Engine controller replacement details behind Artemis 1 launch delay

by Philip Sloss

On December 17, NASA announced a delay of the launch of the Artemis 1 mission until no earlier than March 2022.  Artemis 1 will be the first launch of the agency’s Space Launch System (SLS) rocket, which will send an uncrewed Orion spacecraft to the Moon.

The latest delay was caused at least in part by efforts to resolve a problem with an RS-25 engine controller on the rocket’s Core Stage.  The issue first occurred in late November; after four weeks of troubleshooting, NASA decided to replace the controller, which will take place over the next few weeks.

An updated forecast of when the Artemis 1 hardware will be ready to launch will depend on how long it takes to install a spare engine controller and fully resolve the problem.

Engine controller replacement uncertainty

After four weeks of troubleshooting, NASA announced its decision to remove an engine controller from one of the four RS-25 engines on the SLS Core Stage and replace it with a spare unit. Although some pre-launch preparations continued around the problem and troubleshooting, the issue has delayed final preparations to roll the vehicle out to the launch pad and pushed readiness to launch outside of a two-week period of opportunity in the second half of February.

“NASA is developing a plan and updated schedule to replace the engine controller while continuing integrated testing and reviewing launch opportunities in March and April,” a December 17 Artemis blog entry said. The March launch opportunity is now the soonest that Artemis 1 could launch; Launch Period 19 runs from March 12 through March 27.

Following that, Launch Period 20 opens on April 8 and runs through April 23. Both the March and April launch periods have “cutouts” for Orion eclipse constraints.

Credit: NASA/Jude Guidry.

(Photo Caption: The engine controller units for two of the RS-25 engines installed in the Artemis 1 Core Stage are highlighted in this image taken in November, 2019, when the engines were originally installed. As highlighted, access to the controllers can be seen from the bottom of the Core Stage for engine three (left foreground) and engine four (right foreground). The controller for engine four in the right foreground is the one that began failing to consistently power up in late November during pre-launch testing; it will be removed and replaced over the next few weeks.)

One of the dual digital computer units (DCU) in the engine controller for the number four RS-25, Engine 2060, stopped responding beginning on November 23. All four controllers were being powered up for Program Specific Engineering Testing (PSET) on the SLS Core Stage, which is stacked with the full Artemis 1 vehicle in High Bay 3 of the Vehicle Assembly Building (VAB) at the Kennedy Space Center (KSC) in Florida.

“During this event [the engine 4 controller’s] Channel-B failed to respond to application of power,” an early December agency status report noted. “Each [controller] contains two redundant channels labeled Channel-A and Channel-B. All other CSE CSECs were successfully powered on.”

The dual channels on all four controller units powered up successfully on November 12 during other checkout testing prior to the failure on the 23rd. The engine controller units (ECU) are the major component that was upgraded between use of the formerly-named Space Shuttle Main Engines (SSME) on the final Space Shuttle launches and adaptation of the SSME Block II design for use on SLS as the RS-25.

The engine controller for the SSME and the new unit for the RS-25 is a primary component of the engine. The controller directly operates the engine, from diagnostics and checkout before firing, through thermal conditioning and purges for startup during vehicle propellant loading, through ignition and throttling during mainstage operation, all the way through shutdown and post-shutdown purges and safing.

The controllers accept commands and relay data to and from the SLS flight computers in the Core Stage for launch operations. In addition to the new controller for the RS-25, new control software that runs on the controllers was certified for use on SLS.

The new engine controllers still employ two redundant digital computer units (DCU) that operate on independent control channels, A and B. The DCUs can operate independently to provide full redundancy, but can also share or access some data from the other.

The controller unit is a replaceable item on the engine, and remove and replace operations for controllers were occasionally carried out on engines installed in Shuttle orbiters while stacked in their vertical, launch orientation during that program. Former Space Shuttle Program Manager Wayne Hale, a veteran of a few dozen Shuttle launches as ascent flight director, recently wrote about one instance in which an SSME engine controller failed on launch day and was replaced a few days later.

Credit: NASA/Jim Grossmann.

(Photo Caption: Just to the left of United Space Alliance personnel, an SSME-era engine controller is exposed on the engine installed in the number three position on Orbiter Atlantis while at Launch Pad 39A for STS-135 in June, 2011. In the case of STS-135, a main fuel valve needed to be replaced; however, other replaceable components like the engine controller could also be serviced while the engine remained installed.  NASA will be performing this type of maintenance on SLS for one of the first times in the program.)

The first attempt to launch Orbiter Atlantis and the STS-43 Shuttle vehicle was scrubbed before dawn on July 24, 1991, when the primary computer, DCU A, failed while propellants were being used loaded into the External Tank. “During the countdown for the anticipated launch on July 24, 1991, the SSME 3 controller went to halt when an unrecoverable channel A parity error was detected (Flight Problem STS-43-E-1) prior to the start of liquid oxygen replenish,” the STS-43 Space Shuttle Mission Report said. “Data indicated that a hardware failure of the main engine controller had occurred.”

“As a result, the launch was scrubbed to allow replacement of the controller, and the launch was rescheduled for August 1, 1991. The failure analysis of the controller revealed a broken blind lap solder joint connection of the bit jumper to the half stack, which is not a generic design problem.”

According to contemporaneous Shuttle Status Reports issued by NASA Public Affairs at KSC in late July, 1991, after the launch was scrubbed and the External Tank was drained and inerted, access to the engine area for maintenance was established on July 26. The broken engine controller was removed, and a new one was installed on July 27, followed by testing to verify the new controller on July 28; the three-day countdown was started over from the top on July 29 for the next launch attempt on the morning of August 1.

By the time of the STS-43 launch, the Shuttle Program had ten years of ground and launch operations experience and decided immediately to replace the controller with another unit and launch while continuing to troubleshoot the broken unit in parallel offline.

Circumstances are different in the current situation. In this case, with the SLS Artemis 1 vehicle, although the DCU B power-up failure was observed before Thanksgiving and access to the area was already established, the controller was left in place on the vehicle while troubleshooting continued over the next month.

This is the first SLS vehicle NASA has ever fielded, and with less experience and practice with procedures for removing a controller on an engine while installed in the SLS Core Stage, it is uncertain how much longer it will take to install a new unit and complete the final checkouts of the engines and Core Stage in the VAB.

EGS and prime launch processing contractor Jacobs were in the middle of conducting PSETs on the SLS Core Stage and Solid Rocket Boosters when the controller issue came up. The Booster PSETs were completed, but a few of the Core Stage PSETs were deferred until after the engine controller is replaced and integrated testing moved onto other vehicle checkouts that could be performed independently of the Core Stage engines.

Preparations for removing and replacing the engine controller unit for engine 4 will take days, with removal projected no earlier than next week, just before New Year’s. The uncertainty about how long the engine controller replacement, retest, and other rework will take spills over into launch readiness forecasts for the vehicle and ground systems. SLS can only launch Orion on Artemis 1 when the Moon is in the right position in its orbit around the Earth, but launch readiness is a measure of when the hardware and the people are ready to go.

Credit: NASA.

(Photo Caption: A high-level diagram of the redundancy of the SSME controller from a 1980s NASA paper. The controller has two independent sets of computers, electronics, and power supplies. The new engine controllers that replaced the old 1970s units on the RS-25 have a similar, high-level organization for redundancy.)

This is the first time that NASA is getting an Orion/SLS vehicle ready to launch, so estimates of how long the work will take to complete are more tentative; the extra work on the engine controller now has to be factored into the forecasts/projections of the remaining schedule.

If NASA, its contractors, and broader organizations such as Range Safety were hypothetically all ready to fly in early March, that would be too late for the February period but would provide a few days of margin prior to the opening of the March period on March 12. However, if one or more of those organizations isn’t ready to support a launch until April, then even March won’t be possible regardless of the status of the spacecraft, rocket, and ground systems.

