are the well-known couple 2SC2922/2SA1216 from Sanken. The author has not been watching for any replacement for these. They are relatively linear, fast and at the same time very rugged, a seldom combination. These transistors have been on the market for quite a long time now and are relatively often used in commercial amplifiers, and they are relatively cheap. They come in a rare plastic housing, something that makes it necessary to mount them directly to the heat sink. This is also done for the drivers (Q31/Q32) and the bias transistor (Q30). A better solution would be to have separate cooling for the drivers, but the quiescent point is not drifting much when the temperature has stabilised. Besides additional heat sinks are spared in this way.

The two parallel outputs from the transformer are carried to the rectifiers as shown in the power supply schematics (it should be fuses here in case the rectifiers break down). From the rectifiers, make a connection to a common ground on the chassis. The phono socket ground terminal is connected to chassis (near the input) and the (ground) shield of the phono cable is connected to the circuit board, at the point marked SG (Signal Ground). The inner conductor of the phono cable is connected to the circuit board point marked with 'IN'. From the loudspeaker output the two conductors are twisted and fastened to the circuit board in the two points marked with 'OUT' and 'GND'. The last one is connected to the minus conductor. The loudspeaker minus output is connected to chassis. All connections should be as short as possible. If some sort of instability or noise should occur, the probability is high that the reason is bad wiring (e.g. earth loops).

Start-up and adjustment

It is recommended to use a variable transformer or variable DC voltage generator first time the amplifier is started up. When the power supply voltage is increased, adjust the output-offset voltage by means of the potentiometer P13 to be close to 0 V DC. Also adjust the quiescent current to initially be at a minimum, and increase this slowly by means of the potentiometer P29. If possible, look at the output with an oscilloscope, there should not be anything but noise here if everything is OK. When the temperature is increasing, it is necessary to re-adjust both offset voltage and quiescent current (min. 1.25 A). The offset voltage at the output varies, but should not exceed 30 mV.

About 0.775 V RMS input voltage is required for full output power. This should be sufficient for the most modern signal sources without being forced to use a preamplifier. If higher gain is wanted, R15 is reduced (and vice versa). Please note that the feedback resistor R14 should be unchanged. None of the good properties of the amplifier, like bandwidth, distortion and Slew Rate, are deteriorated by moderate change of R15.

The output power of this amplifier may be increased to 50 W RMS in class A. The cooling requirement, however, is large and should not be underestimated. It is the author’s belief that 25 W RMS real class A is sufficient for domestic use in most cases. A class A amplifier generally is perceived more powerful than a class B (or A/B).

Please notice:
This project description is for non-commercial use, only. Using this document on a site and charging a fee for download is vialation of non-commercial use and prone to demand for payment. So, for commercial use, contact me for agreement of terms. This page, however, can be downloaded for own use, and linked to, not violating term of non-commercial use.



Knut Harald Nygaard