Hacker News new | past | comments | ask | show | jobs | submit login

No, it wasn't, and I agree this is a specious example.

The Wikipedia entry [1] has a good overview, as does this Medium post [2]. In the flight 1951 crash there were actually 2 altimeters for redundancy, but due to human error they were configured so that 1 faulty altimeter affected the autothrottle:

> While on final approach for landing, the aircraft was about 2,000 ft (610 m) above ground, when the left-hand (captain's) radio altimeter suddenly changed from 1,950 feet (590 m) to read −8 feet (−2.4 m) height, although the right-hand (co-pilot's) radio altimeter functioned correctly. The voice recording showed that the crew was given an audible warning signal (landing gear warning horn) that indicated that the aircraft's landing gear should be down, as the aircraft was, according to the captain's radio altimeter, flying too low. This happened several times during the approach to Schiphol. The reason that the captain's radio altimeter was causing problems was the first officer making a mistake when arming the aircraft's autopilot system for a dual channel approach.

> The Boeing 737NG type aircraft has two autopilot systems, which can work independently of each other (single channel) or together (dual channel). Turkish Airlines' standard operating procedure at the time stated that all approaches should be flown "dual channel" when available, but the inexperienced first officer forgot to arm approach mode in the aircraft's mode control panel (which controls the autopilot), meaning that the aircraft thought the pilots wanted to do a single-channel approach using control computer A (captain's autopilot), which had a radio altimeter failure.

[1] https://en.wikipedia.org/wiki/Turkish_Airlines_Flight_1951

[2] https://admiralcloudberg.medium.com/test-article-the-crash-o...




Want to plug your second source admiralcloudberg as a brilliant blog that is worth reading if you're into airline crashes and system failures. He's very in depth and detailed in his writing

https://admiralcloudberg.medium.com/


Thanks to your plug I started reading. Pretty good resource indeed.

I note that the accident report on the Turkish airplane further confirms what we have since learned about Boeing:

* persistent issues with radar altimiters in the entire fleet

* issue classed 'low priority', even if it feeds into autothrottle

* near-crash results in 'fix' that disables autothrotle if two sensors disagree.

* 'fix' is not rolled out to the entire fleet. And it is not 100% effective.

* issue is not mentioned in documentation, pilots can not see which sensor is used

Is it me, or is this exactly the same history as MCAS? Drives home the famous mantra "if it's a Boeing, I'm not going"


Also watch Mentor Pilot on Youtube. https://www.youtube.com/channel/UCwpHKudUkP5tNgmMdexB3ow

He is very through and detailed, an instructor pilot with a lot of experience.


Wow, the use of a single sensor to control a safety-critical feature is reminiscent of the 737MAX MCAS-related crashes, nine years later. (I understand that the pilots should have set the airplane to use two sensors in the above case, and that in the case of MCAS, I'm not an expert but it didn't sound like that was an option in the initial version.)


I’m curious how many instances of the gear warning horn going off is about forgetting to extend the landing gear compared to the plane being about to have a CFIT. Maybe it would be wise to always treat it as a GPWS “pull up” command and only try to troubleshoot at a safe altitude.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: