FMC FAILURE

[emerydc8]

Thanks, Peter. It would be great if you could try it in the sim this weekend and see if the AFDS actually follows the LNAV and VNAV. As you know, once a failed FMC is identified in the sim (you see the amber VTK), action is usually taken so promptly to select FLCH and HDG SEL that I don't know of anyone who has actually let it go to see whether LNAV and VNAV would actually contiue to work. It would also be interested to see if the BA sim gives you the AUTOPILOT EICAS for a failure of the master FMC.

[Hardy Heinlin]

I think, after the FMC failure event, it would be a good idea to wait at least 40 seconds and then see how LNAV/VNAV behave.


|-|ardy

[Britjet]

Indeed, Jon. I will try and fit it in..

Here is my FCOM2 extract:
---------------
In flight, the scratchpad message SINGLE FMC OPERATION displays on only one CDU after loss of the FMC not selected on the FMC selector. The scratchpad message TIMEOUT - RESELECT displays on the CDU with the inoperative FMC.

In flight, the scratchpad message TIMEOUT - RESELECT displays on both CDUs after loss of the selected FMC. The FAIL light illuminates in all three CDUs. The navigation display with the NAV source selector selected to the failed FMC is lost. There is a time delay before the FMC message FMC LEFT, RIGHT displays. When the FMC selector is rotated to the operative FMC, the FMC is available through the CDU. Once an FMC page is accessed, the scratchpad message SINGLE FMC OPERATION displays on the CDU with the operative FMC.

Dual FMC Failure

If both FMCs fail, LNAV, VNAV, and autothrottle are not useable. The CDUs supply route data to their respective ND.
---------------

Of course a single FMC "hiccup" (temporary failure) in cruise is quite common, and it is possible for the system to do a kind of "software reset" which is additional to the outage - if this happens (and it is mentioned in the FCOM as an "if") then you will temporarily lose LNAV/VNAV etc until you reactivate the data.

Interestingly, I note that the 777 has a L/AUTO/R selector, and the QRH stipulates that if you have the switch in other than AUTO you will lose LNAV & VNAV if the selected side fails. The 747 of course does not have the AUTO option but then neither does it deal with loss of LNAV and VNAV in the QRH. Go figure..

And as Hardy says - it is best to sit on your hands for a bit! I don't remember going to basic modes as you mention, but then I don't remember a lot these days!

Peter

[Hardy Heinlin]

Here is my FCOM2 extract:
---------------
In flight, [...] TIMEOUT - RESELECT displays on the CDU with the inoperative FMC.

In flight, [...] TIMEOUT - RESELECT displays on both CDUs after loss of the selected FMC.

Is there a certain delay or an action between these two FCOM statements? Otherwise they're contradictory, aren't they?

Perhaps the definition of "loss of the selected FMC" reads: "FMC is inoperative for 30+ seconds".

Also, I don't understand why CDU C should illuminate its FAIL light; CDU C is never connected to an FMC. I thought the FAIL light illuminates on a CDU if that CDU is connected to a failed FMC via the NAV source selector. CDU C provides standby navigation and alternate nav radio when both FMCs have failed; so dual FMC failure is a state that CDU C knows, but why should its FAIL light illuminate for a single FMC failure?


|-|ardy

[John H Watson]

My understanding and experience is that the flight system does not go to basic modes with a single FMC failure. You have to have a double failure for this to happen. Nor will you get amber lines through the LNAV and VNAV annunciations. They will be normal. Remember that these are not "picture" switches like the NAV source selectors - they show the state that the aircraft is in - it doesn't matter if one or other FMC side has failed.

Also, I don't understand why CDU C should illuminate its FAIL light;

My notes do say ALTN NAV RAD (including on the C CDU) should become available with both FMCs failed or if the master fails in flight, but there were some strange variations during some real world ground tests I did for you, Hardy.

Interesting. On an aircraft today, [on the ground with the IRUs in NAV], I pulled the Master FMC CB and got the following:

Capt: No <FMC prompt available on Menu, no NAV RAD at all, SNS(?) available.
F/O: All normal FMC functions ok (NAV RAD, etc). This doesn't make sense at all, because we are told that only the Master FMC can control radio tuning [and the wiring schematics show this].

