I came across a whole bunch of these recently:
I know there’s data on there but have absolutely no idea how this worked.
I started looking through old pictures of computers to see if I could figure out what machine they went with and, unsurprisingly, I’m stuck. My first guess was this:
But then I found this slide that shows the same kind of reels I have but with a very different keyboard than the one above:
There is no chance whatsoever that I will figure this out on my own so if you know what this is all about, I’d really appreciate your help. Thanks in advance!
Richard Schaefer can probably go into more detail, but what you show are known as “mag tapes” to the in crowd. Mag tapes probably were first used on IBM mainframes, but became de rigeur on many different computers, from mainframes to minis. We used mag tapes for backing up our minis in the EECE Department for several years. Top pic is of a mainframe’s tape drive bank. the bottom one is probably for a minicomputer. The label on the depicted tape suggests it contains pulsar data, probably from someone in Space Science.
Actually the bottom picture probably represents a typical office in the late 70’s. The user has a terminal and keeps a rack of his/her important tapes in their office. Prior to around 73-75 few terminals existed. Most work was via batch card/tape input. time sharing was just becoming available in the mid 70’s. I think my ELEC 220 course (fall of 72) may have been the first to have interactive labs (about 1/3 of the labs were APL using APL terminals in the terminal room). When I took ELEC 421 (compilers) most of the students where still using card decks (some two or three boxes) and reading them thru the card reader (or having the operator do it since they were so big). I was working for ICSA then so I loaded my parser input deck into a file on the 370 and then would edit it via TSO and submit it through the internal ‘card reader’. Still had to wait for the operators to dispatch the output to the bin but reduced the edit/submission time.
ENGI 240 (E.C.Holt and Ed Feustel) in Fall 1970 used the Teletypes attached to the Burroughs 5500. We could not use cards at all or save files; we had to punch paper tape as a backup and reload it if we could not finish our labs during our assigned 1-hour lab session. Saturday mornings were FCFS if you needed more time to finish.
We were the first students taking the 220 (there also was 221 for non-compsci people and 222 for business). It was self-paced and I along with David Dyche tended to move thru it as quickly as Dr. Feustel could produce the modules.
Depending on how they have been stored and how long, they may no longer be readable anyway due to a phenomena called “print-through”. The magnetic image on one wrap of the tape will slowly affect the magnetized coating on the next layer (and vice versa) as all the bits of magnetized material attempt to align with each other. On audio tapes, this sounds like a weak echo, on digital tapes it simply becomes read errors.
And before the EEs, physicists and materials folks out there jump on me, I know that “bits of magnetized material” should be more properly called domains. 🙂
I don’t want to read them. I just want to know how one would have read them.
Just saw your reply above. IBM 2400/3400 series units or DEC-20s, I believe, were the way to go for that
There were even 700-series and 1400-series tape drives that used the same form factor reels before that. The specific tapes would have been manufactured during the 2400/3400 era, though, as they say they are tested up to 6250BPI. The label says the data were recorded at a lower density (1600BPI).
Don’t get too excited about the potentially sexy pulsar data on those 9-track reels. The frustratingly short (and poorly documented) life expectancy of magnetic tape is 10-20 years—at most 30, if you recovered it from inside some sort of mystical force field.
My mystical force field has been acting up lately . . .
Based on the label, the data on the tapes probably belonged to Space Physics. (“PV ORBITER DATA ORBIT: 4458”) One might guess that it was recorded on either one of the computers in Spec Physics or some special-purpose not-quite-computer recording equipment. Somebody who worked over there may recognize the mission and know what equipment was used.
Physically, the tape would have fit on the Burroughs tape drives in the second picture. It might even have been readable, given the right translation software. That form factor was used by a lot of different tape drives and tape formats over the decades. The tapes in the third picture are surely the same form factor.
The computer in the second picture is the Burroughs B-5500 that was in the Basement of Hermann Brown Hall in the early years that I worked there. I probably should recognize the gents in the picture, but I don’t; maybe hardware repair specialists.
