6502.org

Hi All

The following article on Wikipedia gives some information on how the 1530 Datasette worked:

https://en.wikipedia.org/wiki/Commodore_Datasette

Some detail is given on how the analog data on cassette was converted to digital data with which the Commodore 64 could work with.

However some brief mention is also made that when recording data to cassette, the 1530 also contained an audio filter for converting digital data to analog:
I could not get much information on the Internet on how this audio filter worked which was harnessed during recording, except a schematic from the zimmers.net website, which I have attached to this post for reference.

On the above mentioned schematic I it looks some kind of audio filter was implemented indeed during the recording process. However, I didn't had any luck figuring out how this filter works.

Can anyone perhaps give an explanation on how this filter works, or give some pointers into the right direction?

Once I have figured out how this filter works, I thought it would be an interesting exercise to implement this filter as a digital filter on an FPGA and take any *.TAP file and see if you can recreate the original sound as heard when you would play the original tape on a tape deck.

Well, I didn't see much filtering as well.

There is the combination of R23/C6 and C11/R4. R23/C6 will damp higher frequencies. C11/R4 will pass them (10x lower time constant) to the first opamp. This is a weak sort of a bandpass with only one decade of width. But the time constants of these RC combinations appears somewhat strange to me (R23/C6 = 6µs => 160kHz, R4/C11 = 60µ => 16kHz)?
The first opamp should "rectify" and limit any "audio signal" to roughly +/-1V (relative to the bias voltage). This is then fed into the second opamp which only amplifies this signal to bang this into the final transistor Q2.

The wikipedia article describes some different circuit: Your wish to recreate the sound could perhaps easier accomplished when using an audio tool like Wavelab or Audacity to trim the sound (including the more or less dull speaker!) from a .TAP file. It is most likely a bandpass, and a clipper, and then another bandpass for the speaker.


Regards,
Arne

Hi Arne

Thanks for the explanation, but I see you are actually referring to the playback scenario.

The scenario I am actually after is the recording scenario, that is when the switch next to R/W Head is in the REC position and Write data enters from the right, and also via the REC switch position enters the parallel C10/R21, through Q5 and so on.

There is no filter circuitry in the record path. All transistor stages have accellerating caps in there base drive (C10, C9) to switch a little faster and finally there is only C17/R36. To me this looks like DC blocking and attenuation and/or current limiting but not as a "filter".

I wonder why they didn't choose a 40106 (CMOS Hex Schmitt-trigger). One or two in parallel could drive the head for records. The other four could be used to reshape the playback signal. Perhaps a 4049 (hex inverter, CMOS) would even better, as they can be used as amplifiers and as schmitt-triggers as well.


Regards,
Arne

True. In fact, it's a rather poor design; but I suppose they figured that if it worked, that was good enough. C9 and C10 are just to compensate for the following transistors' input capacitance and Miller effect. The record circuit shows no effort to make anything linear or control the gain or frequency response or even the bias. It's just saturation all the way. Wow.