Destructive Imaging of DECTAPE II Media

[Mike Stein]

I can't find it at the moment but somewhere I have a Burroughs digital 
cassette drive, capstan-less and capable of high-speed bidirectional seek. 
AFAIK it uses the standard half-track mono tape format, one track for data 
and the other for the clock.

Don't know if it'd be useful and not sure if I want to part with it 
permanently, but keep it in mind if nothing else works out and you get 
desperate ;-)

mike

----- Original Message ----- 
From: "Mark J. Blair" <nf6x at nf6x.net>
To: "General Discussion: On-Topic and Off-Topic Posts" 
<cctalk at classiccmp.org>
Sent: Tuesday, January 27, 2015 5:36 PM
Subject: Re: Destructive Imaging of DECTAPE II Media


Responding on-list to an off-list message, with permission...

On Jan 27, 2015, at 09:30 , Lou Ernst <louis.ernst at verizon.net> wrote:
>    This is not silly. I think it can work. I had a similar idea about TU60 
> decassettes. My task was easier there though since the TU60 uses standard 
> Philips compact cassette shells. The TU60 however has no capstan and uses 
> the spindle motors to maintain tape temsion. It was designed by the same 
> guy who designed dectape (TU55/56).

Using the spindle motors to maintain tape tension seems like a better scheme 
than that danged drive belt in my opinion. I recall reading somewhere online 
about the heroic measures that some engineers (at HP? Or was it 3M?) had to 
go to to get a DC-100A-like cartridge design working, involving lots of hard 
work in the areas of lubricants, surface finish, and how to make that drive 
belt by stretching out a small disc of plastic. My interpretation is that 
they had to apply heroic efforts to make a crappy design barely work. ;)


>     Long story short, I was able to play TU60 cassettes in a good audio 
> cassette deck and digitize the audio on a PC. I manipulated the audio file 
> in Audacity. I was able to cut single data blocks and dump the blocks to a 
> long .CSV file. I opened these in Excel and created intelligence in the 
> spreadsheet to digest the audio samples back into the block data, even 
> properly decoding the header and checking the checksum. The hardest work 
> in the spreadsheet was tracking the bit boundaries due to the temporally 
> varying bit-rate (no capstan in the real TU60).

Cool!

>     At last I did not have to use this means of recovering data from TU60 
> cassettes. My real TU60/TA11 worked well enough to recover the data on the 
> cassettes I was provided to recover. I developed that tape deck/digitizing 
> method in case there were dropouts unrecoverable by the real TU60.

At least the real TU60 isn't hobbled by that dreadful cartridge design.

>     Perhaps you can build a sort of open frame reel-to-reel deck to play 
> and digitize the tu58 tapes.

That's what I'm thinking. While I could repeatedly use one or two good 
cartridge/belt assemblies with the guts of multiple tapes in order to image 
them in a TU58-XA transport, I don't relish the thought of repeatedly 
tinkering with the danged belt if I can come up with a better solution.

I found a specification that includes the magnetic track dimensions of the 
DECTAPE II on Al's site:

    http://bitsavers.trailing-edge.com/pdf/dec/dectape/tu58/TU58_Engineering_Specification_Jun78.pdf

Also, I found a diagram showing track dimensions for various audio cassette 
formats here:

    http://richardhess.com/notes/formats/magnetic-media/magnetic-tapes/analog-audio/015-cassettes/

in particular:

    http://www.richardhess.com/tape/cass_trk_lrg.gif

Both the DECTAPE II media and standard audio cassette media are 0.15" 
(3.81mm) wide. DECTAPE II has two .057" (1.448mm) wide tracks, centered 
0.046" (1.168mm) apart. Data density is 800 BPI, with 2400 flux reversals 
per inch. At standard read speed of 30 ips, that turns into bits in 41.7us 
increments and flux changes in 13.9us increments. The reels inside the 
cartridge will spin at around 380 to 800 RPM depending on how much tape is 
on them, if I still know how to do math. Or twice that at the scanning speed 
of 60 ips.

Just based on track geometry, it seems to me that it may be quite possible 
to read DECTAPE II media with the inner two tracks of a 4-track recorder 
head or auto-reversing stereo audio deck head. Or even with both tracks of a 
2-track 2-channel head as shown in the diagram I found, though I don't know 
how common those are. 4-track recorders and auto-reversing stereo decks are 
pretty common, though.

Running the tape at a stable speed near 30 ips instead of 1-7/8 ips would be 
the hardest part, I think. My gut feeling is that a normal audio cassette 
capstan and pinch roller assembly may not work well at 16x normal speed, and 
I haven't seen references to any audio cassette tape applications that run 
the tape at controlled speeds that fast (rewind and fast-forward speeds may 
be that fast or faster, but those run the tape at unregulated speeds with 
the pinch roller disengaged).

I don't know if audio tape heads are electrically suitable for this 
application, but I think they might be since the flux change period is in 
the same ballpark as the AC bias frequency used in cassette recorders, so 
the head coils ought to respond ok at those higher frequencies (?).

I'm curious about what folks with deeper magnetic media experience than I 
have might think about this.

-- 
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/