FMC ICAO step size and RVSM

[Hardy Heinlin]

Yes, it makes sense. Thank you, Autopilot :-)

[Phil Bunch]

Somewhat off-topic, and something I've wondered about for years.

As is well-known, the jet stream at cruise altitudes can be as fast as 247 mph (per Wikipedia), and may be as thick as about 3 miles.  It seems to me that the earlier autopilots for sure couldn't be programmed to accurately handle such rapid transitions as an aircraft at cruise speed traversed such winds.

Do most or all autopilots compensate reasonably accurately for such high-speed winds and for such rapid variation with altitude and spatial location?  An aircraft flying at say 600 mph with 250 mph winds seems like a severe aerobatic challenge, unsuitable for airliners!

[Autopilot]

Modern autopilots (i.e. 737 classic, 744 and newer) cope with jetstreams just fine.

A jetstream is classified as a windspeed of 80kts (I believe) or more (fastest I've seen is 200 kts).

The core of a Jet as you said can be 3 miles ( roughly 18000ft) high (or more). but many more miles wide (sometimes 100's) and can obviously be 1000's of miles long.

The wind doesn't generally go from say 30 kts to 150 kts within seconds... It is a very gradual proces.

As long as it happens gradually nothing changes to the forces on the airplane, it still flies the same True Airspeed (or TAS) it's just the groundspeed that changes.

So if the TAS is 450 kts in no wind conditions then the groundspeed is also 450kts.

If the wind gradually ramps up to say a 150kt headwind, then the groundspeed goes down to 300 kts. (airspeed = TAS stays the same...)

TAS (wel IAS actually, but lets not complicate things) has to stay the same because, as you said the maximum force on the airplane is limited by Vmo / Mmo (Max operating speed / Max operating Mach).

And also the stall speed.

So the forces on the airplane are (initially) purely from the speed of the air over the airframe and wings.

So a steady increase in wind doesn't cause any upsets.

What sometimes does happen is that the change aren't as gradually as one would like and this causes turbulence due to wind gusts. This turbulence then causes the g-forces to increase and that obviously affects the structure of the airframe.

However I reckon a pilot would probably only ever see severe (and I mean proper severe turbulence, not heavy moderate...) turbulence once or twice in his / her entire career (30 years?).

Short anwer is that in 99% of the jetstream cases the autopilot does just fine. In those 1% (probably much less) where you might encounter proper severe, there might be issues with the autopilot disconnecting.

I've flown with a collegue who once lost 4500ft in severe turbulence and one other who lost 6000ft. These cases however are more rare than an engine failure, so nothing to worry about.

[jonb]

My understanding is that the FMC will take it that you are level at the new flight level from the waypoint which has the step climb inserted.I also understand that the FMC doesn't recognise if you don't actually do the climb or not, so if you are at FL340 with a step 360S inserted at the next waypoint but remain at FL340 past the waypoint the FMC still thinks that you've done the climb for it's fuel predictions.

Jon, by "next" waypoint do you mean the active waypoint? Are the two waypoints quoted in blue the same in your statement? If so, I don't understand the logic. I assume if the active waypoint is sequenced out (removed from the LEGS), the constraint of this sequenced waypoint will no longer restrict the flight plan, and the FMC computed optimum step profile will be displayed again, starting at the current aircraft altitude.


Cheers,



|-|ardy

Hardy,

Yes, I was referring to the same waypoint.Wether you climb or not at an inserted S climb the FMC will think that you have for the purpose of it's wind calculations, and there will be a "step to FLxxx now" message top right on VNAV.
As you know the S altitude entries are just for performance predictions rather than creating a VNAV profile. You tell the FMC you haven't climbed by reinserting your present cruise FL in 1L, this will remove the "step to FLxxx now" from the top right.
That should be in the honeywell manual somewhere, operationally most pilots won't be aware of these more intricate workings of the FMC as not a great deal of attention is given to the FMC generated arrival fuel, the computer flight plan is used for that. The FMC generated figure comes into it's own when operating in a terminal area when it is more accurate and "live".
I had a play at putting in a S step climb FL at the next waypoint which was way above the present max alt, it simply brings up a "step to FLxxx avail at xxxxz/xxxxnm" message top right on VNAV if there is sufficient flight time remaining , otherwise the message would be a "FLxxx unavailable "
The instant climb to a S climb waypoint altitude I mentioned in an earlier post was actually the way our computer flight plan software used to calculate it, and not in fact the FMC so please disregard that statement, sorry for the miss information on that one.

Re the question on autopilots in strong winds, higher altitudes in jetstreams are generally OK  as you have a larger tolerance of error allowed in track and height tracking.It's the transition of wind speeds, getting into and out of jetstreams  where you get the turbulence, I've been sat in 200kt tailwinds and it's been perfectly smooth. It's when lower down and it's coupled to an ILS it can start to pitch up and down quite badly especially in thermal or windshear conditions.

[Hardy Heinlin]

Hi Jon,

all clear now. Thank you!

Just one cosmetical question:

otherwise the message would be a "FLxxx unavailable "

My current PSX Alpha version displays the word "NONE" in 2R if a FL is unavailable, not the word UNAVAILABLE (haven't seen this in the manuals). I assume you're referring to the associated scratchpad messsage.

...

Re jetstreams. Here's a drawing I made a couple of years ago, while developing the jet stream model for PSX. The source of this principle is from a Jeppesen book called "Aviation Weather".



CAT = Clear Air Turbulence.


Cheers,

|-|ardy

[jonb]

Yes quite right Hardy, none, not ..unavailable. Another prime example of why pilots are no good at this sort of thing, I've been staring at that FMC for the last 16 years but still can't tell you what the correct wording is! Best sticking with the manuals.

Jon

[IefCooreman]

Confirmed by 744 guy: FL360S in legs page means step climb is supposed to start at the waypoint.

Enjoy the code deleting :-)

[Hardy Heinlin]

Done. Thanks :-)

I was wondering whether my prediction algorithm -- I mean the code that calculates the small font altitudes -- should, at the S, start a normal climb, or rather jump directly to the target altitude. If the leg is long after the S, it makes no difference. But if you have some short legs after the S, it could look like this:

3 NM .874/FL350
3 NM .874/FL390S
3 NM .874/FL361
3 NM .874/FL372
3 NM .874/FL390
3 NM .874/FL390

... or like this:

3 NM .874/FL350
3 NM .874/FL390S
3 NM .874/FL390
3 NM .874/FL390
3 NM .874/FL390
3 NM .874/FL390

...

By the way, speaking of autothrottles ...

When VNAV starts a cruise descent, the A/T mode goes to THR, reducing the power until approx -1200 fpm is reached, then changes to HOLD.

How long does it take from THR to HOLD? How smooth is it? My current model takes more than 20 seconds. Is this way too long? In general, PSX is a looooot smoother than PS1. Also on the LNAV axis. But in that THR to HOLD timing I may be exaggerating? It's all for pax comfort! :-)


Cheers,

|-|ardy

[Jeroen Hoppenbrouwers]

Does the A/T really control to a target V/S (so it must measure, correct, measure, correct, until satisfied)?

Or does it look up the throttle change from current to target setting from a table, guesstimating the setting for the target V/S, move there, and go to HOLD?


Jeroen

[Hardy Heinlin]

There is a V/S control by A/T when in FLCH SPD and when the distance to go is less than 4000 feet (at mode engagement). In that case the level change is flown in 2 minutes. (e.g. 2000 fpm for a 4000 feet change, or 500 fpm for a 1000 feet change).

In VNAV CRZ DES the A/T gradually reduces the thrust until -1200 fpm is reached, while the elevators keep the selected airspeed. Then it goes to HOLD. The -1200 fpm are just an approximate target. It may change when the air gets denser or the airspeed changes. The pilot can adjust the thrust when in HOLD mode.

In a well trimmed aircraft, be it a Cessna or a 747, you don't have to do much on the elevators when you want to change the altitude. Just reduce the thrust a bit and let the nose go down.


|-|ardy

[Jeroen Hoppenbrouwers]

Ok, so for VNAV (CRZ) DES is does not just set a presumed correct thrust and then go HOLD?

