Author Topic: Digitizing NASA data tapes  (Read 19409 times)

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Digitizing NASA data tapes
« on: February 12, 2018, 03:20:59 PM »
I am in contact with someone who will help me read NASA data tapes from the 60s to 70s

Some quick looks at the tapes showed Lunar Orbiter, Apollo, Skylab and sattelite 1/2 inch computer tapes.

They Will be uploaded to archive.org as public domain.

I will post findings and relevant parts of our conversation here.

Some tapes cost 80 dollar to acquire, if you want to help finance this you can email [email protected]

Regards,
Apollo16uvc
« Last Edit: February 12, 2018, 03:25:43 PM by apollo16uvc »
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #1 on: February 12, 2018, 03:21:23 PM »
Me: "
Today I played part of the tape through an AKAI X201D. I did not expect to get any usable data, maybe a faint signal.

You can find WAV files and spectrograms here: https://drive.google.com/drive/folders/17vss0VXnpmGchf0DuRh3w2wXxNNphkYA?usp=sharing

Test001.wav through Test003.wav were manually moved along the heads. test004.wav and test006.wav was fed along the heads via the capstan at 3 3/4. test005.wav is played at 7 1/2 inches per second.

Test005.wav has a continuous tone.
test006.wav is the very beginning of the tape, and has weird bursts at the beginning."

Contact: "Having done a number of these for JPL, I'd say that this is a 7-track 800 BPI NRZI tape. Depending on the exact tape drive, it can probably be read on a 729-equipped 1401. Making sense of the data, however, is another story. Early 1960s NASA tapes tended to be 7090/94 ones, with later ones, Univac 1100 series right up through the early 1980s.

Or you can send it to me and I'll extract the data. My contacts at JPL say there are thousands of these things kicking around."

Me: "Thank you very much for your reply.

Can you any sense out of the sound recordings I have made? what are the short bursts on test006 and the tone on test005.


I would be incredible thankful if you could extract the data. How to best ship a tape? I have to send mine from the netherlands to your place, which I assume is America. Any further tapes I get I might have shipped directly to you from a seller in America. Whatever you want.

I know a lot of people will be excited to look at the recovered data.

Is it possible you can try and get the tape identified by the people at JPL? I did scrub off the upper label to reveal the lower label with some more information.."

Contact:"
No, I do not think that I could do anything with your audio. Perhaps an examination with a magnetic developer (such as Kyread) and a low power microscope would verify my suppositions. That would at least tell you how many tracks are involved and define the inter-block spacing, if applicable.

There's a faint possibility that these might also be analog telemetry tapes, but a view with a magnetic developer would verify that. However, what I can see of the partially-covered label seems to indicate a standard digital tape, but the note on the label that says "audio" is puzzling. Were that the case, you should be able to get something using an audio recorder to play these back.

I was under the impression that tapes used for analog telemetry back in the 60s was very different from standard 10.5" 1/2" tape, but you never know.

They're definitely digital tapes if you have blocks of data separated by approximately 3/4" of erased space. That's standard 7 track format.

Finally, a lot of these tapes are simple recertified "scratch tapes"--that is, used tapes returned to the scratch pool and run through a certifier (which erases all data). A tape label that specifies contents and the name of the programmer holds out the best hope, but even so, in the lots I've handled, about 10% were recertified scratch. NASA, like a lot of government operations, re-used tapes heavily, trimming off tape at the beginning when it wore out and applying a new BOT marker. Some of the tapes that I've worked on have several hundred feed sacrificed.

A lot of this stuff is mixed-format data (some text, a lot of binary) data that may be very difficult to suss out without the program that created it. For example, here's one of the tapes that I did: Mostly 7094 floating-point data with a few bits of 7090 BCDIC mixed in."

Me:"Really cool that you worked on NASA tapes before, how did that come to be? and how do the tapes usually end up in the 'wild'?

There are many unknowns, but with small steps I believe we can figure this out.

