40 Years ago: The computer “Time Skew”

April 10, 1981: Columbia appears ready for the first Shuttle launch.

*

*

            Finally, launch day; April 10, 1981.

            Launch day, after a year of delays.  Two years, depending on how you count.

            We’re ready, you and I, to pilot Space Transportation System (STS) – 1.  That’s the official designation of the first flight of the Space Shuttle.

            Finally launch day — maybe.  The first launch of a new vehicle, one without unmanned precursors, do we really expect to get off the pad on the first try?  Especially with a vehicle this complex and revolutionary.

            We’re up at 2:45 a.m., aiming for a 6:50 a.m. (EST) launch from Pad 39-A, the same pad that sent Apollo 11 to the moon.  It’s been nearly six years since the last U.S. manned flight, the Apollo-Soyuz mission.  It’s going to be a long day, even if everything goes smoothly, with a slew of engine burns scheduled in the first seven hours of flight.  Mission rules limit us to a 20-hour day — and that means we’re limited to about 3 hrs. of hold time on the pad, should delays occur.

            We ride to the pad in the pre-dawn darkness of 4:30 a.m., the Shuttle with it’s External Tank painted white blazing in the spotlights emphasizing what an ungainly amalgamation it is — fat fuel tank flanked by twin Solid Rocket Boosters like two Roman candles.  And Columbia, the orbiter with a wing span of 78 ft., attached to the side of the ET — like a butterfly on a milkweed stalk.  Nothing like this has ever flown.  And we’re going to be aboard, a test flight of 36 orbits, 54.5 hours, ending in a gliding landing at Edwards Air Force Base, California.

            We climb inside on our knees, and up to the upper level of the of the 122-ft.-long orbiter, the flight deck, and strap in the two ejection seats.  The count proceeds amazingly smoothly.   After sunrise over the Atlantic, golden light spreads above a a low layer of clouds.  Optimism in the air — the opening of a new era of spaceflight, the beginning of low-cost, routine access to low earth orbit.  The next giant leap, they are calling it.  

            At T- minus 20 min., Columbia‘s four primary redundant computers switch to “Terminal Countdown Mode 101,” and a fifth computer, serving as backup, comes online and synchronizes with the primary ones.  Or is supposed it.  Warning lights flash; CRTs display a string of error messages.  Like nothing we have ever seen.  The backup computer has failed to synch with all four primaries. Perform a reset — is it OK?  

            Now, reaching the planned T- minus 9 min. hold, about 16 min. from actual launch, another problem draws our attention.  The acid levels in the electricity-producing fuel cell #3 are high.   The count is halted while launch control assesses whether this problem is real.

            The team decides the fuel-cell warning is caused by overly sensitive instrumentation.  Meanwhile, the problem with the backup computer reappears — and it’s for real.   The news comes almost exactly at our original launch time of 6:50 a.m. The ground believes the fault resides in the software of the backup, decides to reprogram it, which should only take 7 min.  But then they want to take a detailed look at the program and tells us it will take about 90 minutes.  The count is recycle to a hold point at T minus 20 min.  Launch Control asks us how we feel about waiting that long.

            “We feel like we’ll be here, ready to go.”

            They believe the fault resides in the backup computer, so focus on it.  All five computers are the same IBM AP-101 models developed for the military.  The four primary ones were programmed by IBM.  But to provide redundancy, the backup was programmed with different software developed by Rockwell International, prime contractor for the Shuttle.

            The 90 minutes transpires.  Launch Director George Page comes on the line.  “Hey, I don’t want you to think we’d forgotten about you.  We’re trying to come up with a way to still salvage a launch attempt today.”

            “OK, George. We know you hadn’t forgotten about us.”

            “How are you holding out up there?”

            “Just laying here, you know.”   Incredibility, despite the pressure suits and uncomfortable seats, we’ve dozed a bit during the wait.  No longer.

            “Getting uncomfortable at all?”

            “We’re getting there, George — we’re getting there.”

            They’ve checked out the “BFS,” backup flight software.  Houston radios, “The proper program was in place.  Now we have to determine if that program is working.”

            About another hour goes by — we’re nearing five hours in the cockpit, approaching the time limit set by crew fatigue.

            We’re going to give it another shot — see if the computers will synch up this time, since the programs appear in good shape.   We begin counting down, aiming for a 10:20 a.m. launch.  It’s shortly before 10 a.m.  At T- minus 16 min., the computers should synch up.

            T – minus 16 and . . .  “We’ve just had a repeat of the problem.”   

            “We will scrub the launch for the day.”  They’d struggled with the mysterious computer glitch for more than three hours.

            We tell the launch team, “You all did real good.  We’re sorry we couldn’t go.”

            The earliest we could launch is Sunday, April 12.  If they can correct the problem that swiftly.  Which to us, who know so well how complex this new vehicle is, does not seem likely.

*

            Yet by evening they’ve traced the problem, a “timing skew ” — a once in a million error. Or rather, they estimate it occurs about once every 67 times the computers synch up.  The cause is actually in the primary four computers when they are switched on at T – minus 15 hours and receive a time signal which comes in 80 – millisecond cycles.  The signal came right at 40 milliseconds in that window, which caused, in effect, a rounding error.   That led the primary computer’s time to be 40 milliseconds (one-twenty-fifth of a second) off when the the backup tried to synch up — which has to take place in just 5 milliseconds.  When it didn’t happen, the backup halted operations.

            Changes are such a problem won’t happen again.  But the computers will be switched on at T – minus 30 hrs. and run from then on to guard against a mismatch.

            We’re go to launch at 7 a.m. on April 12.   Which means the new era of spaceflight will open exactly on the 20th anniversary of Vostok 1 and Yuri Gagarin, the first man in space.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s