Input Devices Nintendo Power Glove

Mattel's wearable NES controller arrived promising a revolution in how humans play video games, delivered something far stranger, and spent its afterlife as the unlikely cheap hardware of choice for early PC virtual reality researchers.

What It Was

The Power Glove was a wearable game controller for the Nintendo Entertainment System, manufactured by Mattel and fitted with 3D sensors and a programmable keypad.¹ It wore on the forearm rather than sitting in the hand, and it tracked position across three axes plus rotational roll, giving it four degrees of freedom in total: x movement, y movement, z movement, and roll.⁷ Individual fingers (excluding the pinky, which carried no sensors) bent to distinct values between flat and fully closed.⁷ The grip sensors, however, were the glove's most conspicuous technical failure: the only reliably distinguishable states were 0x00–0x80 for a flat hand and 0xFF for a full grip, with nothing useful in between.⁷

Nintendo Power Flash's breathless promotional copy from Summer–Fall 1989 promised that "a simple upward sweep of your arm" would make game characters jump, planes soar, and villains run, and that a clenched fist would deliver "the knockout punch."¹ Actual play was considerably less cinematic. The sensors were imprecise, latency was visible, and the gesture-to-action mapping that sounded poetic in marketing copy proved frustrating to calibrate in practice.

Reception

Nintendo's own promotional newsletter, Nintendo Power Flash, was predictably enthusiastic. The Winter 1990 issue called it "perhaps the most eagerly awaited controller ever" and claimed it delivered "pinpoint precision."⁴ Reader mail told a more complicated story. A letter published in the Spring 1990 issue from Eric Pratt of Lasalle, Quebec captures the early-adopter experience honestly: he had gone into a store test session expecting nothing special after the NES Max and NES Advantage, reported genuine excitement at slipping it on, and recommended it, but grounded his endorsement entirely in the adrenaline of novelty rather than any specific gameplay achievement.² That gap between the thrill of wearing it and the frustration of actually playing with it would define the Power Glove's reputation.

In the United Kingdom, Mean Machines listed it at £49.95 including postage and packing as late as January 1992, by which point it had become a mail-order bargain item sitting alongside software listings.¹² In the United States, PCVR magazine reported that by 1992 Mattel had discontinued production, and that remaining units could be found at KMart and Toys R Us for anywhere between $20 and $50.⁷ The game software pipeline that was supposed to justify the hardware (a "new lineup specially designed for it" that Nintendo had promoted in 1989¹) had effectively dried up.

The Second Life: PC Virtual Reality

Here is where the Power Glove's story becomes genuinely interesting. Once clearance prices put a four-degrees-of-freedom sensor glove within reach of hobbyists and researchers, a cottage industry of PC interfacing emerged that had nothing to do with Nintendo.

The glove runs on 5 volts, which made it electrically compatible with a standard PC keyboard connector. Pins 4 and 5 of a five-pin keyboard connector supply exactly that voltage.⁵ The glove used a non-standard seven-pin connector and a modified synchronous serial transfer protocol, which initially limited its use; but once the wiring was understood, connecting it to a PC's 25-pin parallel port was a matter of mapping four pins.⁶ The DB-25 to Nintendo connector mapping that circulated in the VR hobbyist community was: pin 2 to pin 2, pin 3 to pin 3, pin 13 to pin 4, and pin 18 to pin 1, with ground on Nintendo pin 1 and +5V on pin 7.⁶

Power could be drawn from the keyboard connector, from an external source, or from an internal peripheral connector, each option trading convenience against portability.⁶ The color coding of the Mattel wiring was noted to vary from unit to unit, which made every interfacing job slightly different and required verifying wires with a multimeter before committing to a connection.⁵

PCVR magazine, published out of Laramie, Wyoming, covered Power Glove interfacing across multiple issues. Issue 1 addressed hardware interfacing to the IBM PC; Issue 2 covered software.¹⁴ One prominent circuit, designed by Ron Menelli and dated November 19, 1991, built a 68HC11-based interface board around a single MC68HC811E2P microprocessor, a MAX232 level converter, an MC34064P-5 voltage supervisor, an 8 MHz crystal, and associated passives, with a Nintendo connector on one end and a DB-25 on the other.⁸¹³ The microprocessor alone cost approximately $50 at the time.⁸

By 1992, PCVR Issue 6 was describing the Power Glove as a recognized low-cost substitute for the Polhemus Fastrak and VPL DataGlove combination, more powerful commercial VR input hardware that most individuals and small research groups simply could not afford.¹¹ The article was direct about its limits: the non-standard connector and protocol had "reduced its use in low-cost VR," and the article introduced a Power Glove Standard Interface (PGSI) intended to provide a standard host CPU interface so that code could be ported between machines with minimal rewriting.¹¹

Software written against the glove on the PC side included a 3D virtual hand rendered in Borland C 2.0 that tracked finger position on screen,⁷ and VR simulation work using the REND386 engine, where developers modified the Power Glove driver to implement different object weights and allow a virtual hand to pick up objects.⁹¹⁰ The REND386 integration added the glove's position and rotation data via pointer_read() and glove_update() calls inserted into the program's main loop.⁹

Specifications

The Archivist's Take

The Power Glove is a peculiar object to evaluate because it succeeded at something it was never designed for. As a game controller it was a commercial and practical disappointment: the sensors were too coarse, the dedicated software library never materialized, and players who wanted to punch a villain on screen found it easier to reach for a standard controller. But as an accidental piece of VR research infrastructure (clearance-priced, 5V-compatible, carrying four degrees of freedom in a form factor anyone could wear) it gave an entire generation of hobbyists and researchers a way into hand-tracking experiments that would otherwise have required equipment costing orders of magnitude more. The engineers at Mattel did not design it for that role. The VR community found it anyway, mapped its wires, reverse-engineered its protocol, and built a small literature around its limitations. That is a stranger and more durable legacy than selling Nintendo games.

References

1. Nintendo Power Flash Issue 5 (Summer-Fall 1989) The Revolutionary New Power Glove (1989)
2. Nintendo Power Flash Issue 7 (Spring 1990) Super Mario Bros. 3 (1990)
3. Nintendo Power Flash Issue 6 (Winter 1990) What's Hot This Christmas! (1990)
4. Virtual reality education and the Macintosh
5. PCVR.Issue.01.1992 (1992)
6. PCVR.Issue.01.1992 (1992)
7. issue2
8. original scanned
9. PCVR Magazine Issue 14 March April 1994 (1994)
10. PCVR.Issue.06.1992 (1992)
11. MeanMachines 16 Jan 1992 (1992)
12. issue2
13. PCVR Magazine Issue 14 March April 1994 (1994)