The earliest tapes at Rice that I remember were on the Burroughs machine in 1968 when the Rice Computer Center was established. Previously I had seen tapes that were recorded at 500 Bits Per Inch and 8 or 9 tracks wide. 800 was a common format. 1600 followed and then 6400. I still have a number of tapes that I got while at Rice, though I left for the Institute for Defense Analyses in Princeton, NH in 1979 – 26 years ago and I doubt that they are readable. Fortunately I copied the files to other computers and have CD versions of everything. For each new media as soon as it is standard and inexpensive, one should copy over anything that is to be part of a permanent archive.
When I worked in the library (late 1960s), the 1401 had 7-track tape drives that recorded at 556BPI or 200(?)BPI. The reels were already the same shape as shown above, except that there were usually perforations so you could hold the end of the tape down with your finger until the second or third layer held it in place.
7-track tapes were early technology. They stored 6-bit data with a 1-bit parity track. They worked great for BCD, but were obsolete when the IBM 360 went to an 8-bit byte. The 9-track tapes store eight bits of data with a parity track.
I’m guessing that there is an EAD format descriptor for this, but I can’t find it in my first few searches.
7-track tapes recorded octal data and 9-track tapes recorded hexadecimal. Univac used octal long after everyone else had switched to hex.
Ed, 1979 was 36 years ago! How time flies!
Before the B5500, there was of course the R1, but its reels looked clearly differently. I also recall an IBM 7040(?) before the B5500, and I think it had tape drives. That doesn’t change any of the conclusions others are drawing.
Yes, the 7040 had tape drives, and the reels would have looked much the same.
9-track mag tape drives were made by many manufacturers. They were the universal interchange storage at the time. Scroll down to “other drive manufacturers” on this page to see a list:
Here is a vacuum-column drive in operation, they were noisy.
There were still a lot of 7-track tapes on campus in the ’70s. Space Physics had a computer room on (I think) the west end of the second floor of their building. And there were several 7-track drives on the various systems.
When ICSA replaced the B-5500 in the second picture with an IBM 370/155 in 1972, it had one 7-track drive (2401?) for backward compatibility with the rest 9-track (3420).
Open reel tapes eventually gave way to cartridge tapes that held more data and took less storage space.
On some of the older tapes, the density of the data was so low that you could spray “tape developer” on the tape (essentially iron powder in a carrier liquid) and visually read the data off of the tape. Yes, Spac’s computer room was on the west end. If I remember correctly, there were a couple of SDS Sigma 6 computers in there. In some sense, computer tape drives were just a fancier version of the reel-to-reel audio tape recorders that were common when I grew up, although using vacuum columns to provide slack on each side of the read/write head(s) instead of capstans.
Correct me if I’m wrong but I think most of the data on the tape came from punch cards. I know most of the time you had to have a punch card to access the tapes to get information off off the tapes. I think there was another way of storing information at this time and that was a multilayer disc disc that cane into play somewhere during this time. Maybe someone can elaborate a little more on these days before the PC’s and Microsoft.
Not necessarily. Initial input might come from cards but mag tapes were the medium used to store data before disk drives became common. When I started at Baylor our financial/HR and clinic billing was run on a Honeywell 120. This was a totally tape driven system. The OS was on tape, all of the data was on tape and most of the process was
1) read input tape(s) [we did data entry directly onto tape]
3) write updated information to output tape(s)
4) write reports/control totals to printer
Our tapes were 7track 200BPI. We had to maintain a 7track tape on the Dec10 for years to read archival Honeywell tapes
almost all of the tape drives used a vacuum column as Richard observed. Some were very short (for the desk top) and some were 4 feet (such as the 3420). (It was loads of fun when a drive died and you had to manually rewind a tape by spinning the supply capstan.
Almost all software was released on tapes. all patches (except for very small patch decks) were released on tape. Some data centers might have thousands of tape reels. Tape management systems were a major application.
