SDS MAGPAK and the RCA Snap-Load tape cartridge

Summary

In July 2019 I obtained a SDS MAGPAK tape cartridge. SDS is Scientific Data Systems; the MAGPAK data cartridge was used on the SDS 900 series of minicomputers. The cartridge form-factor, turned out to be based on the RCA "Sound Tape Cartridge" or "Snap-Load Cartridge" format. The history of that RCA product is described on another WEb page. I purchased some RCA carts as well. It appears that the recording scheme used by RCA for audio stereo, was incorporated by SDS into digital storage in some fashion. I'm determining the methods used, results on this Web page. In Sept 2019 I made some progress on reading some the SDS cart tape, as describe below. Then I acquired a RCA cartridge player and repaired it, as reported on the linked Web page. Recovery efforts of the SDS data using that player will be described here.

The issues of analog audio recording versus digital data recording, are not limited to this ancient cartridge format. In the later Phillips cassette world, digital cassette decks and tapes were also developed. Here's a Web page about an MFE 250 digital cassette drive and tape system, used on a MCM/800 microcomputer of the early 1970's. - Herb Johnson

SDS MAGPAK

The tape is 1/4" wide, the same size as open-reel audio magnetic tape. The cart measures 5 by 7-1/8 inches. The reels of the cart are kept from unravelling in storage, by a braking mechanism. Centered in the cart is a spring-loaded arm with teeth, to engage the edge of each hub.Text molded into the SDS cart contains two patent numbers. I discuss these detail on my Web pages on the history of the RCA tape cartridge, and repair of a tape-cartridge player/recorder.

SDS Sigma 900

According to an Scientific Data Systems brochure, archived by archive.org from Ted Nelson's archive of junk postal mail:

Magpak is a magnetic tape system developed by Scientific Data Systems specifically to fit SDS 900 Series software. It takes seven inconvenient, time consuming, error-causing steps out of small computer operation. With Magpak, budget-limited computer users can have large computer convenience and efficiency at half the cost."

The SDS 900 were a series of minicomputers produced by SDS in the 1970's. Prices started at "$81,000 ($2,350 per month on lease)." SDS was headquartered in Santa Monica, Calif.

"Magpak consists of two independently controlled magnetic tape drives mounted on a 2" by 19" panel that fits any standard SDS 900 Series Computer rack. Each tape drive holds a self-contained, dual track, magnetic tape cartridge with a total capacity of more than 4 million characters.

How is data recorded on a tape?

Computers have used magnetic tape storage for decades. Mainframes and minicomputers used 7-track and 9-track 1/2 open-reel tapes since the 1950's. Minicomputers and later microprocessor-based computers, used a variety of Phillips tape-cartridge methods. Essentially, binary signals are either recorded like floppy-disk signals using an "frequency modulated" scheme; or they are recorded as pulses of one of two time periods (frequencies). For floppy disk methods, check my Web page about floppy diskettes. For Phillips-cartridge pulse-width methods, check my Web page about microcomputer data cassettes.

Playback of SDS data, as audio

Content of the SDS cartridge: open-reel tape recorder

On the assumption that the Sigma 900 cartridge recorder used the same stereo magnetic head as the RCA cartridge - and that is compatible with 1/4-inch open-reel stereo magnetic heads - On Sept 2nd 2019 I attempted to read some of the SDS cartridge tape. For expediency, I simply unwound some of the tape, and fed it "loose" to a Wollensak-3M 1/4-inch stereo tape recorder.

First, I had to physical restore the recorder to operation. It needed lubrication, head-cleaning, new head-pads (they press the tape against the heads). I made numerous adjustments and operations to the mechanism to restore it to function. But by my ear it plays prerecorded tapes slightly slow, likely due to a worn transfer-wheel to the capstan. It's also possibly there's damage to either channel's playback for other mechanical or electrical reasons.

I also set up a Dell laptop with the audio app Audacity, to digitize the stereo tracks. It took a few runs to set up the audio controls for recording without clipping. The tone controls per channel were set to high treble. Experience with audio-cassette (Phillips) 1970's microcomputer data storage suggests that setting.

Observation of the Sep 2 results suggested a sampling rate of 44.1KHz, to obtain several samples of the highest-frequency apparent waveform. (My first recorded session was sampled at 22.05KHz, the second at 44.1KHz). On the imaged waveform, each "dot" represents a digital sample. A 16-bit PCM uncompressed WAV file was saved for each of two sessions of the same portion of the A-side tape. The files are a few megabytes each and run for about 30 seconds each.

At first glance, it appears there's two independent recordings, one for each channel. Each channel shows repeated packets of signals of the same length (.2 seconds) and seperation in time (.25 seconds). The two channels don't appear to be otherwise correlated, and both seem to show similar waveforms. On the second run, one channel shows a repeated dip in volume at various times for each packet; the other does not. Therefore I suspect it's an artifact - for instance, iron-oxide contamination on one channel's head. I'll present further information on the audio patterns as I use Audiology to filter and inspect the results.

Results confirmed: the SDS digital system used the RCA stereo-track scheme, with data stored on one channel at a time, as packets or blocks of encoded data, within the audio range.

While I can perform some analysis by looking at the responses, what I'm seeing is different responses from each channel. This could be due to manual tone settings; junk in each head's gap; or issues with each channel's amplifier. Also: I'm not providing a "test tape", I'm displaying results of the data-as-audio. I'll describe what I see, but I may be mistaken.

Audiology provides a "spectrum" display of a selected part of collected audio. The spectrum shown on the left, is of a segment of over 20 blocks of data from one channel. The result suggest the Wollensak is responding up to several kilohertz; the data as audio falls off rapidly at about 5KHz. While this is not a bandwidth test - the data is a limited number of frequencies - it's generally informative. Examination of individual blocks by spectrum, shows larger peak responses at these frequencies: 334-341Hz, .991-1.018kHz, 1.7lKHz, 2.3KHz, and then dropping off after 4.5KHz to 5.0KHz.

The photo above of the individual waveforms, under manual examination, suggests there's two frequencies of operation, simply by counting samples (dots) from cycle to cycle for wavelength. Fast waveforms are between 4.4Khz and 4.9KHz; slow waveforms are about 2.3KHz. However: the Wollensak may be running slight slow.

Content: RCA cartridge player

By Sept 22 2019, I obtained and repaired a RCA monophonic tape cartridge player, so I could more conveniently play-back the SDS tape and other such tapes. Mechanical repairs were completed and the RCA player was able to play back RCA tapes. When I mounted the SDS tape, it was also able to replay each of the tracks on it, just as the Wollensak open-reel recorder did.

The playback of the SDS cart on the RCA player (Sep 26), seems to have about the same "fidelity" or high-frequency response as the same cart on the Wollensack tape recorder. Audacity's spectrum analysis across several blocks of SDS data, suggests a fairly flat response up to a few KHz, then a falloff from -36dB at 3KHz to -66dB at 9KHz; about 10db per octave. The SDS tape on the (faster) RCA player, shows peak frequencies (by eye) 376-400Hz, 1.1-1.2KHs, a big peak at about 1.9K-2.0KHz, 2.5-2.7KHz, 3.4-3.5KHz; then as response falls off, peaks at 4.2Khz, 5.0-5.1KHz, 5.8KHz and lesser peaks with higher frequency (and less response).

However, there's various artifact that don't look right. I see what looks like dropouts - short bursts of reduced signals. Those appear to be intermittant and less-often when I clean the heads, hence that diagnosis. Also, the ratio of amplitude between the lower and higher frequency waveforms, seems to vary between the two tracks and two sides of the tape. Also, different tracks have different noise levels between the blocks of data - "noise" as in a broad spectrum from 300Hz to several KHz.

In addition: I had trouble with "tape squeal"the physical tape cartridge as it was played. Eventually I determined, it was the tape rubbing the read/write head and causing either the head or a segment of tape to physically vibrate. Proof - a piece of paper between the head and the tape, stopped the physical sound. Of course that squealy probably created an electrical signal as well. Solution? I guess some kind of "tape lubricant". Historically, audio tapes over decades dry out: now I know what that sounds like. ;)

Tentative conclusion: somehow I need to coax more high-end response out of the RCA cartridge player. It may play audio tapes better too, but I need to extract higher-frequency "data" from the SDS tapes, than it appears to be producing. Some combinations of repairs and lubrication may be necessary. I'll report on those, on my Web page on repairs of the RCA 1YB-11 cartridge player.

Oct 16 2019: The cartridge repair Web page (link in last paragraph) describes my repairs and results. The only player-repair I did, was to adjust the azimuth (angle) of the read/write head, so it was properly perpendicular to the recorded track. It was sufficiently mis-aligned, relative to the SDS tape and an RCA prerecorded audio tape, that the difference was noticable to both ear and to a display of the audio signal. With that repair done, the cartridge player played an RCA musical recording with reasonable fidelity.

Engineering of the SDS cartridge tape

Thanks to bitsavers.org and Al Kossow, I found some SDS documents on the Sigma 900 series and their Magpak system. I'll write up the specifications for data recording later. Note: my audio playbacks were done before I read the specifications. We'll see if what I saw, has any correlation to means of recording binary data.

RCA "Snap-Load" Sound Tape cartridge and player

Wikipedia's entry on the RCA cartridge includes the following. My remarks are added in []'s.

"The RCA [Sound Tape brand] tape cartridge is a magnetic tape audio format that was designed to offer stereo quarter-inch reel-to-reel tape recording quality in a convenient format for the consumer market. It was introduced in 1958, following four years of development. This timing coincided with the launch of the stereophonic phonograph record. It was introduced to the market by RCA in 1958."

Wikipedia says the 1/4-inch tape runs at 3.75 inches/second and has four discrete sound-tracks; two per side as with open-reel 1/4-inch stereo tapes. Recording and playback at 1.875 IPS version was also supported. The RCA cartridge measures 5 X 7-1/4 inches (same dimensions as the SDS MAGPAK) and has the same hub-braking mechanism, and the same holes and other physical features. I do not know at this time, if SDS made arrangments with RCA for use of the form-factor.

RCA and other companies produced some number and volume of prerecorded tapes. I discuss RCA and other's prerecorded cartridges on my Web page about the history of these RCA products.

RCA's cartridge players

In late Sept 2019, I acquired a nonworking RCA cartridge player, a model 1YB-11 mono player, A serial number in the 2000's. It's a vacuum tube model, apparently the less expensive model with only two vacuum tubes, a neon-lamp volume display, and a selenium rectifier for the power supply. A view of the tape heads shows a stereo head with two black bars representing the two magnetic pickups. There's an A/B switch on the player's deck to select one of these for the one-channel "monophonic" amplifier and speaker. I completed repairs of the player; See this Web page about my repairs of the 1YB-11 RCA cartridge player.

Herb Johnson
New Jersey, USA
to email me @ follow this for email and business