IRS Drift Example

[Mariano]

Hi,

Here is a good example of IRS drift on the ground and how it affects non-GPS airplanes (until radio updating begins).

On the first and second pictures, we were holding short of runway 18L.

https://www.dropbox.com/s/a64jqu5nt9cpm1a/Photo%20Dec%2005%2C%2008%2017%2008.jpg?dl=0

https://www.dropbox.com/s/v4agyex195etvzv/Photo%20Dec%2005%2C%2008%2017%2011.jpg?dl=0

On the third picture, we were holding position on runway 18L.

https://www.dropbox.com/s/xaqkbd0fwmdjd2n/Photo%20Dec%2005%2C%2008%2018%2051.jpg?dl=0

Notice the 5-knot ground speed (even though we were stationary). By the time I took these pictures, IRUs had been aligned for two and one half hours, with a twelve-minute taxi (IRS groundspeed while parked on the ramp peaked at two knots for quite some time).

Also, note the unusual TERR display (which I assume was caused by the somewhat excessive IRS drift).

Just thought this was worth sharing.

Best regards,

Mariano

[Hardy Heinlin]

Hi Mariano,

thanks for sharing.

Question: The alignment lasted for 2.5 hours? Or the alignment was completed 2.5 hours ago?


Regards,

|-|ardy

[Mariano]

Alignment was completed two point five hours prior to the pictures being taken (we experienced loading and other delays).

Best regards,

Mariano

[tango4]

Very interesting, thanks a lot for sharing.


Charles

[Hardy Heinlin]

I wonder if the drift can be increased during loading when the cargo containers hit the deck platform one after the other -- all these shocks etc., they are small but hard. And they all go into a certain direction.


|-|

[Mariano]

I have wondered about that before.

If the aircraft doesn't have PDUs (as was the case with this one), some of the loading can get very rough.

I know that Boeing says that during alignment, normal loading should not cause IRUs to pause alignment (only wheel rotation), but sometimes I wonder (and look up expecting to see flashing ALIGN annunciators).

Nevertheless, like you pointed out, there is no mention of continuously forward and aftward-smashing ULDs and pallets probably inducing very small but repetitive accelerometer motion.

I will pay more attention at individual IRU ground speeds next time loading gets rough.

Best regards,

Mariano

[John H Watson]

You would think that GPS would be mandated for the EGPWS/Terrain system. It relies on accurate position data.

[Markus Vitzethum]

Hi John,

> You would think that GPS would be mandated for the EGPWS/Terrain system. It relies on accurate position data.

I agree - but then again, what is "accurate" position data? You would think that with VOR/DME updating you have accurate position data?

Here is an interesting near-CFIT event from the times before GPS was widely available on most aircraft (2003), yet EGPWS was already available.

https://www.skybrary.aero/index.php/A320,_vicinity_Addis_Ababa_Ethiopia,_2003

This A320 nearly crashed at Addis Abeba because VOR updating to the (previously correct) INS position introduced a FMGC map shift due to faulty VOR (ground) equipment. This led to a position deviation towards a mountain and the EGPWS did not catch it because it relied on the same faulty position data. If I remember correctly, the classic look-down GPWS finally saved the day.

Markus

[emerydc8]

I have wondered about that before.

If the aircraft doesn't have PDUs (as was the case with this one), some of the loading can get very rough.

I know that Boeing says that during alignment, normal loading should not cause IRUs to pause alignment (only wheel rotation), but sometimes I wonder (and look up expecting to see flashing ALIGN annunciators).

Nevertheless, like you pointed out, there is no mention of continuously forward and aftward-smashing ULDs and pallets probably inducing very small but repetitive accelerometer motion.

I will pay more attention at individual IRU ground speeds next time loading gets rough.

Best regards,

Mariano

20 years ago they used to load DC-8s with nothing but fork lifts. They would raise the ULD and let is slide right off the fork lift into the door of the airplane. It was pretty rough, but even then the IRS alignment (Carousel IV and Litton 92) would simply restart at a higher number in the countdown and continue its alignment. I never saw a case where it hit so hard that we had to take any action. Some of the cans went in pretty hard, like we sometimes wondered if they were going to go out the right side of the fuselage. I don't think any kind of rough loading you'd experience today would even come close to how they did it in the past.

Jon

[evaamo]

Hi all, my ignorant question regarding flight ops and the case depicted in Mariano's screenshots: is takeoff with such a high drift allowed because once airborne the VOR/DME will update the a/c position to an accurate position?.

Mariano: since it wasn't a GPS equipped a/c and you mentioned the alignment was done 2.5 hours earlier, is this a case that would have been reported and sent for review to your carrier's flight ops so it could result in an advisory or notice to crews that it's safer to perform alignment after cargo loading is complete in similar non-GPS a/c? Or is this something that happens regularly and VOR/DME position updating is reliable enough to avoid the need to re-align or perform a late alignment?

Thanks for sharing, interesting topic.

[Mariano]

Evaamo,

Since we were performing an RNAV1 SID, we needed to correct that drift before becoming airborne. This is done automatically when we engage the autothrottle in takeoff thrust by pressing the MCP N1/EPR button in the 767, or the TOGA switches in the trust levers in the 747-400.

