The tape recorder improved when its amplifier misbehaved

In 1940, Walter Weber was trying to improve magnetic tape recording at Germany’s Reichs-Rundfunk-Gesellschaft. During an experiment with negative feedback, his test circuit began to oscillate.

That is normally an amplifier’s way of becoming useless. This time, the tape sounded better.

Early Magnetophon recorders applied a steady direct current alongside the audio signal. The bias helped, but recordings still had conspicuous noise and distortion. Weber’s unstable circuit had accidentally added a high-frequency alternating current instead. The unwanted oscillation moved the tape’s magnetic material through a more nearly linear part of its response, allowing the audible signal to be recorded with much less distortion.

In Weber’s account quoted by audio historian Friedrich Engel, the first test cut noise by about 10 decibels and harmonic distortion from 10 per cent to 3 per cent. Further work produced a frequency range reaching 10 kilohertz and a dynamic range of 60 decibels. Tape had gone from an interesting medium to a high-fidelity one.

Weber was not the first person to encounter the principle. BASF had experimented with alternating-current bias earlier, and Japanese engineers published work on it in 1938. He apparently did not know this. His contribution was to notice what the malfunction had done, identify its cause, reproduce it and turn it into an adjustable recording system.

This is sometimes compressed into the agreeable legend of a lucky accident. But oscillating amplifiers are common; useful recording methods are not. The accident supplied a result that contradicted the experimenter’s expectations. The invention began when he declined to correct the fault and move on.