I make a DIY cassette recorder from a cheap cassette player. I successfully made a stereo playback circuit that works with recorded cassettes, but I'm having trouble with recording; after recording, the signal becomes distorted.

Here is a part of my recorder schematic for 1 channel (there may be minor mistakes in the schematic because I drew it here according to the already assembled prototype):


simulate this circuit – Schematic created using CircuitLab

I have a bias oscillator with output of ~51 kHz at 40-50 Vpp connected to the rec/play head through a 100 pF capacitor, and a line input amplifier with an output range of ~0-4 Vpp connected to the rec/play head through a 2.2 μF capacitor and a 100 kΩ potentiometer.

No matter how I adjust the input and output potentiometers, the signal after recording is distorted. I even tried to reduce the amplitude of the bias, but it did not help.

I could not find any information about the specific level of signals in volts for recording. Some articles say that the bias level should be 10 times the input signal level, but even so, the recording is still distorted.

So, what should be the level in volts of the bias and sound signals? And how should these signals be connected to the head correctly?


Distortion waveforms:

440Hz 440Hz

1KHz 1KHz

2KHz 2KHz

4KHz 4KHz

  • \$\begingroup\$ The absolute voltage levels required depend upon the recording head characteristics. It is impossible to give any figures without taking that into account. What bias level are you getting across the head? \$\endgroup\$ Oct 23, 2022 at 21:47
  • \$\begingroup\$ @KevinWhite my head has resistance about 237 Ohms. It's just a regular no-name streo head from cheap cassette player. I have ~17Vpp bias across the head \$\endgroup\$
    – F3RNI
    Oct 23, 2022 at 21:50
  • 2
    \$\begingroup\$ Can you describe the distortion? A good test signal set would be silence first, then a 440 Hz tone, then a 2kHz tone of the same amplitude, then the sum of both tones at half amplitude. Then an amplitude sweep of a 220 Hz tone from very silent to very loud. Look at the spectrogram of the result (eg in audacity, which you could also use to generate the test signals). What are the things that get distorted? Is it more frequency-dependent or amplitude-dependent? \$\endgroup\$ Oct 24, 2022 at 1:39
  • \$\begingroup\$ I removed the 100R resistor in the bias generator, added 1K resistors after the 100pF capacitors and it got better, but not much. @MarcusMüller At frequencies below 4 kHz, the sine is gradually cut off from above Screenshots: 440Hz 1KHz 2KHz 4KHz \$\endgroup\$
    – F3RNI
    Oct 24, 2022 at 6:16
  • 1
    \$\begingroup\$ @F3RNI reading HP's 1967 Magnetic Tape Recording Handbook, I quote: Up to now our discussion has assumed that the magnetizable medium responds linearly to the magnetizing force of the record head. As might be expected, the perversity of nature asserts itself and the assumption is in error: Actually got me laughing out :) \$\endgroup\$ Oct 24, 2022 at 10:36

1 Answer 1


Thanks to everyone who tried to help! I tried different things for a long time and finally I got a more or less clean signal.

What I did: increased the bias voltage by removing 7805 LDO and 100Ω resistor from bias oscillator power, removed the 1KΩ resistors after the bias oscillator out, added a 20KΩ resistor between the op-amp output and the head, increased op-amp gain to 11.

Thus, the schematic of one recording channel now looks like this:


simulate this circuit – Schematic created using CircuitLab

What else needs to be done: Change all capacitors in the bias generator to >100V, add a resonance circuit to the op-amp to increase output gain on high frequencies (>4KHz)

The result is not perfect, and I don't know how long the 50V capacitors in the bias circuit can withstand, but so far it works. Below are screenshots of various frequencies after recording

110Hz: 110HZ

440Hz: 440HZ

1KHz: 1KHZ

2KHz: 2KHZ

4KHz: 4KHZ

8KHz: 8KHZ

Once again, thanks to everyone who wanted to help.

  • \$\begingroup\$ And yes, to answering my own question about levels: Bias level: as high as you can (I got ~50Vpp output) Sound level: within the supply voltage range (I have ~ 3Vpp at the output without clipping) The bias output is connected to the head via a 100pF capacitor. The sound output is connected to the head through a 2.2uF capacitor and a 20KΩ resistor. \$\endgroup\$
    – F3RNI
    Oct 24, 2022 at 22:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.