With the IRUs off, both <FMC prompts not available. All alternate functions available.

Perhaps the definition of "flight" is IRUs in NAV?

As Hardy says, I would wait for at least 30~35 seconds and then look at what is available.

[emerydc8]

I can confirm that Peter's excerpt also shows up, verbatim, in a Polar Air Cargo FCOM from 2009. Unfortunately, it lists some of the indications you will see if an FMC fails, but it still doesn't describe what you will see on the FMA or whether LNAV and VNAV on the failed FMC (left in most cases) will continue to work without moving any switches.

https://drive.google.com/file/d/0B-WRh0Hf7VdZWktmakRwZThkM0U/view?usp=sharing

[hellowk]

To quote from my FCOM:

FMC Failure
Single FMC Failure
After loss of a single FMC, a resynchronization may occur. The active route may
become inactive, the performance data may be lost, and LNAV and VNAV modes
may fail.

So yes Hardy, I'm having a yellow strikeout through my LNAV and VNAV. The QRH did not mentioned anything about the LNAV and VNAV though. So Peter, you were right at the QRH part. It just mentioned to select a operable FMC, reengage the autothrottle and select a operable FMC on the NAV source selector.
Emerydc8, I'm reading exactly what you are reading with the Polar Air FCOM.

[emerydc8]

Hopefuly Peter can clear this up for us this weekend.

[John H Watson]

Further to what I wrote earlier...

http://www.iinet.net.au/~b744er@ozemail.com.au/PSX/Resync.GIF

[emerydc8]

So, it's looking like amber lines through the LNAV and VNAV, but if close to the MCP altitude, VNAV goes to ALT? This "degraded" autopilot mode and AUTOPILOT EICAS piques my interest.

[Britjet]

Hi Hellowk.

Absolutely true what you say - the emphasis of course being on "may" for each of the items. This happens if there is a software reset, which in itself may be the cause of the FMC failure in the first place, but by no means the only cause, hence the "may"...
Peter

[Hardy Heinlin]

In the current PSX version, the FMC malfunction on the Instructor provides these variations:

Severe: "Persistent power supply failure",
Non-severe: "Bubble-memory case temp exceeds 85°C, auto-shutdown"

This means in the non-severe variation the FMC will re-activate after some random generated minutes.

In the next PSX update I could change this variation to something like this:

"Non-severe: "Software reset and resynchronization in less than 30 seconds."

Would that be a more interesting variation?


|-|ardy

[emerydc8]

Yes, Hardy, that would be nice to have. Hopefully we can get more details about what exactly the FMA will show and what the AFDS does in the event of a single master FMC failure (e.g., does VNAV/LNAV continue to work or does it go to a degraded AFDS mode with an AUTOPILOT EICAS alert?).

Also, and this is a bit off-subject, Peter pointed out above that there is no checklist for loss of LNAV or VNAV. At my company, these failures are actually given more than a single FMC failure, probably because there is no annunciation -- You just have to recognize that LNAV or VNAV didn't go active when it should have.

My company likes to fail LNAV and VNAV on a heavy takeoff and give you an engine fire at rotation. Of course, they will do it on a runway with a special engine-out turn procedure too; so while you are loaded up trying to fly the procedure and dealing with the engine fire, they are hoping you forget to select CON thrust by the five-minute mark. You will usually recognize that VNAV has failed early-on when you are still looking at TOGA past 400', and you will probably have selected FLCH and set V2+100 at 1000'. But if you go one second over the five-minute mark with takeoff thrust still set (you forget to press THR), they fail the other three engines and you will eventually get the red screen of death. It's sadistic, but there's not much you can say because you've exceeded a limit on three engines.

They also like to fail VNAV on takeoff out of HKG where you are limited to 220 knots until established inbound to TD. We usually set a speed on the LEGS page at TD with an at-or-below altitude 220/7000B and just delete this once established inbound to TD. That way, VNAV keeps us from exceeding 220 knots. When VNAV fails on takeoff, our normal procedure is to select FLCH and set V2 + 100 at 1000' AGL. That's almost always more than 220 knots. We had a captain get busted back to FO for a year because he fell into this trap and didn't catch it until he was past 220 knots.

So, having the ability to fail LNAV and VNAV would be nice. I'm sure you have a lot on your to-do list right now and this is pretty low priority, but I thought I'd bring it to your attention that our sim has the option to fail LNAV and VNAV.

[Hardy Heinlin]

You can fail LNAV and VNAV already now in the existing PSX versions.

Just let both FMCs fail.

Or pull one of the probe heat CBs on P6-4 in row D.


|-|ardy

[emerydc8]

Thanks, Hardy. I'll have to try a random, dual FMC failure, if I can do that. It is a bit different because of the stealth way that just VNAV or LNAV can fail without any other warning (no EICAS or alerts and the FMC stays the same). I haven't done any dual FMC failures because the procedure calls for me to move two switches that are out of my view, unless I scroll down to the glare with my mouse. I'll check it out. Thanks again.

[hellowk]

Hi Hardy,

By failing both FMC or the probe heat will no doubt give the same results of a LNAV and VNAV failure.
However, we will probably have to deal with the problems of a dual FMC failure or a HEAT P/S CAPT which are different issues that we have to handle separately.
PSX have been instrumental in helping me understand what kind of faults yield what kind of results.
And to add on to emerydc8, out of HKG 07L, engine failure at 200", followed by a HDG SEL failure. Company procedures require us to fly a engine out procedure. But with the HDG SEL failure, it adds on to their 'fun' ( our torture).
What I'm saying is if the failures can be isolated like just  LNAV failure, or VNAV failure or a HDG SEL failure, it enables us to recreate many possible scenarios for training purposes. It's easy for me to say here and I believe it would probably take tremendous time and effort to implement. But it would also meant that PSX would be even closer to what we deal with in the sims. And this would be my wish list!

[Hardy Heinlin]

In your company sim, can you activate just LNAV fault, just VNAV fault, just HDG SEL fault -- separately -- without getting any other faults?

If so, have they occured in real life and therefore have been added to the sim? (If they happened in real life, they're obviously plausible). Or have they been added just for sadistic fun? :-)

Well, I can imagine an MCP pushbutton may mechanically fail. But once a mode is engaged, it can be disengaged only by pushing another switch (in most cases). That other switch might get a short circuit or something?

[emerydc8]

Our check airmen haven't found the HDG SEL failure button (and that's a good thing).

When they fail LNAV or VNAV in our sim (ex-ANA), there is no indication. All that happens is LNAV or VNAV just never goes active when it should -- it either stays armed or it entirely disappears from the FMA. This was actually a regular failure they were giving to everyone on type rides, although I don't know if it has ever occured in real life.

Usually, they would give a VNAV failure on the missed approach to 31L in JFK, where there is a 1000' level-off. That is one of the few missed approaches where you would select VNAV at 400' to keep from blowing through the 1000' level-off. With a VNAV failure, it won't level off at 1000'.

I think the missed approach may have recently been changed and they have since deleted the 1000' level-off.

[hellowk]

I have not experience a LNAV, VNAV isolated fault myself as compared to emerydc8. However I did experience the isolated HDG SEL fault without other faults. Probably they are trying to simulate like what u have said, a MCP push button may mechanically fail. And if it fails at a time when you needed it, things might can a little interesting.

[Britjet]

Thanks. Peter would know.

Hmmm..thanks for the comment - but it appears I don't!

On the big sim have just given this a go...
FMC selector to L. LNAV/VNAV departure on a SID with left A/P engaged at 250ft. LEFT FMC fail at 800r.

Amber "strikeout" on LNAV and VNAV after a few seconds. Both CDUs blank with TIMEOUT RESELECT (as expected)
Flight system and A/P in what I take to be "attitude stabilising mode" although there is no FMA annunciation of this.
Autothrottle disengages. Captain ND remains blank until switched on the NAV source selector.

Selecting FDs OFF then ON gives HDG (or ATT, depending on bank angle) and V/S modes on the FMA, with LNAV and VNAV showing ARMED (buttons still pressed on the MCP).

Selecting "FMC R" causes LNAV and VNAV to reengage GREEN. Autothrottle can be re-engaged.

I tried it again with an FMC R selection and the indications are what would be expected as per FCOM2.

Sorry if I misled anyone - I seem to remember it differently but obviously not. The FCOM2 and the QRH are pretty useless at giving help on this.

HTH

Peter