magnetic tape analog record/play>ADC>dsp

Go To Last Post
17 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,

I have some analog signals to store in analog form on magnetic tape, and eventually use the AVR ADC for dsp stuff later...

Here is a nice tutorial...

http://hyperphysics.phy-astr.gsu...

Any suggestions? Perhaps you have experience in hacking an off-the-shelf cassette tape recorder, like this one?

thanks!

Last Edited: Wed. Oct 4, 2006 - 07:23 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, that really depends on your signal (amplitude, frequency, output impedance, etc.). If it is in the audio frequency range, it should be straightforward to adapt your signal to get into the mic or line input of the tape recorder. You may need some preamplifier if it's too weak.
If the frequency is higher, you better check the limitations of the tape itself and if it's possible you will have to replace or extend the bandwidth of the amplifier sections which drive the recording head, or find some digital tape recorder which will probably have a better bandwidth.

btw, what's your AVR doing here?

David

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I don't have specs yet, because I don't know the frequency range capability of basic tape recorders. I anticipate the range to be at least within the audio range, and probably sub-audio. Perhaps using a frequency generator and scope to find out the capabilities of the magnetic tape.

The AVR will perform digital signal processing of the analog signal.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Music cassette recorder is 50-15K like FM radio. Dictation recorders had DC bias, had worse freq resp. Those are digital now days.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You should be able to find plenty of information about the subject,
including recording formats etc. when looking some 20 years back.
By that time, recording computer data on consumer magtape records
used to be quite standard, and the computer interface circuitry
usually consisted of nothing else but an analog comparator for the
computer input side, and a small filter for the computer output
side.

Perhaps you need to walk to a library for that, as the Internet
hasn't been a common medium by that time, so most of the articles
about it probably aren't online.

I could certainly dig you up some Z80 code for that kind of audio
interfaces from my old CP/M diskettes, provided my 5.25" floppy
drive still works.

Jörg Wunsch

Please don't send me PMs, use email if you want to approach me personally.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,

Why do wish to record the signals on tape? What is your objective? What sort of data are you recording?

A tape recorder makes many assumptions about the nature of the signal being recorded. The frequency response is not flat, equalization boosts the high frequencies on record and reduces them on replay but that means that the signal amplitude reduces with increasing frequency. You will probably only be able to get a flat frequency response over a range of 200Hz to 3KHz. The phase response will be terrible so the waveform shape on replay will not match what you recorded.

The signal to noise ratio will only be 40-50dB (~200:1) for data. Also you will have to disable Dolby Noise Reduction or equivalent for data recording.

Usually analog data recorders will modulate the signal onto an FM carrier that is then recorded onto the tape. It will then record this carrier - the overall frequency response will probably be less than 1KHz but you can then record all the way down to DC - speed variations of the tape will appear as noise in your signal though so the S/N ration will still only be ~40dB.

Overall I would not recommend this approach these days when there are so many other approaches - if you could explain your problem in more detail we may be able to suggest something.

kevin

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

> the overall frequency response will probably be less than 1KHz

Oh, 2400 Bd used to be standard, 3600 Bd have been possible back
in ``those days''.

Jörg Wunsch

Please don't send me PMs, use email if you want to approach me personally.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

kevin_white wrote:
...
A tape recorder makes many assumptions about the nature of the signal being recorded....

Right, however the $19.00 USD GE tape recorder I just bought from a local store looks like an easy hack to the tape head.

So really what is at stake is the recording head and the magnetic tape. How much simpler can I get for analog signal recording?

dl8dtl wrote:
You should be able to find plenty of information about the subject,
including recording formats etc. when looking some 20 years back.
By that time, recording computer data on consumer magtape records
used to be quite standard, and the computer interface circuitry
usually consisted of nothing else but an analog comparator for the
computer input side, and a small filter for the computer output
side.

Perhaps you need to walk to a library for that, as the Internet
hasn't been a common medium by that time, so most of the articles
about it probably aren't online.

I could certainly dig you up some Z80 code for that kind of audio
interfaces from my old CP/M diskettes, provided my 5.25" floppy
drive still works.

No doubt! DSP literature appropriate for the AVR is best found mid 1960! The best books are from the local university, and are threadbare. Very interesting study however. Most DSP books today don't even consider the Z-tranform may be inappropriate for smove DSP calculations - that discussion by Blackwell is lost in history.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Digital recording and analog recording are two different block diagrams... digital tape is like floppies and hard drives... Manchester Biphase Mark at 1650 bits per inch at 15 inches per second..

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dl8dtl wrote:
> the overall frequency response will probably be less than 1KHz

Oh, 2400 Bd used to be standard, 3600 Bd have been possible back
in ``those days''.

Agreed - that 1KHz figure was when using FM to store an analog signal so as to preserve the DC level and waveshape - the OP said he wanted to store an analog signal in analog form!

Direct recording could go to a few KHz for digital data - the old Kansas City standard used a 2400Hz tone to store data at 300 bps. I modified that to use one cycle per bit to get 2400BPS with no trouble - the bandwidth for biphase encoding (or similar such as Miller/delay encoding have a spectrum that is about 1/2 of the bit rate).

Modern tape drives can squeeze up to 100K transitions/inch and have tracks less than 1mil wide to get very high data densities.

kevin

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Sonos, you don't want to be doing anything at the tape head if you can at all avoid it... as others have pointed out, the audio to be recorded is frequency shaped and has a high-frequency bias signal added to move the signal flux change out of the hysteresis region of the tape.

Assume that a cheap cassette will give you a signal bandwidth of around 150-8000Hz +/- 3dB, with a horrible phase response and a signal to noise ratio of about 35-40dB if you're lucky. It will also be subject to lots of wow and flutter and probably variable capstan speed too... all in all, not a good place to store signals if you have something better to hand.

If your signal is within that bandwidth, you're probably better off recording it straight to your hard disc through the soundcard - make sure you sample at least two and a half times or so more than your highest desired frequency. So to match the cassette recorder, 16 bits at 22ksps or better yet 44ksps.

If your signal of interest is not within that frequency range (e.g. it's around 1MHz) but has a bandwidth of only 8kHz, you could mix a suitable AC signal with it to generate an alias within the audio band and sample that.

Neil :p

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

> Agreed - that 1KHz figure was when using FM to store an analog
> signal so as to preserve the DC level and waveshape - the OP said he
> wanted to store an analog signal in analog form!

Ah, I missed that detail, yes.

Jörg Wunsch

Please don't send me PMs, use email if you want to approach me personally.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

kevin_white wrote:

Direct recording could go to a few KHz for digital data - the old Kansas City standard used a 2400Hz tone to store data at 300 bps.

kevin

Kevin,

That takes me back to 1980/1 when I used the Kansas City format to store 6809 assembler code. While looking for something else last weekend, I "found" the Tandy tape recorder that I used back then. I wonder if it still works .....

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Great discussion, thanks everyone! I did manage to record a variable sine wave onto the tape directly at the tape head, and it was audible when the tape head was resoldered to the component board. But could not view the sine waves directly from the unsoldered tape head on the oscilloscope; maybe I don't have the wireup correctly. Overall better results were obtained using the sound card approach instead.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The tape head output would be in the microvolt/millivolt range, that's why you can't see it on the scope. And unless you used a bias signal while recording, you won't get anything resembling a sine wave back from the tape.