This is the partial circuit of the bias and erase oscillator from a very old tape recorder. The osc freq is normally 50kHz, producing around 70V rms at the marked red point.

Problem: the original erase head has had to be replaced and its characteristics are unknown (L, R). The manual tells nothing either. The replacement erase head is 4.82mH and 5R3

The replacement erase head seems to be responsible for changing the osc freq to 35kHz. This may not be a problem but the bias signal is breaking into the rest of the circuitry at a high level because the bias trap L cannot be set to the osc freq. Which is a problem, not least because it creates a permanent meter deflection.

The trap L does swing through an appropriate range, 2mH to 6mH which along with the 10,000pF cap along side it should tune to the osc freq. Adjusting the core of the inductor, I can bring the bias at V2 down almost to minimum but if I go even a tiny bit too fay it jumps right back up again, to around 700mV. If I adjust L2 so that the leakage is almost minimum, then power down/up the leakage at V2 is once again back to 700mV and I have not touched L2.

Obviously (but perhaps incorrectly) I suspect that the replacement erase head spec is incompatible. If that is the case, how can I change the way the erase head 'looks' to the osc without compromising erase depth AND minimising bias leakage?

 

Is the original head electrically damaged or worn out? If the latter, you can measure its inductance and resistance to compare with.
From the schematic shown, bias will be available at the switch v2 if the original bias frequency is changed. You may have to play with C2-R4 combination to get restore the bias frequency to 50 KHz.

 

It appears that it is the capacitors across the erase head that are tuning it to the lower frequency. So I suggest changeing them to a lower value, so that the original 50 KHz is the frequency. Consider that the bias is also important for the RECORD function as well, and that is also rather important. if the 4K (4000Pf?) capacitor is replaced with either 2700 pf or 2500pf that might get close enough so that 50 kHz could be obtained. And that would avoid having to fix the record bias circuit.