Software bug in Bombardier airliner made planes turn the wrong way

Post Reply
Message
Author
ricardian
Chief Pilot
Chief Pilot
Posts: 5947
Joined: Sat Aug 22, 2015 10:08 pm
Location: 59°09N 002°38W
Gender:
Age: 80

Software bug in Bombardier airliner made planes turn the wrong way

#1 Post by ricardian » Sun May 31, 2020 1:56 pm

Ricardian, Stronsay, Orkney UK
www.stronsaylimpet.co.uk
visitstronsay.com
https://www.wunderground.com/forecast/EGER

User avatar
TheGreenGoblin
Chief Pilot
Chief Pilot
Posts: 17596
Joined: Thu Aug 08, 2019 11:02 pm
Location: With the Water People near Trappist-1

Re: Software bug in Bombardier airliner made planes turn the wrong way

#2 Post by TheGreenGoblin » Sun May 31, 2020 8:48 pm

The two links in that article are well worth reading as well. :-bd
Though you remain
Convinced
"To be alive
You must have somewhere
To go
Your destination remains
Elusive."

User avatar
TheGreenGoblin
Chief Pilot
Chief Pilot
Posts: 17596
Joined: Thu Aug 08, 2019 11:02 pm
Location: With the Water People near Trappist-1

Re: Software bug in Bombardier airliner made planes turn the wrong way

#3 Post by TheGreenGoblin » Mon Jun 01, 2020 2:55 am

Though you remain
Convinced
"To be alive
You must have somewhere
To go
Your destination remains
Elusive."

User avatar
TheGreenGoblin
Chief Pilot
Chief Pilot
Posts: 17596
Joined: Thu Aug 08, 2019 11:02 pm
Location: With the Water People near Trappist-1

Re: Software bug in Bombardier airliner made planes turn the wrong way

#4 Post by TheGreenGoblin » Mon Jun 01, 2020 3:08 am

Rockwell Collins has determined that an error in the design of the Pro Line 4 FMC software causes changes to the procedure-defined turn direction, during a missed
approach when the procedure has been significantly modified The FMS may change the planned database turn direction to an incorrect turn direction when the altitude climb field is edited. The potentially incorrect turn direction is dependent on leg types and geometries of the instrument departure procedure and missed approach procedures. The flight crew may not notice the change during climb.

The FMS may also change the planned database turn direction to an incorrect turn direction if the temperature compensation is activated, which may go unnoticed by the flight crew with the increased workload involved with a missed approach procedure This problem was reported to Bombardier in July 2017.
Bombardier are loth to fix the root cause due to cost and commercial considerations.

Bombardier1.JPG

If they need Ada programmers, there are plenty of them around. Clearly fixing the root cause is the correct thing to do whatever the bean counters think. Yes the software will have to be fully system and regression tested and re-certified. and, yes, that will cost Bombardier & Rockwell Collins plenty. If either company thinks this is too complex then they should get out of this business.

https://code7700.com/altimetry_temperat ... ection.htm
Though you remain
Convinced
"To be alive
You must have somewhere
To go
Your destination remains
Elusive."

Boac
Chief Pilot
Chief Pilot
Posts: 17199
Joined: Fri Aug 28, 2015 5:12 pm
Location: Here

Re: Software bug in Bombardier airliner made planes turn the wrong way

#5 Post by Boac » Mon Jun 01, 2020 9:49 am

With a skimpy article like that in El Reg it is difficult to understand what the problem is. G/A 'turns' are normally in a specified direction (in which case the database should not 'change' due to temp correction) or 'Return to xxx beacon' in which case direction is not specified and therefore not an issue.

I cannot find any approach charts for CYXJ in the Canada AIP - but I am not familiar with that document! It would be interesting if anyone has copies to see what this 'problem' was.

In the case of the R24 approach at Scatsta in the article the turn IS specified as 'left' and logically takes you into the hold or into another approach. That 'left turn' would be FIXED in the database and not subject to change.

Another example of 'journalism' tantalisingly giving only part of the story.

