Analog simulation of tape recording experiments part 2

CD players when recorded to cassette or reel to reel sound fantastic at playback.


Let's connect the line output of a CD player directly to a tape head.


This can be instantly done by connecting a Sony cassette car adaptor to the portable CD player line out.


This output impedance is measured to be 370Ω. (The usual method, the line out is connected to a VTVM while a test CD plays 1KHz at 0dB. Then a variable parallel resistor connected to the output is varied until the signal drops due to load to -6dB. The resistor was then measured to be 370Ω. The internal resistor is thus 370Ω as at that point the voltage is half by the potential devider consisting of the internal and the external variable resistors).


Then a frequency responce across the line out was made using the test CD without anything connected first and then with the cassette head, (car cassette adaptor connected),


Without anything but the VTVM connected to line out the frequency response of the voltage at the line out is flat from 20Hz to 20KHz with half or so dB at 20KHz. (0 dB assumed the -0.5dbm measured at 1KHz).


The inductance of the head of the car cassette adapter is 54.5mH.


With the car cassette adaptor head of 54.5mH the frequency responce at line out (due to the shunt loading) is modified to:


Ref 1KHz (-2.5dBm) assumed 0dB


20Hz        -7dB
23.1Hz     -6dB
63Hz     -1.5dB
80Hz     -1.3dB
125Hz   -0.5dB
160Hz   -0.5dB
250Hz       0dB
315Hz       0dB
500Hz       0dB
800Hz       0dB
1KHz        0dB
1.25KHz   0dB
2KHz        0dB
3.15KHz   0dB
4KHz        0dB
5KHz        0dB
8KHz        0dB
10KHz      0dB
16KHz      0dB
20KHz +0.5dB


Considering the so low inductance of the head there should have been more low cut. Where is the error? Or is there no error? Is this a reason reel to reel sounds better than cassette, by keeping the rising slope soecifucation (constant current with frequency) through many octaves? Reading revealed that using a mu metal core, the inductance of any given number of turns is reasonably constant below 50Hz but above that frequency gradually tends to vary inversely proportional to frequency, so that above about 600Hz the law is such that the inductance has a constant reactance (Babani). So it seems it is mumetal head.


Next a Pleiades filter of 53mH is connected across the portable CD player line out and the voltage at the line out terminal (of one channel) is measured while the test CD is playing.


Frequency response  ref to 1KHz:


20Hz      -30dB
63Hz       -29dB
80Hz       -28dB
125Hz     -24dB
160Hz         -20dB
250Hz         -13dB
315Hz        -10dB
500Hz         -5dB
800Hz         -1dB
1KHz            0dB
1.25KHz     +1dB
2KHz          1.5dB
3.15KHz       2dB
4KHz.         2dB
5KHz.          2dB
8KHz           2 dB
10KHz         2dB
16KHz         2dB
29KHz         2.5dB


This is a -3dB point at 800Hz much closer to the theoretical exoected near 1KHz by the relation, line out Z=370Ω divided by 2 the result divided by 3.14, the result divided by 53mH = 1.1KHz. The frequency response looks much more like a reel to reel inverse playback characteristic.


Next on a Ferrograph series 6 at its recording amplifier, the voltage across the recording head may be measured. Same for the playback amplifier.


A 70μs EQ characteristic as used on 7.5ips corresponds to a turnover in Hz as follows. We inverse 70μs, ie 1/70μs To fine the angular frequency of the phasor (rotating vector) and then divide by 2 and then by 3.14 to find frequency. This gives 2.275KHz. It is the turnover frequency, (3dB point) for the playback integration, or high cut 6dB per octave to stop.


Reference:
Coil Design and Construction Manual - B.B. Babani - page 79