J

[Hardy Heinlin]

Bulfer/Gifford writes "it initially uses a predicted thrust". I don't know if "predicted" means "by database" or "by live detecting". In any case, I find it pretty easy to have the FMC detect the required thrust by smoothly retarding the throttles (also nice for the pax) -- and then HOLD it and use this held thrust as a "predicted" value until altitude capture.


|-|

[jonb]

Hardy,

That slow throttle reduction from THR to HOLD is deliberately programmed to be smooth and slow to reduce thermal shock of the engines after being at cruise power for hours.
We were taught not to manually override the throttle servos during this programmed thrust reduction. I wouldn't know if the timing and rate is different for different engine types though.
I can get you a timing for the GE if you want, but not going near an aeroplane for about 2 weeks now   , on leave, time for motorbikes instead !

[Hardy Heinlin]

I can get you a timing for the GE if you want, ...

Would be nice, yes, thanks!

Do you think, from memory, it takes about half a minute? Just approx. Not a whole minute, but not just 5 seconds either?


Happy ride on the bike!

|-|ardy

[Hardy Heinlin]

As we all know, when the MCP ALT bug is moved, the selected value may appear as a scratchpad message (which may update the FMC CRZ ALT by pushing the MCP knob).

Does anybody know when this "MCP-to-scratchpad" feature is enabled? Apparantly, it's disabled when on the ground. I guess it's enabled only when VNAV is engaged (or when  400 feet above runway). I'm sure it's also disabled when VNAV DES is active.


Cheers,

|-|ardy


P.S.: I think the following criteria are involved also: The new selected MCP altitude appears in the scratchpad only if it's ...

... higher than the current FMC CRZ ALT and VNAV phase is CLB
... higher or lower than the current FMC CRZ ALT and VNAV phase is CRZ*

* and not within 200 nm from T/D

[jonb]

I'll try and film it on my phone or at least time it, but from getting the FMA IDLE the thrust leavers retarding to idle, then getting HOLD on the FMA is probably around 15 seconds at most.

I can get you a timing for the GE if you want, ...

Would be nice, yes, thanks!

Do you think, from memory, it takes about half a minute? Just approx. Not a whole minute, but not just 5 seconds either?


Happy ride on the bike!

|-|ardy

[jonb]

From FMA "IDLE" to "HOLD" takes exactly 26 seconds, or it has done on the 2 times I've timed it. I'll try and get some more observations.

[Hardy Heinlin]

Exquisite, thanks :-)


|-|

[Manumanu]

Hi guys,

I just created this account for 3 things :

1) thanks for the wonderful program
2) thanks for all the info about step climbs (I'm in the process of writing a tutorial about step climbs for my VA)
3) add a little something from the 777 FMS pilot's guide I own about whether 350S is the starting or ending point of the step climb.
Here is a little graph about planned step climbs from this official honeywell manual :


The red line is mine, but as you can see, it seems to suggest that the planned step is ENDING at the waypoint, not starting. I'm sure there's an explanation and IefCooreman's assertion is true, but it's quite funny honeywell themselves are confused about this. And don't forget it is a 777 manual, so it might be different from 747. But it's strange enough to raise a doubt.

I am by no means a real world liner pilot (just some sunday vfr flying) but thought it could interest you guys. And I'll be most interested to read from real world pilots/honeywell-Boeing engineers on this.

Also, following your questions about steps-down, here is what the same manual says :

"If the last step is a step down, that step is maintained until the descent profile is intersected"

Seems indeed to suggest that they are perfectly possible. Same line in the 747-400 FMS Pilot's guide.

Eagerly waiting for PSX's release
Cheers

[Hardy Heinlin]

Welcome aboard, Manumanu.

"If the last step is a step down, that step is maintained until the descent profile is intersected"

Yep, this feature is also implemented in PSX.

3) add a little something from the 777 FMS pilot's guide I own about whether 350S is the starting or ending point of the step climb.
Here is a little graph about planned step climbs from this official honeywell manual :



The red line is mine, but as you can see, it seems to suggest that the planned step is ENDING at the waypoint, not starting. I'm sure there's an explanation and IefCooreman's assertion is true, but it's quite funny honeywell themselves are confused about this. And don't forget it is a 777 manual, so it might be different from 747.

Interesting. Just to confirm: In this printed example, is N61W030 a waypoint in the LEGS with FL390S entered by the crew, -- or an FMC generated phantom waypoint not occuring in the LEGS?


Cheers,

|-|ardy