Magsat Corporation was an American manufacturer based in Harford, Connecticut. They are known for two distinct product lines: Magsat keyswitches, and telecommunications devices for the deaf. Magsat was founded on the 22nd of October 1969 as Satellite Switch, Inc., suggesting that “magsat” is short for “magnetic satellite” (seemingly confirmed in Electronic Design’s Focus on Keyboards article from November 1972). In May 1971, they changed their name to Magsat Corporation. The firm was dissolved in December 1978; the month before, it reverted back to its original name of Satellite Switch, Inc. The MAGSAT TDD/TTY page on the Southwest Museum Of Engineering, Communications and Computation (SMECC) suggests that Magsat transformed into or was absorbed by American Communication Corporation. American Communication Corporation, also of Connecticut, was founded on the 18th of September 1978, around three months before Satellite Switch closed down. An American Communication advertisement shown on the SMECC website uses the same distinctive typeface as used by Magsat for their logo, suggesting that American Communication was based around Magsat (Satellite Switch). Another advertisement, placed in the Deaf American in 1983, notes that “American Communication includes: Amcom, Micon, Magsat, Teletyrm”. American Communication itself was forfeited in 1994; a new American Communication Corporation sprung up the next year, which seems to have become defunct around 2008.
In September 1970, inventor Lloyd J Lapointe filed US patent 3739310 “Electrical switch”; later patents would all be filed by Magsat Corp. These switches employ a novel means of pulse output generation, using a free-moving shuttle containing a magnet, driven by a second magnet in the plunger. This shuttle provides a means to control the switch contacts and provide click and tactile feedback. The term “flying magnet” may not be an official term; it is the name given to the switch in the Focus on Keyboards article in Electronic Design in November 1972, and it serves as a way to differentiate this switch type from the other Magsat patents. The article also calls the shuttle a “magnetic satellite”, seemingly confirming what the company name “Magsat” means.
The magnet in the plunger repels the magnet in the shuttle, driving it into the upper or lower extent of its travel as the plunger is pressed and released. The plunger and shuttle are both responsible for working the switch contacts during a each actuation cycle. This “flying magnet” arrangement is similar to that used in Omron B2H-F7W, although the Omron switch is Hall effect and uses the magnetic repulsion process simply to generate the click and tactile feedback, without the need of a separate shuttle as there are no contacts to operate.
A standard Magsat “flying magnet” switch contains at least two sets of contacts. Each contact set contains two contact pairs, with one pair at the top of the switch and one pair at the bottom of the switch; the top contact also provides the terminal for that contact set. Between each top and bottom movable contact is another metal pressing forming a double-ended stationary contact. The use of two pairs of contacts forms part of the pulse output process, with a special sequence employed to allow the switch to only conduct current for a brief fraction of a second.
In the standard design, all the bottom movable contacts are formed from a single pressing, such that all the bottom contacts are bussed together. The centre contact set provides the common terminal for the remaining contact sets, which provide the output terminals. This centre contact set is arranged to close last so that all other contacts are closed at the point that the switch becomes active. The patent depicts a switch shell with capacity for seven contact sets, and indicates that if this capacity were increased to nine, the switch would have self-encoding ability (8-bit output together with the common terminal). As such, a switch with two form A contacts is not double pole, as the two outputs are connected to a common input. The multiple-contact implementation is well-suited for self-encoding (with nine contact sets) or two-of-N encoding (with three contact sets: common and two form A).
The basic construction of the switch, as shown in a cross section through a single set of contacts, is depicted in Lapointe’s patent as follows:
Note that the bottom contact pair has no terminal; as noted, each contact set has only one terminal, either common input, or output.
