Re: Peruvian 757 crash

Date:         23 Nov 96 03:36:23 
From:         ethan@cs.wisc.edu
Organization: University of Wisconsin, Madison
References:   1
Followups:    1
Next article
View raw article
  or MIME structure

In article <airliners.1996.2270@ohare.Chicago.COM>,
Jonathan Griffitts  <jcg@qadas.com> wrote:
>I've been lurking, waiting to see any more information on the Peruvian
>757 crash a couple of weeks ago.  All I've seen so far is very sketchy
>and self-contradictory information from the newspaper.
>...
>This sounds like it COULD be the ultimate glass-cockpit horror story:
>they suffered an electrical or electronic failure that was so complete
>that it left the plane uncontrollable.  Is this is the event that the
>glass-cockpit opponents have been predicting?

Speaking of problems with glass cockpits, I was looking through the
NTSB database at http://www.ntsb.gov/Aviation/months.htm (someone I
know had a C182 engine blow up and I wanted to see if it was
listed) and found this:

:   NTSB Identification: NYC96IA116
:
:   Scheduled 14 CFR 129 operation of MARTINAIR HOLLAND (D.B.A. MARTINAIR)
:   Incident occurred MAY-28-96 at BOSTON, MA
:   Aircraft: Boeing 767-31AER, registration: PHMCH
:   Injuries: 202 Uninjured.
:
:   On May 28, 1996, at 1421 eastern daylight time, a Boeing 767-31AER,
:   PH-MCH, operated by Martinair Holland, as flight 631 received minor
:   damage during an unscheduled landing at Logan Airport, Boston,
:   Massachusetts. There were no injuries to the occupants, and visual
:   meteorological conditions prevailed. The flight which was destined for
:   Orlando, Florida, had departed Schiphol Airport, Amsterdam, The
:   Netherlands, at 0649, and was operated under 14 CFR 129. The flight
:   crew reported that they had received several false system advisories
:   during the flight. The advisories would appear and then disappear
:   shortly thereafter, with no corrective action being taken. There was
:   no evidence that the actual airplane systems were being affected.
:   These advisories started shortly after the airplane had reached cruise
:   altitude, and continued at an intermittent rate throughout the flight.
:   In addition there were multiple uncommanded disconnects of the
:   auto-pilot. The transponder code window would suddenly display all
:   zeros, and there were changes to the zero fuel weight information
:   displayed on the EFIS. At one time, the airplane flew for about one
:   hour with no problems noted. At 1355, when the airplane was about 20
:   nm miles north of the Kennebunk VOR, Maine, the flight crew declared
:   an emergency due to loss of the EFIS cockpit displays and the inertial
:   navigation units, and requested to land at Boston. The flight crew
:   extended the wing leading edge slats, and received a split slat
:   indication. After checking that the available runway length was
:   adequate, for their configuration and weight, they decided not to
:   extend the wing flaps for landing. The spoilers were armed; however,
:   after touchdown, the flight crew had to manually extend the spoilers,
:   and was unable to engage the reverse thrust. During the ground roll,
:   all main landing gear tires were deflated. Four tires were deflated
:   due to locked brakes, and four tires were deflated due to melted fuse
:   plugs. The passengers were then deplaned with the aid of stairs. The
:   flight crew reported that they were not aware that the thrust
:   reversers, and anti-skid were inoperative, or that they would have to
:   manually deploy the spoilers, and that when deployed, only the flight
:   spoilers would be available. According to a representative from
:   Boeing, the fail safe mode for the air/ground circuitry was the air
:   mode. Once failed, there was no inflight reset capability. The
:   following systems were affected when the air/ground circuitry failed
:   to the air mode: Thrust reversers - Inoperative Engine Idle - Remained
:   at flight idle, did not go to ground idle Spoiler deployment - Flight
:   spoilers only, ground spoilers not available, manual deployment only,
:   no auto-deployment. Auto-braking - Not available Anti-skid - Partial
:   availability The failure of the air/ground logic circuit to the air
:   mode was displayed on the maintenance status page of the EICAS. It was
:   not displayed as a warning, caution, or advisory. There were no
:   messages that the individual systems were either inoperative or
:   working with degraded performance. The airplane was examined and
:   ground runs were made with the electrical system in both the ground
:   mode, and flight mode. The failures have not been reproduced. On June
:   2, 1996, the airplane was moved to the Boeing plant, in Everett,
:   Washington, for continuation of the investigation. No anomalies were
:   reported from the flight.

Although, in this case, the computers aren't directly controlling the
airplane, the failure seems to have affected enough of the systems to
have almost caused an accident.  I'd done some programming and I know
how easy it is to have your program mistakenly switch into the wrong
"mode" under strange circumstances.  I guess it's good that the system
put out some warnings of incorrect behavior, but, since they were
unrelated to the real problem, it didn't help the pilots that much.

I'm not trying to stir up the FBW or software design wars again, but what
something like this happens and there is no off switch?

    Ethan Brodsky

--
<A HREF="http://www.pobox.com/~ebrodsky/"> Ethan Brodsky </A>