Beleaguered NASA engineers have uncovered new problems in space shuttle hardware that could further delay the shuttle's next flight, already postponed until at least August, sources said yesterday.

The new problems -- apparently the result of flaws in the manufacturing process rather than the fundamental design -- turned up in the last few days in the shuttle booster aft skirt segments and the shuttle main engines. The impact on the schedule is being assessed, David Winterhalter, NASA's director of systems engineering and analysis, said yesterday when asked about the problems.

The previous target date of June 2 was scuttled when a segment of the shuttle's booster nozzle failed in a Dec. 23 test firing at the Utah plant of booster contractor Morton Thiokol Inc. That was the second postponement since the Challenger disaster, which occurred two years ago Thursday, and officials have said they will delay as long as they consider necessary in the interests of safety.

Engineers are still uncertain of the precise cause of the Dec. 23 failure in the booster nozzle but concluded that the failed part, known as the outer boot ring, had a fundamental design flaw. They have decided to scrap that design and replace it with a backup that was successfully tested in August. At least two and probably three more full-scale test firings of the boosters, using the backup design, are planned before the next shuttle flight, Winterhalter said.

Top NASA officials had hoped to set a new target date for the next flight at a meeting here yesterday, spokesmen said. Now that decision could come later this week.

The problem in the aft skirt was discovered as a result of improved methods of so-called "nondestructive" testing -- ways of inspecting the parts without tearing them apart physically. The aft skirt is the cone-shaped section of the booster that partially covers the steering nozzle and provides rigid support for the shuttle as it sits on the launch pad.

The new technique, which involved ultrasonic scanning of the hardware, detected cracks, or voids, in welds on two out of six aft skirts that have been scanned so far, out of an inventory of 30, sources said. The flawed aft skirts included one that was scheduled to be used on the next flight after the shuttle launches resume.

Engineers at Kennedy Space Center are scanning the two aft skirts to be flown on the next shuttle flight and so far have uncovered no cracks. As of late yesterday, they had scanned most of one and part of the second, Winterhalter said.

The cracks had not shown up at first under the old technique, which used X-rays, sources said. After the ultrasonic scanning spotted the problem, engineers were able to target their X-rays more precisely, at a different angle from before, and they found the flaws using that method as well.

The problems occurred at the point where the aft skirt is welded to the "hold-down post" that holds the vehicle on the launch pad, according to Ed Medal, a spokesman at Marshall Space Flight Center.

The problem in the main engines was discovered late last week, according to spokesmen at Marshall, when engineers testing an engine for a future flight found "mistracking" in a weld in a part of the engine turbopump known as the first stage blade platform seal, or more commonly referred to by engineers as the "fish mouth."

This alerted the team to check the same part in other engines for similar flaws, Winterhalter said. They may have to remove the pumps from engines already delivered to Kennedy Space Center for the next shuttle flight and replace them, he said. It is possible that this could be done and the replacement pumps could be adequately tested without causing further delay in the flight schedule, he said.