When a pulse-type switch is pressed, the following sequence takes place:
- The downward motion of the plunger releases all the top movable contacts, allowing them to close
- As the magnets in the plunger and shuttle cross over, the shuttle is fired upwards and the plunger is fired downwards; this provides the tactile feedback
- When the shuttle begins moving upwards, it releases the bottom movable contacts, allowing these to close also; the common terminal contacts close last, and the switch becomes active
- When the shuttle reaches its upper travel limit, it pushes the top movable contacts open again; the time between the bottom contacts closing and the top contacts re-opening defines the pulse duration, and the impact of the shuttle provides the click feedback
The release sequence is slightly different, to ensure that the pulse only occurs during the down stroke:
- After the plunger is released, the shuttle remains at the top, holding the top contacts open
- As the plunger nears its home position, it reaches the point where it itself is able to hold the top contacts open
- When the plunger magnet crosses back over the shuttle magnet, the shuttle is driven downwards; the top contacts are now held open by the plunger, and the shuttle opens the bottom contacts as it reaches its bottom position
The process above can be seen in the sequence of diagrams below, based on the written descriptions within the patent:
(The diagrams render strangely in Edge and will likely do the same in Chrome; Waterfox Classic renders them crisply but loses the arrowheads, while Internet Explorer draws hollow arrowheads.)
The lengths of the top and bottom movable contact blades determines which of the plunger and shuttle engage them, allowing for both pulse and steady output modes, as well as both normally-open and normally-closed forms. The form and output mode are defined entirely by the length of the contact blades: each contact set and thus each terminal can be normally open or normally closed, and pulse or steady, independent of all the other contact sets. The switch can also be adapted to provide an output pulse on both press and release strokes.
The engagement between the shuttle and the switch contacts is claimed to dissipate the energy from the movable contacts, eliminating most of the contact bounce. Adjustment to the way that the contacts are formed and engaged also allows for wiping motion, to avoid contact arcing.
The patent offers an alternative method of generating pulse output: the movable contact strikes a piezoelectric crystal (with a CLEVITE High Generating Sensitivity Unit PZT-5 from Clevite Manufacturing Company indicated to be suitable), and from this an output pulse is obtained.
Magazine advertisements in 1971 list the following switch types as available, together with their prices for orders of 100,000 pieces:
- Single Form A (45¢)
- Double Form A (60¢)
- Single Form B (55¢)
- Single Form C (69¢)
- Single Form A (Tactile) (69¢)
- Single Form C (Tactile) (82¢)
- Strobed 2 Form A (Tactile) (94¢)
- Repeat Key ($1.05)
- Push Push Switch ($1.05)
- Solid State ($1.15)
The implementations for strobed, repeat and alternate action (“push push”) switches are not covered in the patent. There is also no information at present to indicate what “solid state” implementation was chosen.
In 1972 and 1973 (and possibly in later years), Magsat were involved with Stackpole, with the switches and keyboards described as “Stackpole/Magsat”. Magsat did file a keyboard circuitry patent, US patent 3755691 “Keyboard clock with timing circuitry for controlling transfer of data between keyboard and processing apparatus” filed in June 1972, but it may have been the case that Magsat preferred to have Stackpole handle keyboard manufacturing.
Stackpole/Magsat keyboards and switches were said to offer a variety of advantages:
- Tactile and audible feedback, an unusual feature in 1972 when keyboards tended to be linear
- N-key rollover is provided directly in the switch, rather than in an “electronic network”, using a 1 millisecond pulse that allows N-key rollover with static encoding (this is the same pulse output NKRO as used by Micro Switch)
- Static encoding, which “eliminates costly RFI problems and time dependency of scanners”
- No stand-by power consumed by the switches, which appears to be deliberately contrasting Magsat with Micro Switch’s Hall effect keyboards and their high idle current
- The switches are “insensitive to temperature, humidity, and stray magnetic and electrical fields”, with the insensitivity to magnetic and electrical fields also a potential advantage over Hall effect
These switches were introduced rather late, because single-chip matrix scanning encoders arrived around the same time as these switches, and matrix scanning put an end to complicated encoding processes. It is unlikely that we will ever know how many of these switches were sold.
- US patent 3920943 “Electrical switch”, filed March 1974
- US patent 3863182 “Magnetically operated electrical switch”, also filed March 1974
The first of these patents depicts a torsion spring contact design very similar to Stackpole LO-PRO, which may have been the result of the two companies’ involvement with each other.
No Magsat keyboards have been encountered to date.
All material was scanned by Bitsavers unless otherwise noted.
- Magsat keyboard switches advertisement, The Electronic Engineer, Vol. 30 No. 9, September 1971
- Stackpole/Magsat keyboard advertisement, Electronic Design, Vol. 20 No. 19, September 14 1972, page 103
- Stackpole/Magsat keyswitches advertisement, Electronic Design, Vol. 21 No. 2, January 18 1973, page 37