Monday, December 24, 2007

Shuttle Launch Update: Fuel Sensor Problem Isolated

Image credit: NASA/George Shelton

The caption reads: Space shuttle Atlantis stands on Launch Pad 39A at NASA's Kennedy Space Center in Florida.

Now that it's Christmas Eve, no doubt you're wondering what's been going on with the space shuttle launch that was delayed due to problems with a fuel sensor, right? As it turns out, they're still working on it.

The fueling test NASA performed last Tuesday revealed a problem with a feed-through connector:

During today's tanking test, sensors 2 and 3 showed open circuit failure indications for a few seconds before returning to normal operation. Sensor 1 failed shortly thereafter and stayed that way for the duration of the test. Sensor 2 worked properly throughout the tanking test but sensor 3 failed several hours after recovering from its initial open circuit indication.

The wiring to all four ECO sensors, as well as the 5 percent level sensor, pass through the same feed-through connector.

"Today we had problems on three of those sensors and we captured the data," Hale said. "And the data is indicating we have a problem at what we call the feed-through connector that leads the wires from the inside of the liquid hydrogen tank in the cryogenic fluid to the exterior of the tank. That's a multi-part connector and our time domain reflectometry instrumentation has indicated that is where our open circuits are occurring at these very cold temperatures.

Atlantis fueling test points to source of sensor trouble

Time domain reflectometry is radar for wiring - send precisely timed electrical pulses down a wire, and what is reflected back tells you where breaks and discontinuities exist in the conductor. In this case, it found a break at the feed-through connector that leads from the external sensor processors to the sensors themselves.

A feed-through connector is a device that leads electrical signals from a cable into the interior of an enclosed space.

NASA's shuttle mission site reports today:

Foam removal operations are scheduled to be completed over the weekend at NASA's Kennedy Space Center as part of the plan to remedy failed readings on space shuttle Atlantis' fuel sensor system. The foam is being removed from a small section of Atlantis' external fuel tank so technicians can get to a pass-through electrical connector that testing pinpointed as the likely source of the sensor issue.

Shuttle Mission Main Page, Dec. 21, 2007

Damaris Sarria, who works at Cape Canaveral for Boeing and has a blog called How I Am Becoming An Astronaut, writes:

[A]t least now we know that the cause from the last couple of scrubs was due to a critical three-part "feed-through" connector. The feed-through connector leads the wires that carries the sensor data from the inside of the liquid hydrogen tank in the cryogenic fluid to the exterior of the tank. We're hoping to find out today how long it would take to fix this issue.

The Infamous Feed-Through Connector ...

The problem with these feed-throughs is that they're in parts, and that some of those parts are at very different temperatures at various times in a mission. To the left is a NASA drawing of such a feed-through larger version here. Before the tank is filled, all parts are at ambient temperature. As the tank is filled with liquid hydrogen and oxygen, the interior is chilled to something like -200 degrees. The outside part will remain at ambient, and will cool more gradually. Meanwhile, the cases the feed-throughs are mounted on will also be changing temperatures, and so will be contracting and distorting their shapes somewhat. As you can see from the drawing, that means the parts connected to one case will be rotated, possibly to the point where the electrical connections no longer meet. Just imagine the dark gray section shifting up or down relative to the light gray section and you'll get the idea. Now imagine that at the same time the electrical leads (in yellow) are also getting shorter, since they are also getting colder, and you'll understand why this thing could go bad.

The good news is that they've found the problem. The bad news is that they have to figure out how to fix it by January 10:

NASA mission managers will meet on Dec. 27 to further discuss plans to fix the fuel sensor system that postponed two launch attempts for mission STS-122 in early December.

The next opportunity to launch Atlantis will occur no earlier than Jan. 10.

Shuttle Mission Main Page, Dec. 21, 2007

Sarria mentions that this launch window closes on January 13, so there isn't much time.

To repair the connector, or even to reach it, is not a simple task:

By the end of the weekend, foam will be removed from around a suspect feed-through connector near a strut on the aft of Atlantis' external tank. Diagnostic tests Tuesday indicated the connector caused intermittent low-fuel sensor readings that scrubbed launch attempts on Dec. 6 and 9.

Foam being removed to reach connector

The link has a photo of someone cutting through shuttle tank insulation, just to give you an idea. Spaceflight Now elaborates:

The connector in question can be accessed at the launch pad but how much time it might take to replace suspect components will depend on whether the fault is found to be inside or outside the tank.

"Some timelines have been developed to change out those parts that can be reached from the outside, and they are on the order of a week to 10 days kind of work," [shuttle Program Manager Wayne] Hale said. "However, the part that's difficult to get to is the socket connector on the inside of the tank and that would be more invasive. You would have to go inside the tank through the manhole cover we've got at the bottom or some other access point and that obviously would be a longer-term operation."

Atlantis fueling test points to source of sensor trouble

I'm sure that what they'll be discussing at the Dec. 27 meeting will include whether the connector design is reliable enough to trust it in another launch attempt, or whether they can test it before they fill the tank to ensure that it will be reliable. At least, those are the questions I'd be asking.

On a much sadder note, International Space Station astronaut Dan Tani's mother was killed in a car-train crash last Wednesday. My sympathies go out to him and his family. I know it's tough being away when these things happen.

UPDATE (Dec. 26): When I read this sentence that I wrote here:

The problem with these feed-throughs is that they're in parts, and that some of those parts are at very different temperatures at various times in a mission.

I can't help but cringe. To somoene who doesn't work with these things, it might seem that I'm saying that the fact that the feed-throughs are in separate parts is a design flaw. I want to emphasize that's not so. It's just part of the problem that the designers have to address. The thing must of necessity be made of separate parts, each of which has its own rate of heat expansion or contraction, and will contract or expand in its own direction. That's the problem.

No comments: