Can RF bias be used to cancel crossover distortion?

One could argue here that the HF bias forces the curve (samples) for low frequencies to average the +Hc to -Hc curves, forming a net (average) curve that always goes through the zero point, so may be in this way that hysteresis effects are reduced. Seems very plausible, now that I have thought about it. The sums are probably a bit messy though.

The author does make a statement that is obviously incorrect though, to wit, "The frequency of the bias is not critical so long as it is greater than the maximum signal frequency ". Essentially, *any* non-linear process will result in sum and difference terms showing up, so if the bias were say,

20khz, and the signal were 19khz, then a 1khz signal would be produced. So, on this basis alone, one needs 2F. Typically, the bias is always > 60kz and that is for a machine with a 15khz signal BW.

Kevin Aylward

On a sunny day (Mon, 13 Oct 2008 19:16:56 GMT) it happened "Kevin Aylward" wrote in :

Yes, that makes sense. However I had a Philips 4408 tape recorder with IIRC 40kHz bias...

The funny thing about that was, that if your recorded silence at 19cm/s, and then played it back at 4.75 cm/s, (1/4 speed) you could hear about 10kHz... It had Ge complementary outputs. So, it had 18kHz bandwidth, 2 x 18 = 36 kHz, and bias at 40 kHz....

This paper corresponds to how it was taught to me, and is from about the same year:

There were so many other issues with magnetic tape. For example 'virgin' tape, tape that was never magnetised, had much less noise. Some recordings for radio broadcast were made on virgin tape (PER525):

()%3B%22%3E%20%7C%20%3C%2Fa%3E&md5=3d41fae36860357fe3fe15f019982eb5

Also the bias symmetry (the waveform) needed to be very precise to reduce noise, some tape machines had a special adjustment for that (Studer?).

snip

noise, some tape

Not on Studer, up to the A80 range, maybe Ampex. The A80 IIRC had a master LC oscillator at 240KHz for bias, but this was divided (SN7492?)down to 80KHz for the erase then through a LC filter for each signal, no funny distortion tweaks, just simple

martin