Finishing integrated testing before first rollout

Although time was lost due to the engine controller troubleshooting, EGS and Jacobs continued working through a series of integrated tests of the vehicle and ground-based launch support systems during December. As a part of the overall Integrated Test and Checkout (ITCO) series, the remaining standalone parts of Interface Verification Testing (IVT) were completed on the Interim Cryogenic Propulsion Stage (ICPS) followed by PSETs on the SLS in-space, second stage.

End-to-end testing of interfaces between Orion, ICPS, the SLS Boosters, and Core Stage, and vehicle to ground system interfaces, was also conducted in the first half of December. Another joint simulation between the launch team, flight control team, and Artemis 1 mission support teams around the country was also conducted on December 13 using emulators to create virtual problems for control team personnel to resolve in real-time.

The “big” end-to-end communications test, which also brought in NASA’s ground and in-space communication networks, started on December 15 and was scheduled to be completed over the weekend. Following that test, EGS and Jacobs were planning to conduct a Countdown Sequence Test (CST) early this week, the week of December 20.

“[The Countdown Sequence Test] is essentially the last two hours of launch countdown, which is very much like a terminal count sim,” Jeff Spaulding, Senior NASA Test Director (NTD), said in a recent interview with NASASpaceflight. “But that’s with the launch vehicle [instead of with emulators]. There will be better fidelity in some areas and not so much in others, but it’s going to be another good test for us.”

Credit: NASA/Isaac Watson.

(Photo Caption: The aft end of the Artemis 1 SLS vehicle is seen from the platform of the Mobile Launcher on August 16. Work to finish checkout testing of the Core Stage engine section, RS-25 engines, and configure them for launch will be one of the critical paths remaining on the schedule.)

Once the big end-to-end communications test and countdown sequence test are completed, only a few other integrated tests remain to complete before rolling out the vehicle to the launch pad for the first time. A Launch Release System (LRS) test has moved around on the schedule and was last projected to be conducted during the holiday fortnight, either the end of the week of December 20 or the week of December 27.

The first part of the final Flight Termination System (FTS) installation and checkout was the last big ITCO test on the schedule, which would leave work in and around the SLS Core Stage engine compartment as one of the critical paths in the remaining Artemis 1 pre-launch schedule. A flight readiness test (FRT) of the RS-25 engines was one of the Core Stage PSETs deferred until after the engine controller problem was resolved, along with a pre-flight checkout of the Core Stage thrust vector control (TVC) systems.

While troubleshooting the engine controller problem was going on, final Core Stage Main Propulsion System (MPS) modifications were completed. The clutch mechanism in all eight MPS prevalves was repaired inside the engine compartment during the Summer and Autumn after the Core Stage was stacked on the Mobile Launcher in June.

Core Stage prime contractor Boeing also installed resized hydrogen bleed system tubes and mounting brackets, and final testing and closeouts of the modifications were in work. Boeing is also finishing final pre-flight trims and painting of the thermal protection system (TPS) cork for the Core Stage engine section and boattail.

Preparations elsewhere around the vehicle were also scheduled to be completed during December, including closeouts of the ICPS and Launch Vehicle Stage Adapter (LVSA) and also configuring the forward skirt and intertank areas inside the Core Stage for rollout. The first vehicle rollout to the launch pad will be for a full countdown demonstration test called the Wet Dress Rehearsal (WDR); the vehicle will return to the VAB for final pre-flight maintenance and launch preparations.

Once final engine retesting and Core Stage checkouts are completed after the engine controller problem is resolved, the integrated operations team can configure the engine compartment for launch, which will be completed before the first vehicle roll out to the launch pad. The work will be very similar to closing out the Core Stage for its Green Run Hot-Fire tests conducted almost a year ago.

Internal access platforms will be broken down and removed, final inspections of MPS and engine hardware will be conducted, and the engine-mounted heatshield (EMHS) blankets that enclose the openings in the boattail for the engines will be installed. Installation of the flight TPS blankets was one of the last major closeout tasks for the engine section.

(Lead image credit: Stephen Marr for NSF)

Related Articles