OGDEN, UTAH, AUG. 28 -- Still recovering from the dire effects of a simple plumbing problem, National Aeronautics and Space Administration officials today postponed a crucial test firing of a redesigned shuttle booster until at least Sunday.

Originally scheduled for 3 p.m. Thursday at the Morton Thiokol Inc. plant north of here, the test was aborted three times in four hours because of a series of computer glitches caused when a water hose slipped off its coupling.

The problems all involved ground support equipment and were not related to the redesigned booster, officials emphasized. The shuttle launch schedule is not yet affected by the delay, they said.

The hose failure, which sent water spraying over the test stand area seconds before the planned ignition, led to two subsequent halts in the countdown due to previously unencountered problems connected with recycling the computers, officials said.

The test teams for NASA and booster contractor Morton Thiokol were running tests today to confirm the computer situation, officials said, and planned to conduct a dry run of the test countdown on Saturday.

In a statement, Morton Thiokol officials emphasized that they "are working at a measured pace and won't proceed with the test until they are satisfied that all problems identified Thursday are fully understood and have been resolved."

A decision on whether to try again Sunday will be made Saturday night, possibly around midnight, they said.

The engineers determined today that yesterday's third countdown was forced to stop just seconds short of ignition because a computer sequencing problem prevented the startup of the booster's steering system, known as thrust vector control.

The engineers traced the problem back to the point when the day's second test countdown was terminated three minutes before ignition. The test team subsequently recycled the countdown to 60 minutes before ignition and in doing so reset the main computer, officials said.

The problem was new to the engineers because it apparently was the direct result of aborting the test countdown, which had never been done before with a shuttle booster.

"We have never before gone through an abort and {reset} the main computer to start over," NASA spokeswoman Sarah Keegan said.

"When that happens, apparently sometimes the main computer will send a random reset command to a second computer. And that second computer reset itself so that the two computers were out of sync and the second one couldn't receive the command from the main one to start the {thrust vector control}."

For the test, the booster is hooked to ground equipment that serves as a stand-in for the shuttle "stack" to which the booster normally would be linked.

"If the water system problem at the test site had not caused the initial test abort and the subsequent recycling of the countdown, the test could have continued," NASA said in a statement. "The {thrust vector control} system would have performed normally."

The purpose of the test, considered a key psychological as well as technical hurdle for the shuttle program, is to verify that the new booster design will work in conditions close to those of a real launch. A poorly designed joint in the booster was blamed for the Jan. 28, 1986, Challenger accident that killed seven crew members.