Height loss in turn with autopilot on

[LecLightning56]

Unless I am doing something very wrong, but I am experiencing significant height loss in a turn with the 170 in the full flap, gear down configuration with the autopilot and autothrottle on.

I took the 170 up to 3000 feet, slowed down to approach speed (about 115 knots), selected full flap and gear down and then changed the heading. This resulted in a very significant height loss of about 500 feet. I recently tried a STAR arrival to London City EGLC with the 175 and experienced the same dramatic behaviour, which approaching EGLC at 2000 feet for an ILS to RW09 (again full flap gear down at 115 knots) is just not safe. With the latter case I was using LNAV and VNAV from a flight plan driven by the CDU. Hopefully I am doing something wrong which explains everything easily and I sincerely hope that this is not a major shortcoming of the product.

[scoobflight]

Height loss?  You mean porpoising as the flaps extend?

[LecLightning56]

No, the aircraft is configured in the approach configuration with flaps fully extended. I am merely executing a turn using the heading select knob with AP and A/T on and the aircraft suffers excessive height loss. I am prepared to accept if I am doing something wrong which is causing this issue, but at the moment the behaviour seems disconcertingly unsafe.

I just tried a fairly major heading change at 3000 feet which resulted in a height loss of 1000 feet. The aircraft eventually recovers back to 3000 feet in due course.

[scoobflight]

Only at Flaps full?

Try increasingly airspeed.

[LecLightning56]

9 minutes ago, scoobflight said:

Only at Flaps full?

Try increasingly airspeed.

Not being a real world pilot but I might have expected that the AP might be man enough to keep the flight parameters within tight bounds. