Hallo Abhishek,

I am not against it, as you stated. But I am not convinced that this is a good way to proceed, today. I often wondered, why recording failed, although everything was done to make it correctly running: Level control, expensive HiFi cassette player, expensive HiFi compact cassette, multirecording technique to increase redundance, a.s.o. Then I wanted to have a look, how information coming from tape really looks like. As I stated in my last reply I was shocked! I did never imagine how clever our ears and brain is, being able to hear a sinus, where actually there is no sinus, at all, on tape. From this moment on, I had the strong feeling, that proceeding with tape recording in future would not be a good idea.

Of course, there are professional tape recording machines arround, still working today, but those brillant recorders are so far away from simple cassette player like good old '8080' form a cygnal chip.

The wrong estimation about drop-outs is, that they are only short time effects. So, transmitting information with the help of Kansas City format, e.g., where '1' is recorded by 8 periods of 2400Hz sinus and '0' by 4 periods of 1200Hz sinus, would help you to avoid data loss caused by drop outs. But that's wrong! Drop-outs can last much longer than some periods of 2400Hz sinus. Much more probable are zones of decreased magnetization sensibility: Amplitude of recorded sinus is changing all the time. Longer periods (longer than a 'classical' drop-out) with drastical decrease of amplitude is very probable.

One good methode to increase redundance is to record same information three times. But you have to guarantee, that there are more correct copies on tape, being identical with your original, than damaged ones. For this, you must have the records separated so much, that a drop-out only affects ONE record. This is possible, when you record bigger data blocks. But at the same time number of drop-outs you have to struggle with also increases!!
When you use tape recorder for data logging, amount of data is so small, that original and copies can be affected by the same drop-out. Then you will not have enough information for reconstruction.
Try first to sample as much as possible datas on your board, and then store them on tape, with 2 copies. But, please keep in mind, that this is no guarantee for always having proper data!

I never remember any cassette which had recorded information absolute correctly. I worked a lot with tape recording when I was much younger. Still someone remembers PET from COMMODORE?

You asked for PLL circuit. I have here some from 1978. It consists of 12 chips, 6 RC time delays and after endless adjusting you are able to kill a man: developer of this circuit.

How to do recording on tape?
In 1984 'supertape' was a rather fast and reliable methode: Bit begins always with a toggling, low to high or high to low. For '0' at half the time of one bit another toggling occured. For '1' no toggling occured. You can read this information by detecting high to low transitions, perhaps with interrupt. Then three possible time intervals between these edges have to be taken into account.
Advantage of this methode is that changes of RPM of player motor is not so dangerous. But, heaven, when there's a transition missing from drop-out or decreased amplitude, then good night!

Finally, I want to focuse, that I do not want to recommend any tape recording. But it's on YOU to decide. And after you had asked for help, I tried to give help.

Good luck,
Kai