User avatar
TheGreenGoblin
Chief Pilot
Chief Pilot
Posts: 17596
Joined: Thu Aug 08, 2019 11:02 pm
Location: With the Water People near Trappist-1

Re: Software bug in Bombardier airliner made planes turn the wrong way

#6 Post by TheGreenGoblin » Mon Jun 01, 2020 10:03 am

Boac wrote:
Mon Jun 01, 2020 9:49 am
With a skimpy article like that in El Reg it is difficult to understand what the problem is. G/A 'turns' are normally in a specified direction (in which case the database should not 'change' due to temp correction) or 'Return to xxx beacon' in which case direction is not specified and therefore not an issue.

I cannot find any approach charts for CYXJ in the Canada AIP - but I am not familiar with that document! It would be interesting if anyone has copies to see what this 'problem' was.

In the case of the R24 approach at Scatsta in the article the turn IS specified as 'left' and logically takes you into the hold or into another approach. That 'left turn' would be FIXED in the database and not subject to change.

Another example of 'journalism' tantalisingly giving only part of the story.
Read the attachments in the article... Sigh!

You can lead a Lightning Pilot to water but you cannot make him drink... =))

https://portal.rockwellcollins.com/docu ... 05f6edfc18

The G/A turn direction is calibrated taking a number of variables into account, turn radius, speed, altitude (corrected for temp etc), height of local topography etc. The bug in the software results in possible deviations from the official approach plates and may be triggered by automatic altitude correction by the software when the temp is well below the ISA temp (see link above) for a fuller picture.

;)))
Though you remain
Convinced
"To be alive
You must have somewhere
To go
Your destination remains
Elusive."

Boac
Chief Pilot
Chief Pilot
Posts: 17199
Joined: Fri Aug 28, 2015 5:12 pm
Location: Here

Re: Software bug in Bombardier airliner made planes turn the wrong way

#7 Post by Boac » Mon Jun 01, 2020 11:16 am

Firstly, my sincere apologies for not reading that link - I must have had my head in the B-Scope (qv) looking for my sandwiches and I missed the 'PDF' link. By the way, you will not find Lightning pilots wanting to drink much on ops lest they need to do a wee-wee after 4 or 5 hours.............

Yes, yes, yes, but one needs the ACTUAL approach plate to see whether the turn direction is specified or FMS generated.

"the FMS will remove the database turn direction (if any) on the immediately following leg. The FMS will turn in the wrong direction after sequencing the “Climb to” leg if the shortest turn direction is different than the required turn direction onto the next leg." and

"The G/A turn direction is calibrated taking a number of variables into account, turn radius, speed, altitude (corrected for temp etc), height of local topography etc." is not the whole story - the turn may well be predicated procedurally as at Scatsa.

The LFBO example is weird! It makes it a major sumo-wrestling nightmare to get onto the inbound course to TOU although the database still reflects that! That is an inexplicable and inexcusable cock-up and I would have tended to ban ALL LNAV operations thus causing major issues for the 'youngsters' :)) until the guilty barsteward programmer could be led to BM's neck massaging machine.. How on earth any programmer can come up with that I do not know.

User avatar
TheGreenGoblin
Chief Pilot
Chief Pilot
Posts: 17596
Joined: Thu Aug 08, 2019 11:02 pm
Location: With the Water People near Trappist-1

Re: Software bug in Bombardier airliner made planes turn the wrong way

#8 Post by TheGreenGoblin » Mon Jun 01, 2020 11:49 am

Boac wrote:
Mon Jun 01, 2020 11:16 am

The LFBO example is weird! It makes it a major sumo-wrestling nightmare to get onto the inbound course to TOU although the database still reflects that! That is an inexplicable and inexcusable cock-up and I would have tended to ban ALL LNAV operations thus causing major issues for the 'youngsters' :)) until the guilty barsteward programmer could be led to BM's neck massaging machine.. How on earth any programmer can come up with that I do not know.
I agree. As for the immediate execution of the programmer rather than his code, I can relate many anecdotes, some personal, but the best related to me by two ex-members of the Sea Wolf missile guidance software team, narrated to me over a goodly number of beers while attending an SSADM course at Sheffield University back in the early 80's. The best part, including the missile doing a full 180 degree turn and returning directly towards the firer's (as opposed to the Fuehrer's) bunker due to a software bug caused by the most elementary of software programming errors, certainly made me laugh I can tell you. =))
Though you remain
Convinced
"To be alive
You must have somewhere
To go
Your destination remains
Elusive."

Post Reply