• Please review our updated Terms and Rules here

9-track magnetic tape unit maintenance and operation?

matthew180

Experienced Member
Joined
Apr 27, 2007
Messages
56
Location
Central Florida
I finally managed to get a 9-track tape unit, a re-branded CDC 92185, and I spent the weekend cleaning it up and replacing the deteriorated foam (full of 1-inch foam to help with noise, I suspect). Being a new tape unit owner, I'm now full of questions.

So I cleaned the tape path as per the manual and fired it up. It runs and passes a lot of the maintenance and diagnostic tests, but fails on some others. One of the errors is "03", which is "ID fault", and the documentation says:

"STU displays Fault Code 03 during Operator Test 1 due to a failure to read or write the PE identification burst at load point. The probable cause of this fault is damaged tape. The tape should be free of defects within the first six inches of tape after the BOT reflective marker."

* STU : Streaming Tape Unit
* BOT : Beginning Of Tape

I get this error when I use brand new tapes (I have sealed new-old stock, 3M 703 and 700 Black Watch tapes). However, when I use a tape (also 3M brand) that came with the unit and had data on it from back when it was in service, it passes this particular test, but fails on a different test later in the procedure.

This makes me wonder if new tapes need to be initialized or formatted in any way? Where, where, how is this PE Identification Burst written? I don't see any mention of such a process in the manuals, but I have never used an actual tape before. The description says "failure to read or write", but it does not say which, and that also hits that the test should be writing the PE Identification? Maybe since the tape with data probably has this ID already, it passes? I'm just guessing, and any information or insight would be greatly appreciated, and more questions to come for sure. ;-)
 
Thanks for the links, that is good info. But... How does a new tape get its ID burst? Is this something the tape drive can write, or do tapes come with the burst already written?
 
They're written at BOT by the drive. Before that, the tape doesn't "know" anything. Regardless of the certification label (800,1600,3200,6250), new tapes are all blank and basically the same (e.g. you can write and read 800 NRZI or 6250 GCR on the same tapes).
 
I'll keep checking the manuals, but so far I'm not seeing any maintenance routines in the tape drive for writing the ID burst to a new tape. Unless the diagnostic routine that is failing to read the burst is also responsible for writing the burst. Strange that it can write a tape full of ones, zeros, or a pattern, but it would not be able to write the ID?
 
I'll keep checking the manuals, but so far I'm not seeing any maintenance routines in the tape drive for writing the ID burst to a new tape.

that is performed by the tape formatter the first time that a new tape is written

does this drive have a Pertec or SCSI interface?

you can isolate if there is a problem with the drive if you can find an already written 1600bpi tape

if you're in the bay area, I can give you one
 
The tape is a re-branded (Honeywell Bull) CDC 92185 / 92181. I don't think it was ever known by a common model number (sure would be nice if it was). I found the correct documentation here:

http://bitsavers.org/pdf/cdc/magtape/92181/

It has an 8-bit SCSI interface with an HD50 connector and identifies itself as a "CDC 92185" via SCSI ID polling.

I am not in the bay area (wish I was), I'm in SoCal. I have some tapes, 3 new, 2 that had some sort of data on them. I have also ordered some of the elusive 3M 777 tapes that are referenced in the manual. I cannot find any information about them or why they are special.
 
777 (or 777GP) tapes are not the best stock to use. I've had numerous problems with 777 tapes written around 1969-1978 with sticky ills. Stay away from 1980s Memorex MRX IV as well. IBM, Graham, BASF, etc. are a better choice.

Basically, if you've got a SCSI interface, set the density to 1600PE (I think there's a manual switch or button for that), rewind to BOT and start writing. You should write an ID burst.
 
I need some reliable tape. Any suggestions on where to get a reel I can trust? I don't know if my problems are the head, the electronics, or the tape. I wonder if it is possible to find a modern tape that is wider, and cut it down to the correct width?
 
Wider tapes exist, but probably not worth the effort. I've gone through hundreds of tapes, both 7 and 9 track; generally, with a few exceptions, they're pretty robust. I note the issues with the 777 tapes only because they appear to be a problem in the jobs that I've gotten.

Does Athana still have tapes in stock? They did, as of a couple of months ago.

http://athana.com/html/ctape.html
 
Very timely thread created here. I acquired a 9-track Cipher tape drive from a co-worker several years ago. I don't recall the model number off-hand, but it looks very much like this unit. It's been in mothball since then, because I never had any hardware to interface it to nor any software to drive it. Stumbled over it in storage a couple of weeks ago, and have been thinking I would pull it out and fire it up again and get on this forum to see if anyone has any PC interface hardware and/or software.

The cable looks like it has a SCSI interface. The document cache on bitsavers.org has a ton of info on how to create an interface cable, so may add this to my list of projects.

I have a couple of 9-track tapes in a closet from college days. They're mid-70's vintage, so could have the problems that Chuck G. mentions. No idea on the widths. Would share one, but they (hopefully) still have content I'd like to preserve. Also, they haven't been used in 35 years, so hardly "known good" specimens.

Would be a hoot to browse the contents of those time capsules!
 
Last edited:
1/2" mag tape is governed by ANSI specs--at least in the matter of width, they're all the same.

Bake them before you read them. Also clean them, if you have access to a cleaner.

It's no big deal to read 50+ year old tapes. It's really amazing how durable they are, even when infested with mold.
 
Last edited:
...

Does Athana still have tapes in stock? They did, as of a couple of months ago.

http://athana.com/html/ctape.html

My eyes!! ;-) That website looks as old as the tapes they are selling. Every page on the site is a new adventure. Haha. I'll have to give them a call and see what the prices are, since they don't show pricing except to say the minimum order is $100 (ouch). They list "re-certified tapes", and it would be nice to have one that I know is good so I can get on with testing / fixing any problems.

I wonder how you clean tapes off of the tape unit?


... because I never had any hardware to interface it to nor any software to drive it. ...

The cable looks like it has a SCSI interface. ...

If your tape unit has a SCSI interface, then just pick up a computer with SCSI built in (I have a SunBlade 2000 in the Market Place section of the forum that has SCSI built-in), or pick up an Adaptec 2940 on ebay. SCSI cables of all kinds are cheap on ebay too, just identify the connector type you need and do some searching (and make sure you pick up a terminator). That should take care of the hardware.

As for the software, I have not messed with that part yet, but I totally expect to be able to read/write data on my tape with Unix (FreeBSD, OpenBSD, a Linux distro, etc.) utilities like "tar" (actually going to use it to write a *tape* for the first time!) or other such commands that were designed to work with tape. Maybe I'm wrong about this part, but I don't think (hope) software will be a problem.

I have a couple of 9-track tapes in a closet from college days. They're mid-70's vintage, so could have the problems that Chuck G. mentions. No idea on the widths. Would share one, but they (hopefully) still have content I'd like to preserve. ...

Thanks, but I have enough *questionable* tapes now. :) You should hang on to yours and try to read the data off of them once you get your drive up and running. I would suggest getting a known good tape from somewhere (like I'm trying to do now) so you will have a way to run the diagnostics on your drive once it is running. The diagnostics will overwrite whatever is on a tape, so you will want one specifically for running the tests.

If your unit is like mine, you can run diagnostic tests from the front panel without needing to have the drive connected to a computer. I won't even bother trying to control mine from a computer until it passes the diagnostics tests without error codes.

Here are a few photos of the one I have. It is branded "Bull" on the front, but the firmware via the SCSI interface identifies the unit as a "CDC 92185 Streaming Tape Unit". Inside there is a label from the CDC/Philips Joint Venture "Laser Magnetic Storage International", and under that label it is a 92181. Poor thing is a physical manifestation of an identity crisis, I mean how many model numbers and names can one tape drive have??
stu1.jpgstu2.jpg
 
Nice! Where/how did you stumble upon it?

If your tape unit has a SCSI interface, then just pick up a computer with SCSI built in

I just acquired a Sun UltraSparc 10 with Solaris 2.6 still loaded up and running. Was wondering what to do with it, and now perhaps I've found its purpose. :cool: I downloaded the manuals from bitsavers.org and browsed through the diag procedures. Will drag out the unit and crank up the onboard diagnostics at least as soon as my daily workload allows. It's a Cipher M990, which is a horizontal front load unit. Size and shape of a large 60 gallon cooler. I don't have a picture of my actual unit, but here's a sketch from the manual.
M990.JPG

Poor thing is a physical manifestation of an identity crisis, I mean how many model numbers and names can one tape drive have??
:lol::lol:
 
You're expecting newly-manufactured tapes today? Offhand, I'd say that the market is very small.

I wonder how you clean tapes off of the tape unit?

You use a tape cleaner/certifier machine. Mine is equipped with a tungsten carbide blade that could neatly sever fingers from your hand, a vacuum pickup and various features, such as leader measurement. (In the old days, if a tape started to get errors, it was usually on the front end; so you simply discarded a hundred feet or so and attached a new BOT marker). I use isopropanol to clean the junk from the cleaner surfaces; in the old days it would have been Freon TF.

I've also got a gizmo that adds a film of cyclomethicone for the really sticky tapes.

They're not common on used market, but they do come up from time to time for sale. Exempli gratia
 
Last edited:
I use isopropanol to clean the junk from the cleaner surfaces; in the old days it would have been Freon TF.

Thanks. This was going to be my next question. The Tech manual for this drive says to use Florochloro-whatsy-who to clean the heads, but as you mention it's Freon based and long gone. My unit will definitely need cleaning before I put any tapes in it, so good to know what works in today's environment (pun intended). ;)

As matthew180 says, I should try to get my hands on a test tape reel before putting in my data-laden ones.

Question for you, ChuckG: My tapes were written under CMS on an IBM 370 mainframe back in 1979/80. No idea of the write density or data format. Was the actual data format standard back in that day? Meaning, did IBM have proprietary data block formats or such back in that day? This is way way down the road after I get the tape drive working and successfully hooked up to something. Just wondering.
 
Probably 1600 PE; 6250 was a bit more common later. As to data format, that was mostly under program control.

That's the wonderful thing about tapes--other than (optional on some systems) standard header and trailer records (e.g. VOL, HDR, EOF, EOF etc.) there weren't really any standards. Here's a reference on current standards. The difference is that instead of ASCII, S/370 would be EBCDIC.
 
Back
Top