Category Archives: Travel

A week ago, April 23rd, I was in Washington to see my first NASA press conference, held at the Newseum for the 25th anniversary of the launch of the Hubble Space Telescope (HST). Afterward, I and the other attendees at the NASA Social were taken out to NASA’s Goddard Space Flight Center. We first learned about the James Webb Space Telescope (JWST), then got some Hubble Space Telescope (HST) history and saw the Satellite Servicing Capabilities Office (SSCO).

We got a peek in through the window into the HST mission control room. Commands are sent up to HST after being carefully checked and double checked to make sure they don’t accidentally instruct the spacecraft to do something stupid and/or fatal.

For many years all of the command consoles were staffed 24/7/365. With the upgrades both on HST and on the ground, many of the operations no longer require constant monitoring. There is an extensive system in place to alert Goddard staff when anything goes “off nominal.” Minor issue will result in a text message or email, more critical problems are met with more aggressive alerts.

In the HST mission operations center we saw where the data from Hubble is coming down and the astronomical observations were monitored. There was data everywhere on dozens of big monitors — heaven of the little kid inside me who still likes to see buttons pushed and lights flashing.

In particular, I asked about the “pickle” diagram, visible on the center monitor closest to us. It shows how Hubble is using its three Wide Field Detectors (WFDs) to track stars in the current field of view. With two detectors tracking stars, Hubble can maintain tracking accuracy to less than 7 milliarcseconds per day.

How good is that? A circle around the sky is 360 degrees. Each of those degrees is split into 60 arcseconds. Now split each of those arcseconds into 1,000 bits. That’s a milliarcsecond. In the real world, 7 milliarcseconds is the size of a dime on the Washington Monument in DC as seen from the Empire State Building in New York City.

The HST is the size of a school bus, weighs twelve tons, and is floating weightless in space. How do they keep pointing it that accurately? (Assuming it’s not some serious black magic.)

These little guys are key. Some of the most accurate gyroscopes ever made are at the heart of these rate sensor arrays (RSAs). Hubble has six of these assemblies, two each on each of the three axes. Usually they run just three at a time, one from each set.

These gyros and RSAs are among the hardest working systems on Hubble and were replaced on three of the five servicing missions. In fact, the failures of gyros were the reason that the third servicing mission was broken into two separate Space Shuttle missions. Four of the six RSAs had failed by November 1999, putting Hubble into a hibernation or safe mode. With less than three working gyros, Hubble could have started drifting and tumbling, making it difficult or impossible to capture by the astronauts on the next shuttle servicing mission. So the planned June 2000 mission was split, with STS-103 going up to replace the RSAs (and other things, such as the primary onboard computer) in December 1999, while STS-109 went up in March 2002.

Hubble doesn’t use thrusters or jets to control its movements in space. For one thing, the gasses used (primarily hydrazine) are very nasty to have around delicate optical instruments. In addition, once the fuel is gone, so is your ability to control your attitude. (Remember the bit yesterday about RRM?) So Hubble is controlled by four massive Reaction Control Wheels (RCWs) which move the spacecraft by gyroscopic action. (For a lot of detail and technical minutia on the Hubble guidance system, see this NASA technical paper.) In high school, did you ever do the experiment where you sit on a bar stool and hold a spinning bicycle wheel, tilting the wheel to make you spin around? That’s how Hubble moves, but with wheels that weigh hundreds of pounds each and are precision machined.

In addition to the gyroscopes in the RSAs, Hubble also senses its position in space using large, sensitive Charge Coupling Devices (CCDs), the same kind of sensors that are at the heart of your digital cameras. Two CCDs are built into each WFD and there are three WFDs on Hubble. (Remember the “pickle” diagram above?) As long as two of the WFDs have a star to track in their field of view, the WFDs and RSAs combined can give Hubble that 7 milliaarcsecond per day guidance.

That’s some seriously awesome engineering there!

Originally HST was launched with three Engineering & Science Tape Recorders (ESTRs) for recording and playing back data. Designed in the 1980s, the ESTRs were reel-to-reel tape assemblies. Data was written and retrieved sequentially and when the tape broke or jammed (as this one did) the ESTR was useless. These ESTR units held 1.2 gigabites of data, state of the art at the time.

The ESTRs were one of the components of HST which were designed from the beginning to be upgraded as technology advanced. They were replaced on the second and third servicing missions with solid state recording units. The upgraded units are like huge, radiation hardened memory sticks. Not only do they hold over ten times as much data as an ESTR, but they can also be read instantaneously instead of sequentially, and sections of memory which become damaged (perhaps by a radiation hit) can be bypassed, leaving the rest of the unit still functioning.

For those interested, I believe these three pieces of hardware were all flight-flown, coming back down from HST after being removed and replaced during a servicing mission.

