Self Oscillating Amplifier for Distortion Testing

Distortion testing can be a painful exercise, since it is very hard to determine how much of the measured distortion is from the amplifier, and how much is from the oscillator.  One of my readers, Alfred Schaub, sent me some info on how it can be done using only the amplifier, with a handful of components to make it oscillate at a predetermined frequency.

Note: Alfred's copyright is limited to his original circuit and my adaptation of it.  All text and other diagrams are copyright (c) Rod Elliott.

This allows the home constructor to make quite accurate measurements, without having to spend a lot of money on a low distortion oscillator.

The circuit is based on a simple inductor-capacitor filter circuit, and needs only a pot and a small light bulb to set and stabilise the oscillation.  The frequency is fixed, and with a good inductor should be capable of very low distortion.  This circuit is a slight modification of that submitted by Alfred Schaub - I added a 100 ohm fixed resistor and used a 1k pot (rather than a single 100 ohm pot), and I also added a 1k resistor to improve the circuit's Q and reduce externally generated distortion to the lowest possible.  The amplifier must have a minimum gain of about 8dB for the circuit to work, but since typical amp gain is in the order of 30dB this will not cause a problem.

The circuit is shown in Figure 1, and as you can see there is not much to it.  The inductor should be a speaker crossover type, and preferably air-cored to ensure that there is minimal distortion.  Because the power level is so low, you may find that a ferrite cored inductor is quite usable, but be aware that it will introduce some distortion due to core non-linearities.  This can only be determined by experiment.  Some care is needed to prevent mains hum pickup from the inductor as this will affect the reading quite badly, but this will not be too much of a problem if the components are all mounted in a steel box for magnetic and electrostatic shielding.  Make sure you position the unit well away from power transformers or other mains powered devices when measuring the distortion of an amplifier.

Figure 1 - Amplifier Oscillator Circuit

The lamps need to be the smallest (in current) you can find.  The suggested units should be quite easy to find, or (if you can find one) a single 24V lamp can be used instead.  This was Alfred's original idea, but 24V lamps are uncommon.  The lamps will create a small amount of distortion themselves, but at frequencies above 500Hz this will be very small indeed.

This circuit will only work with 'conventional' non-inverting power amplifiers.  Simple inverting amps (such as the Zen) cannot be tested, as the feedback will be negative instead of positive, and the amp will not oscillate.

When the unit is connected to the amplifier under test, make sure that the pot is set to the minimum (fully anti-clockwise, or wiper at ground).  Power on the amp, and carefully advance the pot until oscillation starts.  Set the level to just below the clipping point (or any other level you want to test at), and the lamps will stabilise the level at the preset value.

As shown, the circuit will oscillate at a little over 1kHz, and this seems to be the most common test frequency for distortion measurements.  Frequency is determined by the formula

f_o = 1 / 2π × √( L × C )   where L is inductance in Henrys and C is capacitance in Farads

Figure 2 shows a simplified version of the distortion meter project, and is designed to operate at the same frequency as the amplifier-oscillator combination.  By very careful adjustment, it will be possible to remove the fundamental frequency completely, leaving behind the distortion to observe, listen to or measure.

Figure 2 - Simplified Distortion Meter

For the selected oscillation frequency of about 1kHz, the values for the resistors and capacitors are ....

This provides enough range to match the distortion notch filter circuit to the amp's oscillation frequency, but it may be necessary in some cases to slightly increase (or reduce) the value of the 2.2uF capacitor in parallel with the inductor if the component tolerances move the frequency too far from the nominal frequency.

The opamp requires a +/-15V supply, and is used to increase the Q of the circuit.  If this were not used, the reading at 2kHz (2nd harmonic) would be well below the true value, making the reading useless.

To measure the distortion, set all tuning pots to the mid position, and switch to the CAL (Calibrate) position (the switch is open).  With the input level at minimum, apply the signal to be measured from the power amp.  The applied voltage must be greater than V RMS.  Adjust the Q control to about half or less.

• If you are using a millivoltmeter (not digital!), set it to the 3V range, and advance the input level until the meter reads full scale.  Set the switch to READ.
• Carefully adjust the  fine tuning controls (they are interactive) until the  minimum possible voltage reading is shown, adjusting the range on the millivoltmeter as you get lower  readings.  Advance the Q control and repeat until Q is at maximum, and you have the minimum voltage  reading.
• Make sure that the input level control is not changed during the measurement, as the resistance affects  the notch filter tuning, and you will have to re-tune the filter.

If you were to obtain a final reading of 7mV, you can now determine the distortion + noise ....

THD% = ( V2 / V1 ) x 100    Where V1 is the initial voltage and V2 is the lowest reading
THD% = ( 0.007 / 3 ) x 100 = 0.23%

I strongly suggest that you use a small amp and speaker (or headphones) to listen to the distortion + noise output.  Without an oscilloscope, this is an excellent way of deciding if the distortion is 'nasty' or 'acceptable'.  Nasty distortion has a hard edge to it, and tends to sound ... well, nasty!  Acceptable distortion will be an almost pure tone, without any sound of buzz.  There will always be a certain amount of hiss, and a really good amp will have you listening hard to hear the tone buried somewhere in the background hiss.

So there it is.  All the circuitry can be installed in one enclosure to make a complete compact test unit, without having to go to all the bother of making (or the expense of buying) a low distortion oscillator.  Make sure that the distortion meter section is well shielded so the amplifier output cannot be capacitively coupled to the filter or opamps, as this will give an incorrect reading.  My thanks to Alfred for the idea, and I hope you can get some valuable use from it.

Main Index

Projects Index