I looked around for Kyread, and buying from their website the shipping cost alone is 150 dollars. I can buy magnetic viewing film much cheaper, do you think it has enough resolution to view the tracks?

If it does have analog telemetry, is there a special type of audio recorder we need to play them on? or will any 7 track multi track recorder do? I do not know if a 1/2 inch 7 track audio recorder exist, 1/2 inch 8 track tape recorders were more common. Today I spooled the tape through my 1/4 inch 4 track stereo recorder again, this time there were 4 to 5 distinct frequency tones in the spectogram instead of one. Could it be overlapping tracks?

Is it possible rectifying a tape causes this continuous tone?

I may get some NASA tapes soon that have a greater potential to hold data."
*Pictures of tapes for sale*

Contact:"My tapes came directly from NASA JPL in Pasadena. I do not own them--and they were returned after data retrieval, along with the data retrieved from them. In other words, it's JPL's property, not mine and I treat it that way.

Kyread is nothing more than 1 to 3 micron iron particles in a fast-evaporating fluid (it used to be a type of Freon, but since that was banned, methyl perfluoroisobutyl ether is used). It is relatively inert, so it does not affect the coating or base material. You shake the bottle up and drop a drop of the suspension on the tape and allow the solution to evaporate. Since the iron particles are so small, you can visualize very small features in the tape. With a decent chemistry lab, it should not be difficult to mix some of this up in your location, assuming that the carrier liquid is legal in the EU.

When I have a new tape, unless I'm very certain about the content, this is always the first step. Tapes labeled as 9 track often turn out to be 7 track and vice-versa. What are you going to believe, your eyes or some lying label?

The other thing is preparation of the medium. Normally, the procedure is to "bake" the tape, then run it through a tape cleaning machine. Additional treatment may be necessary, such as lubrication, to get the tape to read. After all, you have no idea of the storage conditions during the 50 or so years that the tape has been in storage. It's not uncommon for splicing tape (used to attach leaders) to dry out and let go, so you have to be prepared for that. BOT and EOT markers similarly dry out and fall off...

The tapes you pictured above are certainly data tapes and definitely 7-track (9 track tapes don't usually occur with even parity). But you can't get your expectations too high. The Apollo program involved hundreds of subcontractors who generated probably tens of thousands of tapes. The GE tapes you show above may be nothing more than engine test data--and like such things, you need a Rosetta stone to interpret it, since it's unlikely that the data will contain any clues. Before the days of high-speed telecom and big disk drives, tape was king. Outfits like NASA ordered tapes by the truckload.

But, if you're willing to take a shot, I'm willing to have a go at it. I'll PM you my shipping address."

Our conversation continued in private and got quite technical.

I will try to get the tracks visualized on the tape currently with me and am looking into getting it played on a multitrack 1/2 recorder, if it is analog telemetry we might get something.

As for the data tapes, my contact will handle those.

Regards,
Apollo16uvc
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #2 on: February 12, 2018, 03:22:23 PM »
"You can see an HP 7970 tape drive demonstrated at about 7:11 in this video:



Mine is a 7970B, modified to use a 7-track head instead of the standard 9 track one. The interface is "raw", which feeds into an STM32F407 microcontroller, which writes to a micro SD drive. While I could probably write 7 track tapes, mine is configured as read-only.

For 9 track tapes, I have several drives, but the one that I used most is a Fujitsu M2444 drive. It's impossible to see it working, as the door must be closed for it to operate. That is fed by a 16-bit ISA controller in a PC. The whole affair is rack-mounted."

The first tape has been send: https://www.ebay.com/itm/332544367534?ViewItem=&item=332544367534

As for the "Audio tape" I will be getting someone locally to play it with a multi track audio recorder. It is this tape: https://www.ebay.com/itm/Vintage-NASA-Audio-Data-Tape-Reel-From-The-1960-039-s-Who-Knows-What-May-Be-On-This-/332455504748
« Last Edit: February 12, 2018, 03:24:45 PM by apollo16uvc »
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #3 on: February 17, 2018, 02:40:00 PM »
Tape: NASA Automatic Data Processing 4141 has arrived.

Update written by contact:

"Niels,
 
 The tape arrived today--the photos don't do it justice(?).    The reel is chipped and absolutely filthy--it looks as if it spent the past 50 years on the floor of someone's garden shed.
 
 At this point, it's difficult to tell how much of the tape is intact.   This is my plan:
 
 1.  Bake the tape for a day or two
 2.  Run it through the cleaning machine a couple of times to remove as much of the dirt as possible
 3.  Clean the reel and re-spool the tape
 4.  Try to see if there's anything readable
 
 How does this sound?"

"Well, I unreeled a bit of the tape; I didn't find a BOT marker, so I suspected that the first few meters of the tape are history
 
 Here's what I see when I look at the tape:
 
 9-track, not 7!  If the date on the tape is correct, this means it was produced on an IBM System/360, as almost nobody else had working 9 track drives then.    Even in 1965, IBM was shipping the first S/360s out with 729 7-track drives.   It took them awhile to get the bugs out of the 2400 series drives.
 
 Testing the very beginning of this very wrinkled tape shows data on the first few cm, so if there was a leader, it's been long gone.
 
 I'll bake and clean the tape and attach a BOT marker about 3 m in and see what I can see with my 9-track drives.  This may take a few days."
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #4 on: February 18, 2018, 07:36:10 AM »
The tracks on the tape visualized with a magnetic viewing solution:

Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #5 on: February 25, 2018, 06:13:08 AM »
A even greater magnification closeup of the tracks:
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #6 on: March 03, 2018, 07:30:42 PM »
The NASA ADP tape may not be readable with conventional computer tape drives, as the ADP contains telemetry and not data. We may figure out a way to digitize it eventually.

We have had a succes with two other tapes though! Switch-Action TP 1820 and 2909 have been read and successfully converted to .tap and ASCII text. I have had a look and think it is absolutely fascinating.

When I have the time and some more info I will make a public archive and put the files there.

Cant wait to share this with you!

Regards,
Niels
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #7 on: March 04, 2018, 05:30:22 AM »
The tapes with their files can be found here:

https://archive.org/details/SpaceData



All thanks go to Chuck for reading the 7-track tapes for me. This would not have been possible without his help. More tape recovery may come from him later so stay tuned.
« Last Edit: March 04, 2018, 06:02:08 AM by apollo16uvc »
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline ajv

  • Venus
  • **
  • Posts: 52
Re: Digitizing NASA data tapes
« Reply #8 on: March 04, 2018, 01:43:58 PM »
12 12 L DI 0276  SIC LOX BUBBLING DUCTS 1 + 3  M DO 0276 ON

Cool!

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #9 on: March 17, 2018, 01:29:59 PM »
Two more tapes have arrived this week, a quick update below:

"The second set of tapes arrived in good form. At first blush, they seem to be more switch panel data, but we'll know more later.

Unlike the last batch that were equipped with poly "tape hanger" strips, these were in 60s-era 3M hard styrene cases. That was good news and bad news. The good news is that the cases use a weatherstrip-type rubber foam to seal the edges of the two case halves. The bad news is that the foam has gone the way of all old rubber and deteriorated, essentially gluing the case halves together. With some coaxing and hard pulling, however, they came apart. I removed the bad foam so that this won't be a problem in the future.

I baked and cleaned the tapes (1179 was a bit sticky) and did a quick run-through of them.

2090 reads perfectly; no issues--the data looks much like the previous tapes; 132-character records of what I'm supposing are switch legends. Only one single parity error (I can isolate it to a single byte). 14,404 records amounting to about 2 MB.

1179 starts off the same way, but after about 663 blocks, everything goes to hell. Parity errors, short blocks, incomplete records. The only way I can think of resurrecting anything is to simply read without retrying all the data I can, and then trying to make some sense of the result. My suspicion is that the tape was stored (for years) close to a magnetic field and so became partially erased. Physically, the tape is in fine condition; not shedding or displaying any other physical damage."
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #10 on: March 18, 2018, 04:20:50 PM »
Uploaded 2090
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #11 on: March 20, 2018, 12:32:25 PM »
We are currently brainstorming on the best way to digitize 1179. When processing the BCDIC data to ASCII we get mostly garbage and the occasional readable string.

The best way might be to read it without any retries and try and make something out of it, this may requirements a firmware hack of the drive. The .TAP format does support forward and backwards retries but it reads the same errors each time.

Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #12 on: July 08, 2018, 07:38:18 AM »
Working on four other tapes.

I'm starting here with the tape QK7992H from NASA in the lot of 4. I suspect that it's the best as it's probably the original.



The SIMH file can be downloaded here: https://archive.org/download/SpaceData/QK7992H.zip

At any rate, count this as a test only--there appears to be only one uncorrected parity error in block 130.

Note that any text in these is EBCDIC, not ASCII. The data structure appears to be fixed length records of 83 (total) bytes, each record starting off with its number in EBCDIC. Initial records are entirely in text; subsequent records are binary data.

For example, the first physical block (translated from EBCDIC to ASCII) is:

Code: [Select]

20 20 20 20 |..Block 0:.. |
00000010 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 | |
00000020 20 20 20 20 30 30 30 31 30 30 30 31 30 37 36 38 | 000100010768|
00000030 30 37 36 38 20 20 20 31 20 32 20 32 30 32 30 20 |0768 1 2 2020 |
00000040 31 35 33 30 20 20 20 20 20 20 20 20 20 20 20 20 |1530 |
00000050 31 20 20 43 20 20 20 20 20 20 2a 20 20 20 20 20 |1 C * |
00000060 20 2a 20 20 20 20 20 20 2a 20 20 20 2a 20 20 20 | * * * |
00000070 2a 20 20 20 37 32 30 2a 20 20 20 20 20 20 2a 20 |* 720* * |
00000080 20 20 2a 20 20 2a 20 2a 20 2a 20 2a 20 2a 20 2a | * * * * * * *|
00000090 20 20 20 20 20 2a 20 20 32 20 20 43 4c 57 52 31 | * 2 CLWR1|
000000a0 35 33 30 20 20 20 56 45 53 54 41 2c 31 32 4d 49 |530 VESTA,12MI|
000000b0 4e 2c 53 4d 2e 41 50 2e 2c 4c 4f 2d 44 49 53 50 |N,SM.AP.,LO-DISP|
000000c0 20 20 5a 45 4c 4c 4e 45 52 2f 4d 41 59 20 32 31 | ZELLNER/MAY 21|
000000d0 2c 31 39 37 38 20 20 20 20 20 20 20 20 20 20 20 |,1978 |
000000e0 33 20 20 43 43 41 4d 20 32 20 49 4d 41 47 45 20 |3 CCAM 2 IMAGE |
000000f0 31 35 33 30 20 20 20 20 20 20 20 20 50 52 4f 43 |1530 PROC|
00000100 20 56 45 52 53 4e 20 32 31 31 20 50 41 52 41 4d | VERSN 211 PARAM|
00000110 20 53 45 54 20 4e 4f 20 38 20 20 20 20 20 20 20 | SET NO 8 |
00000120 20 20 20 20 20 20 20 20 34 20 20 43 20 20 20 20 | 4 C |
00000130 20 20 20 4e 4f 20 4c 41 4d 50 2d 45 58 50 20 54 | NO LAMP-EXP T|
00000140 49 4d 45 3d 37 31 39 20 53 45 43 53 20 20 20 20 |IME=719 SECS |
00000150 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 | |
*
00000170 35 20 20 43 |5 C

I think you should read the text like this:

530 Vesta, 12 MIN, SM.AP., LO-DISP

ZELLNER / MAY 21 1978 3

CCAM 2 IMAGE 1530 PROC VERSN 211

PARAM SET NO 8 4 C

NO LAMP-EXP TIME = 719 SECS


The "ZELLNER" and "VESTA" surprised me, as neither Pioneer 10 nor Pioneer 11 did any imaging in the asteroid belt (Vesta is a big asteriod).