Dr. Marc Kuchner wants your help in looking for targets for HST, and later for JWST. There’s a crowd sourced (Zooniverse) project called “Disk Detective” in which you can do real, honest-to-god science in your spare time.

The short version is that stars with planets are stars surrounded by dust due to the planets, asteroids, comets, and other assorted objects crashing into each other every so often. Because of celestial mechanics and conservation of angular momentum, the dust tends to flatten into a disc or ring. Conversely, we’re finding that when we see a star with a dust disc, we often find planets there.

It’s time consuming and inefficient to have have telescopes like Hubble look at every single star looking for planets, so we would like to improve our odds and find another way to narrow the search. A key tool here is the Wide-field Infrared Survey Explorer (WISE) mission, a space-based telescope which looks at big chunks of the sky at once, but at lower resolution than telescopes such as Hubble.

Dust is bright in the infrared part of the spectrum since it absorbs starlight and re-radiates the energy in the infrared. Therefore, if we look at a star in WISE images and find a dust disk, that’s a good candidate to look at more closely using Hubble or another big telescope.

But how do you look at all of the stars in the WISE images? Computers? Not really. It turns out that computers aren’t that good at examining images and “looking” for certain telltale characteristics. But the human eye and brain are pretty good at that. But there are billions of stars. So what if a whole lot of people each looked at a few dozen or a few thousand stars each?

That’s how Disk Detective works. If you go to the site it will show you a series of ten images for a single star, each image in a different wavelength. You can flip through the images as often as you want, then answer six simple questions about the images, such as if the object is moving. This only takes a minute or two, you submit your answers, and go on to the next image.

If multiple people independently judge a particular image to be a good candidate to have a dust disc and possibly planets, then the pros can take a look at it, possibly moving the observations up to a much bigger telescope, or even up to Hubble. It’s a piece of cake, and beats the hell out of playing Solitaire on your computer! Give it a try, maybe you’ll be the one who finds another new planet.

Finally, we got to talk to two of the engineers who helped to design and built the tools needed to perform the delicate tasks of the Hubble servicing missions. (Again, I apologize for not getting their names – if anyone can fill in the blanks for me in the comments, I would appreciate it.)

Here we see a panel that was designed to go over a series of over a hundred screws and bolts that had to be removed, without allowing any of them to get away or fly loose. A single nut, bolt, washer, or screw that escaped and drifted into the telescope could cause an electrical short or damage a lens or mirror, causing an incredible amount of damage. And remember, this work was being by astronauts wearing spacesuits with very limited mobility and dexterity, floating weightless, often with poor or little lighting. The astronauts described it as being like performing brain surgery while blindfolded and wearing oven mitts.

In order to safely open panels and instruments that were never designed to be opened while simultaneously preventing any loose bits from drifting away, some very complex tools were designed and built. For example, the blue panel above was attached in place with bolts screwed in using the big handles (easy to use with gloves). The clear holes lined up over the screws that needed to be removed, and the holes in the plastic were just big enough to allow a screwdriver tip or other tool to get through while still being small enough to not let the screw escape.

Another problem was dealing with sharp edges, knives, and cutting tools. In a pressure suit surrounded by hard vacuum, NASA doesn’t want astronauts handling a knife or anything sharp. When this top panel had to be removed with a whole series of cuts, this tool was built to screw on (again, big handles, easy to use in stiff spacesuit gloves), cut the top of the compartment off with blades that were recessed and not a danger to the spacesuit or gloves, then pull off with all of the loose bits captured inside.

Other new tools had be developed to make the Hubble repairs feasible. For example, the standard Pistol Grip Tool (PGT) which we saw in the  SSCO is used to remove and install screws and bolts – but it’s really slow. For something like the job above with the blue panel, which had 100+ screws, that would take forever. So other faster, more lightweight tools were developed for the Hubble repairs. In addition, since they would be working in the dark a lot, let’s put some LED lights on there so we can see what we’re doing while looking out through that spacesuit helmet. Right?

Yeah, this is the one they developed (I believe it was flight-flown), and yes, they let me hold it.

Geek joy!!

FYI, it’s heavy and awkward to hold even there in the auditorium using just my bare hands. We can’t ever give enough credit to the astronauts who pulled off five astonishingly successful service missions, giving us one of the landmark scientific instruments of our generation.

Flip through the images that we’ve gotten from Hubble.

Go see one of the IMAX 3D movies about Hubble, or watch the PBS Nova special from last week.

Read about some of the discoveries that Hubble has allowed us to make in the last 25 years.

Study how the initial problems with the Hubble optics were overcome. It’s a classic study in recovery management, how an initial critical error, particularly a very public and very expensive one, can be faced head on and resolved, leading to one milestone achievement after another.

That’s why we’re celebrating twenty-five years of Hubble’s observations.

