The Atlas rocket program looks to return to its successful ways this week, launching a powerful U.S. military communications satellite Wednesday night with a tried-and-true engine valve instead of a newer design blamed for creating serious problems during the last ascent four months ago.

An onboard camera shows the Centaur RL10 engine firing during the troubled June launch. Credit: ULA

Atlas enjoyed a spotless 14-year consecutive success record covering 80 flights, but that score card was hit with a major blemish June 15 when a leaky fuel valve caused an Atlas 5 rocket to fall short of its intended altitude.

A pair of classified U.S. national security spacecraft deployed from the rocket's Centaur upper stage into a lower-than-planned orbit that day. Government officials said afterward, in essence, that the launch wasn't perfect but it wasn't a complete failure either.

Throughout the summer, sky-watching hobbyists who routinely track all sorts of satellites reported the secret duo from the launch were gradually boosting themselves higher and higher. Ted Molczan, a respected satellite observer, says the craft have made up about half the distance toward the expected launch altitude so far.

While the satellite recovery work has been underway, rocket-maker United Launch Alliance, engine builder Pratt & Whitney Rocketdyne and the Air Force have investigated the ascent incident and ordered a corrective fix to ensure upcoming launches won't suffer a similar mishap.

Several factors had coincided to create the problem, officials concluded, including an unusually lengthy firing by the cryogenic RL10 main engine during the launch sequence. But the liquid hydrogen fuel valve design -- a new version with only a half-dozen previous flights to its credit -- was at the heart of the problem.

"The investigation determined that the fuel inlet valve on the RL10 engine system failed to completely close after the first main engine cutoff on the Centaur upper stage. That happened because its closing force was not sufficient," Jim Sponnick, ULA's vice president for Atlas programs, said in an interview Friday.

"We learned that we had a combination of higher than expected friction forces inside the valve, as well as the effects of a very cold temperature on the helium that is used to operate this valve. This particular mission required a longer than average time for the first burn, which resulted in the liquid hydrogen flowing to the valve for a longer than typical period of time, which then decreased the temperature and increased the frictional forces within this valve.

"And those factors in combination led to the point where the valve did not have sufficient force to close itself."

After the Atlas 5 rocket's first stage had fired and dropped away, the Centaur ignited its engine four minutes into flight as the booster soared northeastward from Cape Canaveral on a trajectory experts believed illustrated that the clandestine payload was two ocean surveillance spacecraft.

The initial firing by the RL10 powerplant lasted about 15 minutes to reach a temporary parking orbit. As the cryogenic engine shut down, the fuel valve tried but failed to close completely.

The valve reached "near-closure but not full-closure," Sponnick said.

That left a pathway for the rocket's precious liquid hydrogen fuel supply to slowly escape as the Centaur quietly coasted in space for nearly an hour.

When the Centaur restarted the engine to deliver a final propulsive push for its payload, there wasn't enough fuel left in the tank to complete all of the mission's second burn. Although the firing seemed only fractionally short, roughly four seconds, the rocket simply could not achieve the desired orbit.

The satellites separated from the rocket and officials in the Atlas control center cheered, not fully aware that the launch had been crippled by a fuel leak.

"We did have some indications of that condition in real-time. There were several parameters on the vehicle that were indicating temperature differences from normal. It was not known at that time that there would be an affect on vehicle performance," Sponnick recalled.

But the realization soon hit the team and efforts commenced to figure out what went wrong.

"We've literally spent thousands of hours conducting very high fidelity tests to thoroughly understand the cause, as well as to test the replacement valve," said Sponnick, adding that the testing occurred at Pratt & Whitney Rocketdyne facilities in West Palm Beach and ULA's sites in Denver.

"We both developed very high fidelity test rigs to subject the valves to very flight-like conditions. With that high fidelity testing, we were able to replicate the failure-to-close scenario. And we've also been able to use those test capabilities to test the valves that will be flown."

The next Atlas 5 rocket launch is planned for Wednesday night from the Cape. Credit: Pat Corkery/Lockheed Martin

The remedy ordered for upcoming Atlas 5 and Delta 4 rocket launches -- both rocket families use versions of RL10 engines to power their upper stages -- is reverting back the previous fuel inlet valve design that officials fully trust.

"The upcoming missions are not flying a new valve design, it's actually the valve that has flown for many years on prior Atlas flights and also on the Delta flights that have flown to date. Our testing, as a result of the investigation, has shown that the heritage valve design has larger closing force margins," said Sponnick.

"And we've also taking several additional measures. We are performing a very high fidelity test of the heritage valve design that ensures that we are confirming prior to flight all of the opening and closing force margins for the valve. So we're doing a very specific test and we're also implementing some additional cycles of the valve in flight to provide even greater margin to close following main engine burns."

Some may question why there was a switch from the trusty old design to the new valve in the first place. But officials explained that they were forced to find another valve because the vendor was stopping production of the old device.

"The manufacturer that had supplied the valve in the past was essentially phasing out that product line. So we were developing a new design," Sponnick said.

"For the next several flights we do have in inventory the heritage valves that are being tested in the manner we've described. We are developing several options for longer term."

The Atlas 5 rocket will resume launches Wednesday evening with a planned 8:22 p.m. EDT (0022 GMT) liftoff from Cape Canaveral's Complex 41. Riding atop the 19-story rocket will be the U.S. military's first Wideband Global SATCOM communications satellite.

"Mission success is our absolute No. 1 priority. And that's exactly what we applied to this corrective action. We continue to be driven by technical success, not launch dates. So we are doing what we feel, jointly, is the right thing to make sure our customer -- the Wideband Global SATCOM program -- is going to get a great launch come (Wednesday)," Col. Michael Moran, the Air Force's Atlas group commander, said in an interview Friday.

"From my vantage point, I am more confident now than ever on the capability of that valve to perform as it's expected in flight. I'm absolutely confident that the processing that we've gone through, the reviews that we've gone through and the technical analysis that we've gone through responding to the anomaly in June has been comprehensive and effective. I have absolutely zero concern about the performance of that vehicle making its way to orbit."

"And I'll echo that," said Sponnick. "We're extremely confident in the understanding of the anomaly and the thoroughness of the correction actions that have been put in place."

This will be flight No. 11 for the Atlas 5 rocket fleet, and everyone involved hopes and expects to see a successful outcome when the massive satellite is deployed into a supersynchronous transfer orbit a half-hour after liftoff.

You can follow Wednesday's countdown and launch in our Mission Status Center. We'll have live play-by-play reports throughout the evening.

The status center also has an update on the launch delay from Tuesday to Wednesday.