floppy drive music player

[jesrandall]

I am trying to create a driver circuit to allow recording and playback of
sound to a 3.5 inch floppy disk using a standard floppy disk drive...

I know that normally information is recorded to the disk in a modulated
format (specifically MFM - modified frequency modulation) and that it is
normally stored in sectors and other logical structures. However, for my
project, I want to record "raw" sound (no file format) directly to the
disk (similar to the way information is stored on a record or cassette
tape)

I don't want to use exiting floppy drive controller chips, although I
don't think these would help because they are designed around PC formats
anyway (perhaps they could be modified), but I would like to avoid using
them.

If it helps, here is some of my current progress.

I have set up a simple PIC microcontroller programmed to keep track of
which track the drive is on and automatically increment the track after a
revolution. This way I can record (starting on track 0 all the way to
track 79 and then switch sides and go from 79 back to 0) Since the disk
spins about 6 times per second (360RPM) that turns out to be about 26-27
seconds of recording time. The microcontroller will eventually allow for
simple control of record, pause, and playback functions. The problem I've
run into and hopefully only, is that I am unable to record!! lol however I
am able to "change" the information on the disk and presumably read it.
Since I don't have an oscilloscope I have connected the read output of the
disk drive to a piezo speaker and I listen for sound changes. The write
process is set up like this: when the user wants to record, write enable
is set low (which is the "go active" value (negative logic)) so then the
drive writes to the disk whatever is on the write input pin....

So....lets say I connect the write input to ground and write enable is
enabled. It would seem like it would just record a blank space on the
disk, wouldn't it??? But apparently this doesn't happen, instead it
records a steady hissing noise (frequency is unknow, i don't have a
frequency counter either...) when I connect the write data pin to positive
the same thing happens. Using a 555 timer to generate tones I was able to
make beat sounds during playback like hissssssssss-pop-----hisssss....etc.
If I set the 555 timer to a very high frequency it produces almost
complete quitness, but not quite there is still a little background noice.
And finally, Once I have recorded that quiet space, I can tap the input on
negative or positive while recording over it to produce clicks and pops
that exactly match my tap rate.. similar to just connected a speaker to a
battery. It would seem if all this happens I would be able to easily
record music or atleast tones (faster beats/sec , thats determines the
pitch of the tone), but it doesn't work.

Does any body know anything about what actually happens when you record or
write/read to a disk??
I've read a lot about flux reversals and stuff like that but I was
convinced music/sound could still be recorded in my way mentioned above.
Has anyone else ever tried or succeeded doing this?? I would really like
any information or suggestions.
Thank you very much. I hope I was clear with my question. If you need any
specifics answered about what I've already done I'd be happy to answer.

 

jesrandall wrote:

thanks thomas, your suggestions are very helpful, i hope it works.
Are
there any sites or books that I can learn more about floppy drive
details
such as the bits/track you posted.

You suggested I use a digital signal, but I actually already am... I
havn't tried using analogue audio yet...I'm just trying to record
digital
tones produced by 555 timer, right now. The FM idea sounds good...
but why
can't I use the low frequency (2000-3000hz) audio sound im using
write
now... is there a physical attribute with the floppy drive heads that
prevents this.... I don't mind using FM but I'd like to understand
why I
have to. Also, the strange effect I got... when I held the write
input low
but still got a tone on the read pin. I know according to MFM format
all
zeros are mixed with a clocking signal (explainging the tone I got),
but
does the drive automatically convert incoming signals into MFM
format???
Thanks a lot for helping me...

-Jesrandall

I faintly recall that floppies had a very simple write circuit- don't
recall the part-number, but the read circuit companion was the Motorola
3470. The head was center-tapped and the two-phase drive was generated
by the 34???.
Even after floppies' electronics were reduced to One Big Chip, the
original parts' functionality was kept.
Sorry.... it's been a while.
Wade

 

[Michael A. Terrell]

jesrandall wrote:

I am trying to create a driver circuit to allow recording and playback of
sound to a 3.5 inch floppy disk using a standard floppy disk drive...

I know that normally information is recorded to the disk in a modulated
format (specifically MFM - modified frequency modulation) and that it is
normally stored in sectors and other logical structures. However, for my
project, I want to record "raw" sound (no file format) directly to the
disk (similar to the way information is stored on a record or cassette
tape)

I don't want to use exiting floppy drive controller chips, although I
don't think these would help because they are designed around PC formats
anyway (perhaps they could be modified), but I would like to avoid using
them.

If it helps, here is some of my current progress.

I have set up a simple PIC microcontroller programmed to keep track of
which track the drive is on and automatically increment the track after a
revolution. This way I can record (starting on track 0 all the way to
track 79 and then switch sides and go from 79 back to 0) Since the disk
spins about 6 times per second (360RPM) that turns out to be about 26-27
seconds of recording time. The microcontroller will eventually allow for
simple control of record, pause, and playback functions. The problem I've
run into and hopefully only, is that I am unable to record!! lol however I
am able to "change" the information on the disk and presumably read it.
Since I don't have an oscilloscope I have connected the read output of the
disk drive to a piezo speaker and I listen for sound changes. The write
process is set up like this: when the user wants to record, write enable
is set low (which is the "go active" value (negative logic)) so then the
drive writes to the disk whatever is on the write input pin....

So....lets say I connect the write input to ground and write enable is
enabled. It would seem like it would just record a blank space on the
disk, wouldn't it??? But apparently this doesn't happen, instead it
records a steady hissing noise (frequency is unknow, i don't have a
frequency counter either...) when I connect the write data pin to positive
the same thing happens. Using a 555 timer to generate tones I was able to
make beat sounds during playback like hissssssssss-pop-----hisssss....etc.
If I set the 555 timer to a very high frequency it produces almost
complete quitness, but not quite there is still a little background noice.
And finally, Once I have recorded that quiet space, I can tap the input on
negative or positive while recording over it to produce clicks and pops
that exactly match my tap rate.. similar to just connected a speaker to a
battery. It would seem if all this happens I would be able to easily
record music or atleast tones (faster beats/sec , thats determines the
pitch of the tone), but it doesn't work.

Does any body know anything about what actually happens when you record or
write/read to a disk??
I've read a lot about flux reversals and stuff like that but I was
convinced music/sound could still be recorded in my way mentioned above.
Has anyone else ever tried or succeeded doing this?? I would really like
any information or suggestions.
Thank you very much. I hope I was clear with my question. If you need any
specifics answered about what I've already done I'd be happy to answer.

How much recording time do you think you will get? at 300 rpm you
get 1/5 of a second per track. There are 80 tracks per side, so it
would be 16 seconds per side. You need a high frequency bias to record
low frequency audio. The bias should be 50 to 70 Khz and the level has
to match the magnetic media. You need to study the circuits in common
reel to reel or cassette recorders to wrap your mind around how analog
recording works.

Digital recording is simpler. Saturate the magnetic media without
burning up the head then reverse polarity and repeat. After that it
becomes the problem of timing the transitions.

--
Beware of those who post from srvinet.com!

Michael A. Terrell
Central Florida

 

[worldcitizen]

jesrandall wrote:

I am trying to create a driver circuit to allow recording and
playback of
sound to a 3.5 inch floppy disk using a standard floppy disk drive...

I know that normally information is recorded to the disk in a
modulated
format (specifically MFM - modified frequency modulation) and that it
is
normally stored in sectors and other logical structures. However, for
my
project, I want to record "raw" sound (no file format) directly to
the
disk (similar to the way information is stored on a record or
cassette
tape)

I don't want to use exiting floppy drive controller chips, although I
don't think these would help because they are designed around PC
formats
anyway (perhaps they could be modified), but I would like to avoid
using
them.

If it helps, here is some of my current progress.

I have set up a simple PIC microcontroller programmed to keep track
of
which track the drive is on and automatically increment the track
after a
revolution. This way I can record (starting on track 0 all the way to
track 79 and then switch sides and go from 79 back to 0) Since the
disk
spins about 6 times per second (360RPM) that turns out to be about
26-27
seconds of recording time. The microcontroller will eventually allow
for
simple control of record, pause, and playback functions. The problem
I've
run into and hopefully only, is that I am unable to record!! lol
however I
am able to "change" the information on the disk and presumably read
it.
Since I don't have an oscilloscope I have connected the read output
of the
disk drive to a piezo speaker and I listen for sound changes. The
write
process is set up like this: when the user wants to record, write
enable
is set low (which is the "go active" value (negative logic)) so then
the
drive writes to the disk whatever is on the write input pin....

So....lets say I connect the write input to ground and write enable
is
enabled. It would seem like it would just record a blank space on the
disk, wouldn't it??? But apparently this doesn't happen, instead it
records a steady hissing noise (frequency is unknow, i don't have a
frequency counter either...) when I connect the write data pin to
positive
the same thing happens. Using a 555 timer to generate tones I was
able to
make beat sounds during playback like
hissssssssss-pop-----hisssss....etc.
If I set the 555 timer to a very high frequency it produces almost
complete quitness, but not quite there is still a little background
noice.
And finally, Once I have recorded that quiet space, I can tap the
input on
negative or positive while recording over it to produce clicks and
pops
that exactly match my tap rate.. similar to just connected a speaker
to a
battery. It would seem if all this happens I would be able to easily
record music or atleast tones (faster beats/sec , thats determines
the
pitch of the tone), but it doesn't work.

Does any body know anything about what actually happens when you
record or
write/read to a disk??
I've read a lot about flux reversals and stuff like that but I was
convinced music/sound could still be recorded in my way mentioned
above.
Has anyone else ever tried or succeeded doing this?? I would really
like
any information or suggestions.
Thank you very much. I hope I was clear with my question. If you need
any
specifics answered about what I've already done I'd be happy to
answer.

I would think that the spacing between the concurrent "tracks" would
cause considerably crosstalk between tracks. I also concur with the
person who suggested that you must use a bias signal to enable you to
record to your chosen media. This is very common practice on analog
audio recorders. You won't record squat without that bias signal.
Good luck on your project.

 

[Zak]

jesrandall wrote:

thanks thomas, your suggestions are very helpful, i hope it works. Are
there any sites or books that I can learn more about floppy drive details
such as the bits/track you posted.

Look for floppy controller data sheets. The Western Digital WD1770 and
similar (1771, 1791, etc) are early models, and with early technology
the sata sheet may give a bit of an explanation.

You suggested I use a digital signal, but I actually already am... I
havn't tried using analogue audio yet...I'm just trying to record digital
tones produced by 555 timer, right now. The FM idea sounds good... but why
can't I use the low frequency (2000-3000hz) audio sound im using write
now... is there a physical attribute with the floppy drive heads that
prevents this....

I guessed you want to keep the drive electronics intact. That said the
rotationa speed is too high and the head is too small to give
appreciable bass response. Actually I thing the drive acts as a
differentator: you feed it a signal which it will record, but it will
respond with a pulse for every transition that you sent it.

I don't mind using FM but I'd like to understand why I
have to. Also, the strange effect I got... when I held the write input low
but still got a tone on the read pin. I know according to MFM format all
zeros are mixed with a clocking signal (explainging the tone I got), but
does the drive automatically convert incoming signals into MFM format???

No, you have to send data and clock yourself and separate them on the
controller.

FM would write the clock, then a data (or not), another clock, another
data and so on.

MFM would double the speed and leave out the clock is a data was just
written if I remember correctly, thus keeping the frequency reasonable.

But the drives are made to read and write digital signals only -this is
why I suggested FM.


Thomas