Bjorn’s Corner: Flight tests

By Bjorn Fehrm

By Bjorn Fehrm

19 February 2016, ©. Leeham Co: Last week I described how Mitsubishi Aircraft Corporation (MAC) issued a press release on Christmas day communicating MAC would be doing structural reinforcements on their test airframes before continuing flight testing. MAC was perhaps overzealous when informing the world that they would do minor reinforcements to two ribs and a few stub spars in order to pass Ultimate strength tests for the aircraft.

I rightfully thought this is the Japanese culture at play; there must not be a big problem behind it.

There was one more area of that press release that intrigued me. Here what it said: “The first flight and the subsequent flight tests have confirmed the basic characteristics to be satisfactory. However, we also have recognized several issues as we attempt to accelerate our development.”

Time to decrypt this as well and compare to what has become standard industry practice.

First test flights

The released sentences smelled of troubled flights. At my January meeting with MAC’s Marketing Director, Hideyuki Kamiya, I asked him what was the meaning of this part. He answered the test flights were normal, nothing special was discovered and they were satisfactory. I told him “once again your language is other than your colleagues.” Instead of dwelling on MAC’s language, let’s look at what MAC’s colleague OEM’s typically say.

“It was a joy to fly, it really flies well,” “The aircraft flew beautifully,” “It was a perfect flight, we only had one small problem,” and so on. Behind those upbeat expressions after the first flights, one cannot really read anything else than that the flight was not interrupted for some reason. I know that during several of these flights, there were real problems found, some of which merit to stop the flight and going home in emergency mode.

Today’s aircraft are extremely well analyzed and tested aerodynamically before first flights. Detecting major flaws in how the aircraft flies as an aircraft is less probable. On the other hand, the systems side and especially avionics and flight controls have grown in complexity to a level that makes test crews really nervous.

A modern airliner has between 10 million to 15 million lines of software code in its systems. The likelihood of some of that software going into wrong modes or outright stopping to work is high. One of the calmest first flights had the air-conditioning flying to New York when the aircraft was circling the Pyrenees. Another had the whole electrical system shutting down for a while.

Two very critical system areas are the Fly-By-Wire (FBW) and the computerized engine control, the FADEC. For first flights, the FBW is virtually taken out of the equation. Civil Airliners have to be naturally stable also when equipped with FBW. This means the aircraft will sort itself out of troubled flight situations without FBW intelligence and with minor pilot input.

Consequently the FBW is set to work as if the contact with the rudders was made with steel wires, the so-called direct mode. It means a minimum of involvement from the complex FBW software and therefore a minimum chance that things could go wrong. It’s also the preferred mode from the test pilot’s viewpoint for first flights. First flights are when you explore the base behaviour of the aircraft.

You don’t want an FBW or autopilot which adds or subtracts to or even masks the base aircraft’s quirks during such flights. It’s the only way to add the human input to all the data when the flight gets evaluated after flight. FBW intelligence is added much later in the flight test programs.

For the engine’s FADEC, it’s even worse. You don’t fly it first on the test aircraft. The engines are test flown on the engine manufacturers own four engine test aircraft, a Boeing 747 or loaned A340/A380. One of the four engines is replaced or there is a special fifth pylon added for the test engine. Only with such a setup can a complete FADEC run-away be handled. You stop the engine by shutting its fuel supply in worst case and fly home on the remaining engines.

Only when the engine and its computerized control are rock solid is it allowed to fly on the airliner test aircraft. And even then it can go wrong, as it did for the half military-half civil Airbus A400M. A new FADEC software version which was tested to be solid did not install correctly on the target aircraft. The consequences in the ensuing test flight were fatal.

I would say the least worry I would have as a test pilot today would be how the base aircraft flies. At least as long as one follows the mother’s advice “Fly slow and low.” Once one get to high flight levels and transonic speed it gets more interesting.

11 Comments on “Bjorn’s Corner: Flight tests

  1. Historically Aircraft pitch trim has been hard to predict perfectly correct in wind tunnel tests especially for low wing fighters.
    Flow around the aft body is not so easy either, just look at all the 737NG voretx generators they are now cleaning up for the 737MAX.
    Air conditioning coolers and APU air inlets have also shown hard to make perfect. Noice from landing gear turbulent flow is another challange that maybe the A320neo has solved better. Flow under the wing at low speed high Power thrust reverse can be added to hard to predict items list. Another is the Nose Wheel spray getting sucked into the Engines. Even the top of climb Engine fan blade flutter is hard to discover for a particular Engine/Aircraft combination. This is just the tip of the aero mountain of possible pitfalls needed to be tested and resolved.

  2. Very interesting article as ever.I don’t think MACs problem in communication was anything to do with the level of honesty, I know very little about test flying but I always take “its the sweetest bird I have ever flown”thing with a very large pinch of salt especially if its a government programme which could be under threat.It’s much more to do with the language used. In any industry a release like that would cause alarm, it just reeks of disaster.

  3. As I pilot when I see the it was perfect statement, I snort.

    I believe the 747-8 had some serious turbulence issues around its gear door areas.

    While I never did test flights, I would also say I would not want anything automatic turned on simply from the fact that it might do something completely bizarre.

    Per Bjorn, find out what the aircraft does and does not do, then only at altitude turn on single things and test before next.

  4. I do have mixed take on the Japanese culture.

    From what the reports at Yuasa were on the 787 battery, i.e. the manufacture facility was filthy and the prosecution process was grossly deficient in quality control, there seems to be two sides to that.

    They made no effort to clean up the facility or correct defects when the batteries started melting down and apparently denial was the operative word.

      • The last 15 years of decline in Japan has decimated their small to medium businesses, they are still run like the 60s-70s with the same equipment as their is no money for new investment. And any new factories by the international firms have been in China or similar.

    • The batteries thing was really strange.
      Same design prismatic cells from the same line seem to have performed unremarkably in previous use cases.

      IMU the Yuasa people must have been convinced that
      the apparent ( and longstanding ) issue was a case of misuse by Boeing ( and the charger people ). Doing (dis)charge cycles up to final limits reduces available lifetime massively and any tolerance miss will kill the cells for sure.
      ( flightblogger at the time had an article on early battery issues even before first flight that could require an ameliorative chemistry change )
      IMHO the filthy workplace allegation is a distraction.
      good for blame but not good for an explanation.

      Beyond the Boom Box the final fix on Boeings side included reducing max and min charge margins.
      ( What they should have done from day one.)

  5. So to decode what they found was that the basic handling characteristics were OK but they now will try to turn on more advanced software. Again sounds like nothing to worry about. Typically the software is “on” but not connected to anything. That way you can analyze the output and see if everything seems OK. Typically they fly a few dozen times to get an envelope with basic handling then start turning things on with a cutout switch just in case.

  6. Boeing does not fly in “direct” mode on first flights. The plane is launched with full FBW authority

  7. Well, gain enough altitude to get back to the runway, if no problem initially gain enough altitude to recover from upset.

    Otherwise good article, thankyou.

    • Which reminds me of the engine fire warning logic on the first 767s.

      Limited warning until something like 400 feet (often an initial flap-retraction altitude), as Job 1 is to get safely into the air.

      At that altitude crew have some time to deal with it, and a minimum safe altitude to return to the runway.

      (Digressing, that requires circling for most airports, SEA-BFI-PAE and YVR-YYJ/Abbotsford being exceptions with a runway only a few tens of miles ahead. 😉

Leave a Reply

Your email address will not be published.