Tuesday, October 3, 2017

Magnetic tape saturation with no moving parts


Or instantaneous analog simulation of magnetic tape recording and reproduction.


Tape recording reproduction simulation with no moving parts.


Investigating on how to magnetically saturate not only low frequencies in signal transformers...


But instead all frequencies by constant current feeding instead of constant voltage feeding the transformer with respect to frequency as is done on magnetic heads for magnetic tape recording. Constant current feeding with respect to frequency on magnetic heads is usually done by feeding the head from a large output impedance. For example on Ferrograph Series 6 electron tube reel to reel tape recorder the recording head is fed by an ECC82 electron tube from the high Z anode circuit. If memory is correct there may be a series resistor in other circuits before feeding the head.


The great advantage doing saturation magnetically is that the rising frequency responce that results then needs low cut de emphasis which will round off the instateniously peak limited waveform.


Inductors made from coils of insulated wire usually wound on a magnetic core behave like the mass of heavy doors. They try to keep electric current constant (flow of electrons) the same way a mass tries to keep velocity constant. In fact mass is the analogous quantity in mechanics to inductance on electricity, magnetism. On doors if we try to oscillate them forth back forth, the higher the frequency of excitation the less the amplitude. Or if the amplitude has to be constant wrt frequency, the force has to be increasing with frequency. This is Newton's second law, force = mass x (rate of change of velocity or acceleration). The equivalent on electricity magnetism is that voltage across an inductor = inductance x (rate of change of current). The higher the frequency (rate of change), the higher the force, tension or voltage on doors or inductors needed, resulting in 6dB rise per octave. Mathematically this can be proved with differential equations.


So to get back to flat responce a high cut filter at -6dB per octave is needed. This by cutting high order harmonic distortion is rounding the instantaneously limited peaks at almost the speed of light?!


It is the electron population itself that saturates. Magnetism is produced by moving electrons and in the case of magnetization by their circular motion.


Recording tape has relatively few billions billions?! of electrons as it is relatively thin and one layer.


To do this at ordinary levels with transformers means less iron, mu metal, or Nanocrystalinne etc core. For example less length of mu metal tape on a tale wound ring magnetic core.


On this easier to saturate tape wound core a primary and a secondary winding can be wound. Exactly as a transformer inside a Coles 4038 but with less iron. How much iron needed will depend on the operating level of saturation or threshold of signal level.


This simulates a recording and a playback head kissing each other at their air gap.


And hopefully sound thickening tape saturation at all frequencies could be produced without the need of ac bias or nasty moving parts, pitch roller, capstan, belts, cleaning, lubrication, noisy moving parts, power consumption of motors, bulk, weight etc.


The high cut bright sounding waveform as is done by the high cut woofer on guitar amplifiers can then feed an Analog to Digital converter protecting it from overload almost at any level by the tremendous peaks of real live music.


The sound should be big and loud.


What turnover playback EQ should be used for the high cut, 70μs or higher?


Summary:


An inductor in order to saturate at the full frequency spectrum must have the voltage applied to it rising with frequency. As when trying to oscillate a door at high frequency. Inductors and doors (in general mass) have an impedance that rises with frequency. It is called inductive reactance.


This is convenient as a low cut filter is needed to bring back to fast frequency responce and its other action is rounding any saturated waveform.


This is the basis of magnetic tape recording acting euphonically for small amounts of distortion, thickening and adding volume to the sound. This is done by instateniously peak clipping or limiting the waveform and rounding it by the low pass high cut de emphasis filtering action.


At any frequency the higher harmonics produced are reduced by the low pass high cut action of the de emphasis EQ filter needed.


So only 2nd and low order harmonics are left to thicken the sound. Mike Oldfield (can be read on interviews) produces his guitar sound using similar preamplifier and EQ techniques. FM broadcasting signal processing as pioneered by Bob Orban uses similar techniques to make radio statins sound loud and great.


Tape recording has the fantastic additional property that the more it is overdriven the bandwidth is reduced. This acts as an automatic high frequency limiter. Will this also take place on the analog simulation?


The more distortion is produced the more the rounding of the waveform. High harmonics are filtered this by definition or Vourier analysis means the edges of the waveform are rounded.


Reducing the high order harmonics and leaving only the first order ones produces a waveform that sounds nice, loud, full devoid of abrupt very high peaks (a live music waveform can be full of them with a ratio higher than a million or 60dB) that will make life difficult on ops amps, A to D converters and to our ears. The requirement is not nessasarily preserving an unlistenable 60dB dynamic range that requires an 1 million watt power amplifier to reproduce edges of the waveform when the average level is 1 watt but condensing a 60dB or more dynamic range to say 20dB without audio distortion to our brain or departure from naturalness.


References:
(General on analogous systems, a beautiful book), Introduction to System Dynamics - Shearer, Murphy, Richardson - Addison Wesley


www.261.gr on instantaneous peak limiting


Bob Orban on FM Optimod


Electricity and Magnetism - Berkeley Physics Course - Purcell - McGraw Hill


Electromagnetism - Grant, Phillips


Engineering Electromagnetics - William Hayt - McGraw Hill


Music, Physics and Engineering - Harry Olson (RCA) - Dover

No comments:

Post a Comment