Since it was launched in 1990, the Hubble Space Telescope has undergone eye surgery to correct its flawed vision, an electronic brain transplant and numerous other upgrades and repairs to replace aging or broken components and to install new instruments. On Wednesday, the $2 billion observatory faces its riskiest operation yet, the equivalent of open heart surgery to replace its central power system controller.

For the first time since it was launched, ground controllers will shut Hubble completely down so astronauts John Grunsfeld and Richard Linnehan can disconnect its six batteries and replace its power control unit, or PCU, a 160-pound black box that was not designed to be replaced by spacewalking repairmen.

Once Hubble is shut down for the PCU change out - about the time Grunsfeld and Linnehan exit the shuttle Columbia's airlock at 1:27 a.m. - an imaginary "thermal clock" will start ticking. The astronauts must complete the PCU replacement and restore power by the end of the day, or Hubble's critical systems could be damaged by the effects of the normal temperature extremes of space.

While NASA managers typically downplay the drama associated with key events like this one, Anne Kinney, director of astronomy and physics at NASA headquarters, said she was "nervous as hell" about the PCU swap out.

The PCU has 34 closely spaced, hard-to-reach connectors in two rows down its left side. Two more connectors and ground straps are located on the bottom. The box measures one foot by two feet by four feet long and it is buried in a mass of cables and other equipment.

"What makes it difficult is, as you're facing the PCU, those connectors are on the left-hand side, they're not staring right at you, they're on the left face," said lead flight director Bryan Austin. "And that's on the side that that bay door is hinged. For the suited crewman to reach his hand in there, he's pretty much reaching in there blind.

"I kind of equate it to changing out spark plugs on your car. There's always those spark plugs down there where you sort of can't see real well, you've just got to go down and feel and make sure you're oriented such that you're unscrewing it without a lot of offset force."

Grunsfeld and Linnehan originally planned to start the spacewalk an hour earlier than usual because the job is expected to take at least seven-and-a-half hours. While a half hour of contingency time is built into that schedule, the spacewalk could run even longer. But nine hours is the limit on NASA's current spacesuits.

But Austin said Tuesday the spacewalk would begin at the same time as the others in Columbia's mission - 1:27 a.m. - because it had become clear the astronauts simply could not get prepared in time to leave the airlock an hour ahead of schedule.

Given the challenge - and risk - associated with replacing the PCU, one might ask why NASA would order such a challenging bit of orbital surgery. Hubble project scientist David Leckrone provided a detailed explanation during a briefing Tuesday:

"In 1993, we like to say we performed eye surgery on Hubble to correct its poor vision," Leckrone told reporters. "And then in servicing mission 3A in 1999, we successfully changed out the central spacecraft computer in Hubble and a lot of us referred to that as brain surgery. And I guess carrying this analogy to (Wednesday's) activity, you could say that tomorrow Hubble gets a heart transplant."

"You can think of the PCU as the central pump that circulates electricity through the spacecraft. I don't want to carry this analogy much further than that, but in a sense that does describe our attitude and approach and feelings about what's going to happen tomorrow. The two doctors, John Grunsfeld and Rick Linnehan who are going to be performing this surgery, have studied hard, trained long and hard and are probably the world's most skilled surgeons for this particular kind of operation.

"But despite that, of course, any major surgery entails a certain degree of risk," Leckrone said. "And that will certainly be the case (Wednesday), a certain degree of necessary risk to do a necessary job.

"The PCU has been with us since Hubble was first launched in 1990, it's roughly 12 years old now and it's had a number of gnawing problems all these years that have given us a little anxiety from time to time and that we've sort of had to nurse along," Leckrone said. "There's one in particular that is particularly threatening to the future science capabilities of Hubble and in my view is the most important factor going into the decision to change the PCU out.

"The technical term for this is the 'bus bar fault.' And what that really means is there's a piece of metal that is physically attached to structure within the PCU that's held down by one or more screws. And one of those screws has come loose as far as we can tell. The consequence of that is the electricity that's normally supposed to pass through this joint from the spacecraft's batteries to the scientific instruments and other components of the spacecraft is impeded by the resistance because the contact is not a perfect contact between the bus bar and the structure it's attached to. Some people say Hubble has a loose screw, or at least the PCU has a screw loose. And that's a correct statement.

"The danger or risk that this poses to science can be viewed sort of step by step," Leckrone continued. "At the present time in this mission, we're putting in a new Advanced Camera for Surveys, which uses a good deal of electrical power, we're putting in the NICMOS cryocooler, which uses - the cooler itself - uses as much electrical power as a stand-alone instrument. It wil revive the NICMOS instrument, which also uses a fair amount of electrical power, and then STIS (spectrometer) and for that matter, WFPC-2 (Wide Field Planetary Camera), use a fair amount of electrical power.

"With the present bus bar fault in the spacecraft, we cannot at the present time operate and adequately power all of these instruments at the same time. So if we didn't change out the PCU, we would have to go into a power management mode where some instruments would have to be cycled on while others are cycled off. This is a very laborious, time-consuming and painful way to operate the spacecraft and would surely cause a reduction in our overall very high science observing efficiency.

"The present impedance or resistance of this joint is 22 milliohms. At that level, even if you shut off all the instruments you could not do infrared science with both the NICMOS and the cryocooler, there would not be enough power to allow that set of components to be operated. Even at the current level.

"Now since this is a loose screw and it could become more loose with time and the resistance from impedance could go up further, when we got to the 110 milliohm level we could only power one Hubble instrument at any given time and all the rest would have to be shut off. And if we got up to 165 milliohms, no instruments could be operated. And if you want to really get scary, if you got into the 2-ohm to 4-ohm resistance level, which means the screw has come really loose and the thing is just sort of sitting there barely making contact at all, then potentially if we went into one of our hardware safe modes, two of our batteries would overheat and potentially rupture. And that would be just catastrophe.

"So I think you can see there are strong scientific implications for failing to correct these problems and therefore, as with any beloved relative, you're worried about sending them in for bypass surgery or even a heart transplant, but you realize the risk of not doing it is severe. And in this case, the risk of not changing out the PCU is severe. And so that's why we're doing it."

Grunsfeld will begin the spacewalk as a "free floater," installing thermal covers on the new solar array diode boxes and deploying thermal shields over equipment bays five and 10 that were put in place by Newman and Massimino. At the same time, Linnehan, riding the robot arm, will disconnect Hubble's six batteries. Three batteries are located in equipment bay two and the other three are in bay three.

Grunsfeld will deploy light shields put in place earlier to protect the telescope's sensitive star trackers. The astronauts then will open equipment bay four and begin the tedious job of disconnecting and removing the PCU.

After latching the door open and installing a plastic cable holder to keep the connectors in order after they are removed, Linnehan will begin disconnecting the PCU, starting at the top and working down the double rows of cables.

With six cables to go, Grunsfeld will take Linnehan's place on the arm, complete the PCU's electrical disconnection and remove it. He then will install the new PCU and make all 36 electrical connections.

"These are not our typical fully EVA- friendly connectors with the big wing tabs on them that make it convenient for an EVA-suited crewman to manipulate," Austin said. "We've worked really hard to have a special tool that we're going to use to get a good grip on these and get through this."

In addition, the left side of the new PCU is canted slightly, giving Grunsfeld slightly better access for re-attaching the electrical cables. Assuming all goes well, the astronauts will reconnect Hubble's six batteries and retract the thermal covers, clearing the way for anxious ground controllers to re-power the observatory.

It will take six minutes or so to power down the Hubble Space Telescope, but it will take several hours to power up its myriad electrical systems. The work will begin as soon as Hubble's batteries are reconnected. At that point, anxious engineers will find out if all of those system do, in fact, power back up normally.

"I think the chances are extremely low that we're going to lose any equipment," said program manager Preston Burch. "The other thing to remember is that Hubble has tremendous redundancy. We've got at least two of just about everything on the vehicle. If something does go kaput on us, it's certainly not going to put us out of business by any stretch."

Austin said mission planners designed all five repair spacewalks with various "breakout" points to ensure the telescope can be left in a safe configuration overnight in case any given task cannot be completed during a single EVA.

"There is a period of time, though, on EVA-3, the PCU, where it's kind of like open heart surgery except we don't have it on a heart-lung machine, we've just taken the heart out and we've got that period of time before the telescope is even awake again, before it's even viable to be powered through the orbiter," he said.

"So we've got a minimum amount of work we have to do that day once we take the PCU out, get it back in and get some minimum amount of connectors hooked back up so at least we can apply power from the orbiter back through the telescope to revitalize it. So that day will be extensive."

Various contingency plans have been developed to provide enough power to Hubble to keep the observatory alive overnight in case of problems completing the PCU installation. Eleven of the 36 connectors are needed to keep Hubble safe overnight with its batteries still disconnected. At least 23 of the connectors must be hooked up before Hubble can be released from the shuttle.

The bus bar fault first surfaced after the 1993 servicing mission to correct Hubble's flawed vision. Since then, the impedance caused by the loose screw has varied from zero to as high as 40 milliohms. An ohm is defined as the resistance in a circuit in which a potential difference of one volt produces a current of one ampere. At present, the impedence is roughly 22 milliohms. Burch said the severity of the problem is unpredictable in the long term.

"There's no way to predict the future on something like this," he said Tuesday. "It's sort of a 'Dirty Harry' question, you know, are you feeling lucky today. And we don't like to trust to luck for something as important as HST. We feel that the consequences of not doing something about this problem are potentially much more profound."