Also, the ability to write to a tape was by the presence of a removable write ring (plastic ring) just outside of the backside of the hub. Without the write ring a tape could not be written on. Most organizations would remove the rings unless the tape was about to be written to
I think it is likely that the data on that tape came directly from telemetry equipment, despite the fact that the label mentions 80-byte records (the length of most punch cards).
Not that this is strictly germane, but audio archivists run into trouble trying to play old reel-to-reel tapes all the time. Ampex, a commonly used brand of audio tape in the ’70s and ’80s, are notorious for separating their oxide from the Mylar strip, or having the oxide adhesive seep through and glue two sections of oxide together. When converting the tapes to modern formats (sometimes a challenge because of the difficulty of finding well-functioning tape players) it is common to bake the reels in an oven at about 200 degrees F for a bit to “reset” the adhesive. If you are lucky you will get one usable pass. I remember transferring a reel tape once where it was raining oxide as it passed the playback head and we had to keep lubricating the head with oil so the tape wouldn’t stick.
Back in the 1980s some of the oil companies kept backup 9-track data tapes in salt mines for the low humidity and constant temperatures. Unfortunately salt has a way to get into anything. Often when the tapes were retrieved and read small crystals of salt would be on the tape. The number of read heads that needed replacing on the drives went through the roof. They soon found better places to store the tapes.
Writing data directly to tape was common by 1970 for high-volume data generators like seismic exploration, telemetry, and such. At Shell Oil, analog acoustic data from geophones was captured at the exploration site on tapes about 1-meter in length and about 4-inches wide. The analog data was replayed in the data center and digitized to 9-track tape for processing.
Applications that captured or generated large amounts of data wrote to tapes because online disk storage was too small in size and to expensive to capture and permanently store large quantities of data. The commonly used disk drive on IBM mainframes in 1972 only held 100 MEGAbytes (although the media could be swapped). That amount of data could fit on a couple of 9-track 1600-bpi tapes.
This mag tape is hanging on my wall at Chegg. You can read the labels, but it is a BCPL distribution tagged by two Rice grads, Tw Cook and Kevin Pei. The fragment of fabric in the upper right corner is a Rice pennant. My desk also sports an HP46 calculator and a 20″ slide rule.
My first job out of Rice in 1982 was at Ampex. The oxide “tape developer” liquid was first developed to help one splice 2″ videotapes. You wanted to make your cut between tracks which were almost perpedicular to the long side of the tape. You would put developer on both pieces of source tape, then try your best to cut between the tracks, then splice the tapes together. This was to avoid the “jump” in the playback when only a portion of a track was played back.
In the seventies with the advent of helical scan videotape recording, it was impossible to physically cut a tape between tracks, which were now almost horizontal with the long tape edge. The videotape equipment manufacturers started building simple electronic editing features into their videotape machines to do simple insert and assemble edits.
A personal anecdote about peeling tape oxide: When I was at Ampex, I once had to run an overnight test of a videotape recorder in a hot/cold box (to verify proper operation through the whole stated temperature range). My test played the whole tape, then rewound back to the beginning and started over. (I was testing operation of the control electronics, not the tape handling mechanics). I started the test at the end of the day to run overnight. Sometime during the night, the cooling in the hot box failed. I came in the next morning and the interior of the box was like an oven. The tape machine was still running and functional. However, the 1-inch videotape had gotten so hot that the magnetic surface had lost adhesion with the mylar backer, and there was loose tape oxide flopping all over the place as the machine dutifully tried to read the “tape” from end to end and then rewind.
Some time in the early 90s I was working for the Environmental Science & Engineering department and for one of our projects we had a reel tape (called a half-inch tape) from which we needed to get data. We ended up taking it to a local IBM office (if I recall correctly, though maybe it was Sun Microsystems) which had a half-inch reel drive that they let us use. That was when I learned the UNIX ‘dd’ command.
I Know its been some time since your first post of the Data reels, or mag tapes, but I have one of the machines that your tapes go onto, I also want to sell it, i’ve currently got it listed on Gumtree so if you search for “digi-data corporation” which is who made it, it should bring up the listing. this machine also have a 80 pin scsi interface but i’m unsure of what computer it would connect to.