A little Googling makes this clear:
https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19780021064.pdf

"Dr. Benjamin H. Zellner
Lunar and Planetary Laboratory University of Arizona
Tucson, Arizona 85721 "

"B. ZELLNER. Geography of the Asteroid Belt ..................... 99"

Benjamin H. Zellner is an astronomer who was in the UofArizona at that time. His specialty was asteroids. As far as I know he still lives - hopefully he can tell us what's on the tape?

In the document he talks about UBV, in my opinion color or intensity data of asteriodes observed with an observatory. This explains the exp time of 719 seconds. Can CCAM stand for a CCD sensor?

How is that for an imitation of Sherlock Holmes? Give me a role rusty mylar and I find the owner.
Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon

Offline ka9q

  • Neptune
  • ****
  • Posts: 3014
Re: Digitizing NASA data tapes
« Reply #13 on: July 09, 2018, 01:06:34 AM »
Keep up the good work!!

Offline apollo16uvc

  • Mars
  • ***
  • Posts: 396
  • Where no telescope has gone before.
    • Patreon
Re: Digitizing NASA data tapes
« Reply #14 on: August 04, 2018, 12:25:46 PM »
The last Switch Action tape that was having tons of reading problems has been digitized!

"I had another go at 1179 (new code, re-calibratomg read ampliffiers).
Much better results this time, though not absolutely perfect, but pretty
remarkable for a 46 year old tape"

Switch Action Tape 1179, rocket identification SA-511, Apollo 16

 




De band in de drive:


Info, SIMH bestand en ASCII omzetting: https://archive.org/download/SpaceData/SwitchActionTp1179.rar

The data is comparable with other Switch Action tapes, with some interesting things like it mentioning tthe UI and other specific parts of the rocket. Here is a sample:

Code: [Select]

R/L 40M17360-11   E/O 8S-0411    REV K    D/I 10/11/71   ECP 10-3218   E                                                           

Block 10:

R/L 40M17360-11   E/O 8S-0412    REV L    D/I 10/27/71   ECP 10-3216   E                                                           

Block 12:

R/L 40M17360-11   E/O 8S-0413    REV M    D/I 10/28/71   ECP 10-3206   E   511 FRT-1                                               

Tape mark at 13

S     02 02 L DI 0026  EDV SII INT HY STBY HY CHG C  M DO 0026 ON  C A                                                             

Block 53:

S     02 03 L DI 0027  EDV SII INT HY STBY HY CHG O  M DO 0027 ON  C A                                                             

Block 54:

S     02 04 L DI 0028  EDV IVB AFT SLINE WD SYCHG C  M DO 0028 ON  C A                                                             

Block 56:

S     02 05 L DI 0029  EDV IVB AFT SLINE WD SYCHG O  M DO 0029 ON  C A                                                             

Block 57:

S     02 06 L DI 0030  EDV IVB AFT HY WD GN SYCHG O  M DO 0030 ON  C A           

Block 674:

S     19 21 L DI 0453  INT AUDIO SELECT POS NO. 1    M DO 0453 ON  C B                                                             

Block 676:

S     19 22 L DI 0454  INT AUDIO SELECT POS NO. 2    M DO 0454 ON  C B                                                             

Block 677:

S     19 23 L DI 0455  INT AUDIO SELECT POS NO. 3    M DO 0455 ON  C B                                                             

Watch me at: YouTube
Experience the past: Flickr
Support me on Patreon