That’s why we’re looking forward to many more years of Hubble’s observations.

It was a great NASA Social!

After rubbing elbows with Space Shuttle mission commanders (sorry, still squeeing) at last Thursday’s NASA press conference for the 25th anniversary of the launch of the Hubble Space Telescope (HST), the attendees at the NASA Social were taken out to NASA’s Goddard Space Flight Center where we learned about the James Webb Space Telescope (JWST) and saw where it was being assembled.

First of all, I want to apologize to all of the wonderful researchers and engineers at Goddard whose names I did not get. They all deserve all of the credit in the world for the amazing work they’re doing, and have been doing for decades. If anyone from Goddard or any fellow NASA Social attendees can fill in any names that I missed, I would greatly appreciate a note in the comments with the information.

This is a 1:5 scale model of HST. The see-through panel at the back end is there to show where the instruments are. HST was visited five times by Space Shuttle crews who replaced and upgraded control systems, instruments, cameras, and optics. Any one of those visits could have ended disastrously. One loose screw, one bolt that wouldn’t come loose, and HST could have been permanently crippled.

The fact that the five servicing missions were 100% successful despite glitches and problems is one of the reasons I like crewed missions. Things never go the way they’re planned and humans are highly adaptable. Check out the PBS’s Nova program, “Hubble’s Amazing Rescue” for a great description of what was at stake, how nearly impossible to get it right was, and how spectacularly triumphant the results were.

Ben Reed runs the Satellite Servicing Capabilities Office (SSCO) at Goddard, where they design and build tools and robots to work in space. Here he’s holding a Pistol Grip Tool (PGT) which is used to screw and unscrew nuts, bolts, and screws in microgravity. This particular one is actually from the Neutral Buoyancy Laboratory (NBL) where astronauts train for their space walks.

One of the projects being worked on at the SSCO is the Robotic Refueling Mission (RRM). I remember seeing these experiments being run  on the International Space Station (ISS) a while back. This test panel is identical to one flown to Low Earth Orbit (LEO), as is the next one shown below.

Thousands of satellites have been sent to orbit and then become useless as they aged. Sometimes a critical piece fails, maybe a power relay or the onboard computer. Sometimes satellites get hit by meteorites or other orbital debris. More often than not though, the satellite fails when it runs out of maneuvering fuel.

Due to various factors such as atmospheric drag and gravitational pull from the Sun, Moon, and other planets, satellites tend to not stay where you put them. They’ll all use some sort of small thrusters to “station keep.” When that fuel runs out, they drift, lose attitude control , and become useless. Oh, and 99.9999% of these satellites were never intended to be serviced or refueled, so when they’re dead, they’re dead.

Unless…

Let’s say you built a robot. A very clever robot with a bunch of nifty tools. Maybe the robot is autonomous, or remote controlled from the ground, or a little of both. (Yeah, I would send a crewed mission since humans are incredibly dexterous and clever monkeys, but that’s just me.) Maybe these nifty tools and clever programming would allow securing wires to be snipped, locked on bolts to be unlocked, thermal blankets to be peeled back, and sealed access ports to be unsealed on dead satellites.

Now you need to test your robot. It’s expensive (and a whole different engineering problem) to accurately locate the broken satellite, rendezvous, and attach your refueling robot to it. Especially when you don’t know yet if you’ll actually be able to do the undoable when you get there. So you want to test the breaking and entering and refueling before you spend all of that other money on finding and rendezvousing and capturing. This is where the RRM and ISS come in.

These panels have fueling ports and connections that are the same as those used on numerous commercial satellites. The test is to see if controllers on the ground can remotely use the tools built into the robot to do the job.

The RRM mission on ISS uses the Canadian-built Dextre two-armed robot along with tools built for this task. Dextre can be held at the end of Canadarm2 or it can be docked at one of the Mobile Base Stations on ISS. Seen here is a heavy-duty version of a remotely controlled robot similar to Dextre. (Dextre’s tough, but lightweight and designed to work in microgravity. This one has to be bigger and heavier to work at 1G.)

So far the RRM tests on ISS have gone very well. There have been no “show stoppers” in proving that a remotely controlled robot can break into a satellite with empty fuel tanks, refill the tank, reclose the port, and release the satellite.

The issue, as it always is, is money. A robotic refueling mission like this might cost $500M or more, plus the initial development costs. (Please note, all of these numbers are Wild Ass Guesses (WAGs) and yes, I am trying to see how many Three Letter Acronyms (TLAs) that I can put into this article.) Would it make economic sense? That was one of my questions for Ben Reed.

The answer is, not surprisingly, “It depends.” It’s not a single calculation but a spectrum. On the low end, if you have a cubesat that only cost $50K to build & launch, spending hundreds of millions of dollars to refuel it would be stupid. On the high end, if you spend $7.998B to build and launch JWST, perhaps a couple hundred million dollars to double or triple its lifetime might make sense.

To be quite clear — NO ONE at Goddard said ANYTHING about possibly refueling JWST. In fact, whenever the concept got brought up, it was politely but firmly squashed. JWST has been designed from the get-go to be a “fire & forget” spacecraft with no possibility of being repaired or refueled. I’m just speculating wildly here.

But that being said, remember, it’s an economic spectrum. Even if you can’t refuel JWST, what about if you’ve got a $1B+ weather satellite or communication satellite. If your choice is to build & launch another $1B+ satellite or try spending $250M to refuel the old one (again, WAGs!), which do you choose?

More to the point, if you build your robot with lots of different capabilities and a big tank of fuel, maybe you can refuel ten satellites. Or fifteen. Or thirty. Now your cost per satellite is maybe $20M each. So if you have a perfectly good, functioning $1B+ satellite that’s out of fuel, do you build another, or risk $20M or so to double its lifespan? If the technology’s there, the answer seems obvious.

The SSCO isn’t developing this technology just because NASA necessarily has any plans to use it. The key word in SSCO is “Capability.” Yes, NASA would like to be capable of doing this if they need to, but it could be another NASA spinoff that saves billions of dollars.

That’s not all the SSCO is working on. Here you can see “test rocks” of different sizes, weight, and composition. Some are hard as a rock, some are more like clumps of sand barely holding together. Some are metallic, some not.

These test rocks are being used to see how current robots and their manipulators (“hands,” if you will) can deal with different objects they might encounter on the surface of a comet or asteroid. (Remember OSIRIS-REx from yesterday?) We don’t know what we’ll find on the surface of any given object (that’s the reason we’re going, to find out – right?) so the robot spacecraft we design and build have to be prepared for a range of options.

Also on the left in this picture you can see a grey-white object. I believe this is a model of a device currently on ISS in the Japanese Experiment Module (JEM, better known as Kibo) and used to launch cubesats from ISS. What I want you to see however is that this has been made by “additive printing,” otherwise known as “3D printing.”

3D printing is going to be HUGE! For example, in a case like this, you can prototype an object for hundreds or  thousands of dollars instead of spending hundreds of thousand dollars or more. Then you can test it, tweak it, break it, refine it, and print another. You work out the bugs using the cheaper 3D printing route, then you build the final product the old fashioned way.

I’m sure I’ll be ranting at length about 3D printing here at some point.

In the other corner of SSCO they’re working on the other part of the robot refueling problem – rendezvous and capture of a satellite that might not be under control. If the fuel’s gone (or there’s some other problem) it could be spinning, rolling, tumbling — or doing all three at once. The more it’s moving, the harder it’s going to be to latch onto.

This mockup of a satellite panel is mounted on a rig that can move it through a wide range of motion. The (again, heavy duty, designed for 1G, not microgravity) robot arm on the right is being used to develop software and techniques to learn how to approach and grapple a tumbling satellite, or to figure out when it can’t be done and back off.

These three tools were on display in SSCO, demonstrating how much smaller, more capable, and more elegant our tools for working on orbit have become.

The big blue on on the left was used in the rescue & repair of the Solar Max satellite in 1984. The center tool is that PGT used by shuttle astronauts and now by ISS astronauts on their space walks. The far right tool is one of the tools that was used by Dextre during one of the RRM experiments.

Again, I apologize for not getting the name of this woman who did such a great job of walking us through many of the discovery highlights from HST. From planets, to planetary nebula, to galaxies, to black holes, Hubble has spent the last twenty-five years revolutionizing the way we look at the universe around us.

Getting observing time on Hubble is non-trivial. There are, of course, far, far more requests for observing time than the telescope can perform. Observing time is allocated by a committee that picks the most worthy proposals.

In addition, there is a certain amount of time that is set aside as “director’s discretionary” time. One thing that this time is used for is Hubble observing time when something transitory or unexpected happens. Maybe there’s a supernova, or a comet’s going to slam into Jupiter. With discretionary time, these objects and events can be observed without kicking anyone else off the schedule.

In fact, the landmark “Deep Field” image by HST was done using director’s discretionary time. After all, it was quite literally a blank, empty piece of the sky, without any known stars or other objects in it. Why would the committee give anyone ten continuous days of observing time to look at “nothing”? Yet this image, covering the area the size of a dime as seen from seventy-five feet (let that sink in) contains over 1,500 galaxies.

1,500 galaxies, each with billions of stars, many of them with planets, almost certainly some of the planets containing life, almost certainly some of that life reaching a level of intelligence equal to or better than ours. All that from an “empty” area that small.

The universe is a big place!

Next, we see how HST is operated every day, how it points so stinkin’ accurately, and how it got fixed and upgraded on the servicing missions.