When we do this, the FMC corrects its position to the stored coordinates of the threshold of the runway we entered in the CDU during preflight (if performing an intersection takeoff, we enter the runway followed by a slash and the taxiway designator (example: "R18L/S10")).

As we do this, we check to make sure that the runway symbol "jumps" to the triangle symbol (or viceversa? ;-) (it takes about two seconds). If this does not happen, we can continue the takeoff, but we must inform ATC that we are unable to continue on the very precise RNAV1 SID on which they cleared us, at which point they would give us a heading (not a preferable course of action, though). I have never seen the FMC position not update to the runway threshold position (or the coordinates at which we told the FMC we would be engaging the autothrottles to takeoff thrust).

As you pointed out, if the update does not happen on the runway, after airborne, and if sufficient and valid VOR-DME or DME-DME signals are received, the situation will correct itself, but it would be too late if we were supposed to be following a very precise RNAV SID path (we could very easily turn and converge onto the departure track of a parallel runway in a busy airport).

If this was one of our ETOPS-certified 767s (which it was not) maintenance would need to know about this excessive drift. We have a procedure when flying ETOPS-certified 767s, by which we are required to record each IRU's drift rate (NM/HR) from the FMC MAINT>IRS MONITOR page in the ACARS post-flight report after every single flight.

This way, maintenance can track individual IRU performance of each and every ETOPS aircraft, allowing them to detect negative trends (wether they are GPS-equipped or not).

GPS-equipped airplanes do not exhibit this issue, unless GPS NAV is manually disabled (I believe TOGA runway position update is disabled whenever GPS NAV is enabled).

Best regards,

Mariano

[evaamo]

Thanks for the detailed reply, Mariano! 

>>This is done automatically when we engage the autothrottle in takeoff thrust by pressing the MCP N1/EPR button in the 767, or the TOGA switches in the trust levers in the 747-400.

Fantastic stuff!  I should read the FCOM throughly one day ;-)

cheers
-E

[Mariano]

Anytime ;-)

Best regards,

Mariano

[Avi]

Couldn't you find a "quiet" 30 sec and performed a fast alignment?

[Mariano]

Avi,

That's a good and valid point, nevertheless, in order to do that we would need the coordinates of the expected takeoff position (runway threshold or intersection), to which we do not have access, but which the FMC automatically applies to itself when we press N1/EPR (or TOGA).

Moreover, manually entering coordinates before taking the runway has a significant potential for error, which would be even more critical before performing an RNAV 1 SID.

Best regards,

Mariano

[Roddez]

Many years ago, I recall being in the cockpit in the early 90's (When you still could) of a Qantas 767 flight from Wellington (NZWN) to Sydney (YSSY).  As we were crossing the Tasman sea so there were no ground aids available and we were on inertial nav only.

We were on descent into Sydney when we came within range of the ground stations in Australia.  The pilots were expecting this and said "Watch This" and the aircraft performed a correction.  It wasn't big but it was noticeable.

How did this manifest itself in the older generation of aircraft (747 Classic, DC8, etc) that relied on both INS navigation for Overwater sectors and Ground based aids for land navigation?  I wonder if anyone on our forum remembers those days?

Rod.

[emerydc8]

How did this manifest itself in the older generation of aircraft (747 Classic, DC8, etc) that relied on both INS navigation for Overwater sectors and Ground based aids for land navigation?  I wonder if anyone on our forum remembers those days?

At Polar in the early 2000s on the classics, we had some Litton 92 INS units that only had DME updating. I remember it was like clockwork going into ANC from Asia, as soon as it picked up a certain VOR (I don't remember which), it would turn almost 90-degrees to the south. You had to be ready for it and if you weren't the next question was WTF is this thing doing now? Then the pilot would reach for the turn knob on the center pedestal.

On the DC-8s, during the entire nineties, most of ours at Emery Worldwide had the Tracor (Trimbal) 7900 units (VLF/Omega). Toward the mid-nineties they added a full-time GPS update feature but I don't think it was ever certified for us (We weren't allowed to use it, but we all did.) So the DC-8 either had full-time integral GPS on the Tracor or it didn't. There was no DME updating. On some earlier DC-8s in the late 80s, we had the Delco Carousel IVs and Litton 51s (the ones with a thumb-wheel and only 9 waypoints), that was just a straight triple-mix operation and they did a pretty good job despite no updates at all, other than you could press HOLD when flying over a VOR and manually update it with the VOR Lat/Lon.

Things have come a long way since then.

Jon

Tracor 7900


Carousel IV


Litton 51

[Balt]

Very interesting Mariano! From an operational perspective it seems odd that you can initiate the takeoff when the RNP requirements for the SID aren't met until it's too late to abort...?

[Jeroen Hoppenbrouwers]

I agree that this is a bit of a gray area, but the position update happens when you hit the "set thrust" function which by definition you do before you are even in the high-speed segment. So you can also just retard the throttles and take the first exit on the left to sort it all out.

Hoppie

[Hardy Heinlin]

When a map shift exists prior to takeoff, what else should the "set thrust" function do aside from setting the thrust? Just do nothing? What if you just want to takeoff now regardless of the entered SID?

Can this 767 generate the message "UNABLE RNP" before takeoff?


|-|ardy