How did you become a test engineer?
I come from an aviation family. My grandfather was a World War II fighter pilot. I have an uncle that was an airline pilot, a cousin who flew for the Air Force, and my brother works for the FAA, so it’s kind of in the blood. I went to college in Spokane, which is on the other side of Washington state, and I began working for Boeing less than a month after graduating from college. I started out as a flight deck engineer, working on the 777 pilot seats to ensure they conformed to the new 16g rules that were coming into effect, with the 777 the first Boeing aircraft to be fitted with 16g-compliant seating. I participated in some development testing and supplier qualification testing with crash dummies. It was quite an experience – actually getting to stand next to a crash dummy and watching what actually happens during a test is a lot of fun. Then almost 20 years ago, a good friend of mine who was working as a flight test instrumentation engineer suggested that I come over and join the group – I’ve been here ever since!
What was the most important lesson of your early career?
I was only 22 years old when I joined Boeing. I remember working my first day with engineers from the 707 and Apollo programs. I am thankful to them for creating an environment that expects and delivers great products. Their absolute focus on safety made the greatest impression on me. My first flight test program was the 737 Next Generation. In all the pre-flight meetings, the first conversation was always about safety, and every single conversation that followed would examine each particular maneuver and how prepared we were for it. I remember sitting on the aircraft at my station by the instrumentation rack and just feeling extremely safe. In everything we do, we are trying to be safe and we are trying to make a safe product, and there’s a lot of great people dedicated to making that happen.
Describe your current role.
Every day in flight test, we ensure Boeing airplanes are the safest and most efficient to fly. My team works with design engineers to collect flight test data that they can use to continually improve Boeing products.
It can be challenging work. Some days it means diligently checking the performance of thousands of airplane system components while at cruising altitude. Other days it means monitoring systems while the airplane flies in ways that would spill a lot of coffee if paying customers were on board.
If the aircraft doesn’t already have the measurement equipment on board, my team can usually get it – it’s just a question of time and planning. In fact today’s technology allows us to collect and analyze hundreds of times more data than when I started at Boeing 23 years ago – so much data, that we remove all passenger accommodation to make space for our equipment and build our own onboard intranet.
And what are you working on right now?
Right now we’re starting the transition from the build phase into the testing phase for the 737 MAX. We have two aircraft in final assembly and I’m the lead instrumentation engineer on the first of these. What we are doing daily is primarily functional checkouts, where we make sure that every single one of our channels is functional, calibrated and operational in the way it is supposed to be.
My job is to coordinate these onboard calibrations while liaising with our manufacturing organization to ensure the airplane continues to get built at the same time. We’re getting the aircraft ready for roll-out this year and flying next year.
What are some of your day-to-day tasks?
Currently we’re doing calibrations of primary control surfaces. We’ll put additional sensors on an elevator tab, for example, and then we will define our own calibration standard for that elevator tab and sweep it through its range of motion to generate a transfer function that we then use to correlate the voltage that we’re getting out of our transducer with the actual position of that surface. We’ll do the same for the aileron wheel position – putting in higher accuracy, higher sample rate measurement sensors, and then sweeping the control column through its range of motion to get a transfer function for calibration purposes.
We’re also working with the production team on the functional check-out of the avionics, which features quite a few ARINC 429 databus components, to make sure that our connection is functional and safe. So, when we start preflight testing, we will be recording all the data that’s required to show that the aircraft is ready for first flight.
There are four aircraft in the 737 MAX 8 flight test program, and each one of them will have a slightly different set of instrumentation with different channels depending on what the requirements are for testing. The first aircraft will be used more for controls testing, while the second will do more engine testing and the third will focus on systems testing. You can’t fit all
the instrumentation on one aircraft efficiently, so you spread it out between the different flight test vehicles.
What are the key challenges presented by the 737 MAX?
The 737 is a good and established platform that is well understood from a test and evaluation perspective. However, the main challenge is instrumenting anything that’s new on the aircraft, such as the high-definition (HD) video display system in the flight deck, for example. In response, we’ve had to adapt our instrumentation system to be able to record HD video. The challenge is to work out how best to instrument that system and provide the data, and then get that data from that system onto the recorder, so our team can examine the data after the test or even view it live during the test. That can be challenging, but so far it’s looking good – we’re pretty pleased with how things are going on this program to date.
And by today’s aircraft in general?
We all have laptop computers that we use all the time and we think that all the computers in the world can talk to all the other computers in the world. In reality, every single one of them speaks a different language. I liken it to the Tower of Babel. Every single model has slightly different hardware, using slightly different languages, and has software that allows it to interpret data from other computers. So the big challenge for any instrumentation engineer is how to record all of this different data from different computer models onto a single recorder while maintaining time synchronization. You could buy every manufacturer’s recorder, but you would have to change media for multiple systems during the flight and it becomes unmanageable. We record all of our data onto a single, really high bandwidth recorder – but to do that, everything has to be in the required format.
For example, a new display system may generate data in a format for a new avionics standard of fiber-optic HD video. We would then have to define a mechanical interface to connect the new fiber-optic bus with our computers and then design software to interpret that data and translate it into a format that’s acceptable for our recorder.
So every time we put a new avionics system on board an aircraft, we go through this challenge. And it’s not just avionics. The transducer and signal conditioning manufacturers within our own test equipment industry produce hardware and software that all ‘talk’ slightly differently. As a result, you often can’t just go and put a new sensor on the wing of an aircraft without first designing a system to interpret that data and get it on the recorder.
What are the current capabilities of the data recorders you use?
I don’t know the exact numbers, but they record a lot of data. Our current recorders are capable of recording more than 400Mbps, and our media holds a terabyte.
Is data collection easier or harder than it once was?
Distributed signal conditioning architecture has made data collection much easier. When I started in flight test, if you had a particular parameter that you needed to record, you would have to run a wire from the relevant sensor, which could be anywhere on the aircraft, all the way into the cabin for signal conditioning on individual cards mounted in units in the racks.
For example, if you had an accelerometer on the wing tip, you would have to run the wire all the way down the leading edge of the wing, through a wing-to-body penetration, into the cabin, and to a rack-mounted multiplexer full of individual accelerometer cards for every sensor.
However, if we want to add an accelerometer out on the wing tip today, we just run the wire to a small piece of equipment mounted on the wing’s leading edge, which does the conversion from analog to digital right there on the wing. We can collect multiple measurements in that one spot. All we need to run from there is a power cable and a communications cable to the cabin. We can build communications networks exterior to the pressure vessel of the aircraft, with smaller penetrations through the side of the body.
My first flight test job was on the very first 737-700 some 19 years ago, and in order to run all those wires out from the wing into the fuselage, we had several 3.5in-diameter penetrations on each side of the aircraft. Now, we can get away with penetrations that are 1.5in in diameter as they only have to take a single bundle. This makes it easier and cheaper, and we can do functional checking of airplane components before they are installed in the aircraft, saving a lot of time. A distributed data system makes things a lot simpler, and we get a lot more data.
Modern data recorders are also far more advanced. For some of the programs just before I joined flight test, we had relatively limited bandwidth, so you would have to decide what measurements you were going to record for a particular test, because you didn’t have the bandwidth to record all of them, all of the time. Every day, our job was to go out and reconfigure the data systems to record the measurements that were required for that particular test. Hence our overnight pre-flights would focus on reconfiguring the data system and making sure all the measurements worked before we could go and fly. We now spend a lot less time reconfiguring the data system due to the greater bandwidth of the recorders. The computers have made it easier – not only can we record more data at better sample rates, it’s also easier to do so on a daily basis.
Has anything got harder?
Trouble-shooting digital data systems can be quite an opaque task – it’s hard to see where you are having communications problems on a computer network. It requires a new set of tools and a new set of training. We need digital network communications experts in our crew, which is new. Trouble-shooting an Ethernet bus can be tricky because it could be either the hardware or the software that’s running a particular box that has a bug in it – determining where the issues are is a lot more difficult than it used to be.
What happens to all the data?
It is recorded onto a pair of recorders in the test instrumentation racks – we use two recorders for redundancy purposes. We monitor the data in flight via an onboard suite of data monitoring tools to validate the test conditions, and then we download that data into an archive after the flight. Once it’s in the archive, various Boeing company personnel can make data requests and examine what happened on each day’s flight.
The recorders were tape-based when I started, but now they are solid-state media. They are taken to a special archive computer facility for storage and access by Boeing engineers.
We also use telemetry, but we tend to limit this to safety-critical flights where we have an extremely limited crew, as telemetry limits where the aircraft can fly – you have to maintain a certain distance from base to stay within the available telemetry range.
How much data is captured during a typical flight test?
It will depend on the aircraft. If it’s a brand-new digital aircraft like the 787, where a lot more digital measurements can be recorded, the data rates are pretty large, more than 200Mbps – we can fill up a recorder in about eight hours. But airplanes with fewer digital channels and fewer required analog measurements will be significantly less.
What is one of the hardest tests to instrument for?
Load survey testing is one of the hardest, as we have to put pressure sensors over the entire surface of the wing. This kind of testing is challenging because of the sheer quantity of pressure measurements being taken and their location on the aircraft. It can be in the range of a thousand individual pressure measurements spread over the wing to measure pressure distribution, which can be challenging. It takes a lot of pre-planning and organization.
What’s the most rewarding aspect of your job?
The most rewarding aspect for me is working in a real team environment. Every day, we get together with the pilots, test ops and the mechanics that maintain the aircraft to plan testing. Then we go and test the aircraft. It’s great to be part of such a highly qualified team, where everybody is focused on safety and making the product better by going out and testing it and finding its limits. I find that incredibly rewarding.
How do you see your role changing in the future?
Fundamentally, it’s still just installing a sensor, changing the signal from the sensor into a digital signal, and then recording that signal. What has changed is the digital technology. The increased bandwidth, distributed signal conditioning and component miniaturization continues to develop at a rapid rate. The data system racks that we put on the aircraft today are nowhere near as full as they used to be when I started here. We’ve seen less hardware installed on the aircraft over time and I think that trend will continue.
Anthony James is editorial director at UKIP Media & Events Ltd, publisher of Aerospace Testing International
December 8, 2015