Editing Mariner 1 (section)

== Launch failure ==

[[File:Atlas Agena with Mariner 1.jpg|thumb|Atlas Agena with Mariner 1]]

The launch of Mariner 1 was scheduled for the early morning of July 21, 1962. Several delays caused by trouble in the [[range safety]] command system delayed the beginning of the [[Countdown#Rocketry|countdown]] until 11:33&nbsp;p.m. [[Eastern Time Zone|EST]] the night before. At 2:20&nbsp;a.m., just 79 minutes before launch, a blown [[fuse (electrical)|fuse]] in the range safety circuits caused the launch to be canceled. Countdown was reset that night and proceeded, with several holds, planned and unplanned, from 11:08&nbsp;p.m., through the early morning of the next day.

At 9:21:23&nbsp;a.m. on July 22, 1962, Mariner 1's Atlas-Agena lifted off from LC-12. The launch went entirely according to plan up to booster separation. During sustainer phase, the guidance system began issuing improper steering instructions, causing the Atlas to fishtail left and right. Its flight trajectory began to point downward and to the left of where it was supposed to be, creating the danger that it could impact in the crowded Atlantic shipping lanes. At 9:26:16&nbsp;a.m., just six seconds before the Agena second stage was scheduled to separate from the Atlas, at which point destruction of the rocket was no longer possible, a [[range safety officer]] ordered the rocket to self-destruct, which it did--the Atlas Flight Termination System was also designed to destruct the Agena if activated but Agena had no Flight Termination System of its own and could not be destructed following Atlas SECO. Telemetry signals were received from the probe for another 1-1/2 minutes. Mariner program director Jack James believed the destruction of the rocket was unnecessary and it would not have landed anywhere but the middle of the ocean.<ref name=report/>{{rp|87}}

===Cause of the malfunction===
Because of the gradual rather than sharp deviation from its course, JPL engineers suspected the fault lay in the flight equations loaded into the [[SM-65 Atlas#Guidance|computer that guided Atlas-Agena]] from the ground during its ascent.<ref name=avweek1962a/> After five days of post-flight analysis, JPL engineers determined what had caused the malfunction on Mariner 1: an error in the guidance computer logic combined with a hardware failure.<ref name=ceruzzi/>

The Burroughs guidance computer used data transmitted to it from the rate beacon on the Atlas and used this information to issue steering commands. The guidance program was supposed to contain a hyphen which instructed the computer to ignore data coming from the Atlas's rate beacon if it failed in-flight to prevent incorrect commands from being sent but it had been accidentally left out of the program, which a technician at Cape Canaveral entered into the computer as it was without realizing the program he was sent had a mistake in it.{{r|ceruzzi|vice}} (The very same mistake that would afflict [[Phobos 1]] 26 years later).
 	
During its ascent, Mariner 1's booster briefly lost [[Missile guidance#Remote control guidance|guidance-lock]] with the ground. Because this was a fairly common occurrence, the Atlas-Agena was designed to continue on a preprogrammed course until guidance-lock with the ground resumed.<ref name=avweek1962b/> When lock was reestablished, however, the faulty guidance logic caused the program to erroneously report that the "velocity was fluctuating in an erratic and unpredictable manner", which the program tried to correct for, causing actual erratic behavior, which prompted the range safety officer to destroy the rocket.<ref name=ceruzzi/>

The incorrect logic had previously been used successfully for [[Ranger program|Ranger]] launches but the rate beacon had not malfunctioned on those so the problem didn't show up there. The Mod III-G guidance system used on Atlas-Agena vehicle was a persistent source of trouble and malfunctioned on many launches since Atlas-Agena began flying in 1960. It was an adaption of the Mod III guidance system used on Atlas B, C, and D missiles which had the original vacuum tube electronics converted to transistors, but the modification had been done hastily and was unreliable. After repeated Atlas-Agena guidance failures, the Mod III-G was redesigned during 1963 to properly accommodate transistor electronics.<ref name=ceruzzi/>

The catastrophic effects of a small error "summed up the whole problem of software reliability" and contributed to the development of the discipline of [[software engineering]].<ref name=ceruzzi/>

Subsequent popular accounts of the accident often referred to the erroneous character as a "hyphen" (describing the missing component of the symbol) rather than an "R-bar"; this incorrect mischaracterization was fueled by [[Arthur C. Clarke]]'s description of the malfunction as "the most expensive hyphen in history".<ref name=vice/>