eileen-collins-ksc-99pc0176-mediumIt’s Women’s History Month in March, so I wanted to highlight the contribution of astronaut Eileen Collins, NASA’s first female Shuttle pilot and mission commander.

And then I recalled which mission Collins had flown in July, 1999 and thought it better to summarize how close we came to losing another STS mission and her crew that night.

STS-93 launched in the early morning darkness just a bit after July 23, 1999 had begun. It’s primary mission: Deploy the Chandra X-ray observatory, the heaviest thing that the Shuttle has ever shoved into space at 25 tons.

I’m not an engineer, and others have written or shown the more technical details more succinctly than I can do. I’ll link to those articles at the end of the post. Here, I’ll translate for other layman space enthusiasts what broke and what could’ve happened.

You might want to get a cup of coffee or whatever beverage or snack that tends to soothe you before reading. I might end up going all Up Goer Five here with simple language to explain a complex series of events.

Past Space Shuttle Program Manager and Flight Director Wayne Hale has a blog where he recounts his work with the launch of many an STS flight, including this one. He’s the best source for several elements involving this crazy launch, so here goes my take of it all.

Bad Thing #1: Something shot an SSME engine bell

Inside the each of the three Space Shuttle Main Engines is an injector, the central part where the liquid oxygen (LOX) is carefully forced into the combustion chamber, mixed with hydrogen and burns to make thrust. Now, that injector is a series of tubes, 600 posts in all. NASA inspects every one of them because, if any one of them breaks, an overload of LOX could fill the combustion chamber and cause the engine to explode. If they found a bad LOX post, they literally plugged it up with a gold pin.

On STS-93, one of those pins came loose. It shot like a bullet out of the injector and put several holes in the internal nozzle of the engine. Now, to keep the SSME from melting its own nozzle, liquid hydrogen is pushed through it. The nozzle is really a series of tightly packed tubes.

A large image that shows the telltale holes inside SSME #3 in center. (NASA)

So at launch, the right-most SSME engine bell was venting a few pounds of hydrogen uncontrolled into the combustion exhaust.

Not only was this bad as the nozzle could’ve failed (potential explosion), but a weak LOX post was now operational again (potential explosion) and the venting had caused a series of events that led to more oxygen flowing into the engine, which caused the turbopumps (which pump fuel) to heat up. The extra LOX getting pumped in was cutting into the total supply, threatening to short the vehicle’s velocity to get to orbit.

At that rate of leak, Columbia would either had failed to reach orbit and be forced to abort to a risky landing, or use a boatload of precious orbital maneuvering system fuel.

But something else happened that avoided that–and that involved another SSME.

Bad Thing #2: Failing SSME Computers

Each SSME is managed by two computers, one primary, one backup, called the Digital Computer Unit, or DCU. The A computer gets all the best data and access to sensors.

The center SSME’s A computer failed and stayed dead (see why in the next section). The B computer took over, but it has limited access to similar data and less telemetry to send to Mission Control to keep them updated.

Then one of the two sensors used by the A and B computers on the center engine to measure the engine’s average combustion chamber pressure stopped working.

B computer, working on only one sensor, believed that the center SSME pressure was 12% higher than it should be, and so throttled down the center engine.

That automatic change offset the fuel leak problems on the right SSME, leaving the vehicle with the just enough fuel to reach orbital velocity.

Not to be outdone for getting attention, however, the right engine’s B computer decided to fail, too. Thankfully, the primary A computer kept the still-leaky right engine going to cutoff. If right engine A computer failed then, Columbia would have had to make an abort.

Did I mention that all of these Bad Things were going on in the first minute of flight?

Now there were bonus problems that looked grim as well while the engines were being spooked.

Bad Things #3 and 4: A Bad Fuel Cell and Weirdness in an SRB

Right after liftoff, Commander Collins called out a “Fuel Cell PH” warning.

Wayne Hale mentioned a previous post on the NASA blog about a time when a fuel cell on the second STS flight was mixing its reactants very, very badly. Badly as in “kaboom.” STS-2’s flight was shortened because they had to shut down a fuel cell.

But fortunately, in a fashion faintly similar to the Apollo 13 incident, the issue wasn’t the fuel cells in STS-93 in themselves but one of the electrical buses shorting, causing a series of events that made the fuel cell instrumentation ratty and caused some power loss down the chain as the cell worked to recover.

One of the victims of the bus shorting was center SSME DCU A computer, which died altogether.

Also adding weird was one of the two hydraulic steering units on the right Solid Rocket Booster. At least, the data that the Flight Dynamics Officer had seen indicated something was terribly wrong.

If both hydraulic units failed, the SRB would stop steering and the whole launch vehicle might have careened out of control.

Again, only the telemetry was bad, not the SRB thrust controller. Still, FDO and others were trying to sort out and resolve multiple alarms all occurring simultaneously, working out what was real and what was bad instrumentation.

The Rest of the Story

Despite bad data and power losses and hydrogen leaks, STS-93 was only a few feet-per-second short of its needed velocity, easily resolved with an OMS burn later.

I like listening to flight director loops–the conversations of all the controllers and the flight director while also hearing the air-to-ground conversation from CAPCOM to the crew. Wayne asked the cool guys on the NASA SpaceFlight news and forum site (a favorite hangout of mine) to release to the public a video from their reserved forums. It’s STS-93’s loop but with graphics and captions included with the audio that show the bat-crap craziness in the first two minutes of flight and into MECO.

You can also read Wayne’s two blog posts on STS-93’s launch woes in general here and the specific engine LOX post problems here.

I’m going to go take a sedative.