Jump to page content

Control Devices

Contents

History

Not much is known of Control Devices. Control Devices, Inc. was founded on the 16th of October 1969 and was a Massachusetts company, like various other keyboard manufacturers of the day. Three people are named on their sole keyboard patent: Donald C Gove, Barry W Mullins and Edmund G Rousseau. Donald Gove was responsible for the IKOR self-encoding capacitive system, filing patents for it in 1967. The Articles of Organization for Control Devices, Inc. names Donald Gove as one of the three directors as well as the treasurer; it would seem that he left IKOR to co-found Control Devices.

The business is listed as having a date of involuntary dissolution of the 7th of July 1980. The keyboard patent was assigned to Plessey Overseas Limited in April 1982; the circumstances surrounding this acquisition are not known.

CDK series

CDK was Control Devices’ capacitive keyboard implementation. Described in US patent 3660838 “Multiple point switching apparatus” (filed in April 1970), the design as patented uses a capacitor-controlled transistor within each key, itself connected to a row in a diode matrix.

The advertised advantages of this design were as follows:

Implementation

The switch itself is somewhat involved. In addition to the self-contained plunger assembly, each key station requires a transistor, two resistors and a capacitor. The sense lines—the diode matrix columns—are held high by a fixed capacitor in each key position, which is charged from a separate shared feed. Pressing a key causes the base of the transistor to receive an AC signal delivered from the variable capacitor formed from the key’s capacitive assembly. With each AC cycle (suggested to be around 100 kHz) the transistor briefly activates and drains some of the charge from the capacitor, faster than it can recharge (as the charge line is fed through one of the two resistors). Once the capacitor is drained, there is a current path around it and through the transistor to ground, pulling that sense line low. The core portion of this circuit, specific to each individual key, is shown below:

Each column of the diode matrix is fed into a sense amplifier, and these in turn are synchronised by the strobe circuitry to ensure that no output is signalled until the diode matrix columns have stabilised. Additional circuitry is used to suppress output when two or more keys are held simultaneously. Non-encoded keys still have the same transistor–capacitor–resistor arrangement, but they are fed directly into a simplified sense amplifier that is not involved with the strobe or lockout circuitry. Although it is not specifically stated as such, some of the circuitry (such as in the sense amplifiers) may serve the purpose of converting the AC signal from the oscillator to a DC signal that forms a stable output: the per-switch transistor will be cycling at the same 100 kHz rate as the AC feed through the variable capacitor.

The patent does not detail how overtravel is managed. Most capactive keyboards use an overtravel pad made of foam, that allows the movable plate (a foil layer on the bottom of the pad) to directly reach the PCB. Control Devices instead use a thin metal sheet attached to what is seemingly a rigid plunger. Possibly the switch is able to actuate at a capacitance level lower than that of full travel. Claims that these keyboards are “foam and foil” suggests that possibly the design changed, or that the object in the advertisment photo above the conductive plate is formed of a compressible material.

Models

The only known model is the CDK-3, advertised in Electronics magazine in January 1971. This is an ASR-33 format keyboard, described as “low weight” at 1.5 pounds (equivalent to 0.68 kg) and thus considered suitable for portable equipment.

Matrix scan

The only known Control Devices keyboard is the type originally used in the DEC VT05 (introduced in 1970). This is a matrix scan keyboard although the implementation is similar. Per the technical material for the VT05, each key only has one transistor and one resistor; the collector and emitter are wired into the matrix and read by a conventional SMC KR2376 single-chip encoder (custom part KR2376-17), typically used with conductive switches.

Whether the matrix scan type is still part of the CDK family has not been determined.

Documentation

All documentation was scanned by Bitsavers unless otherwise noted.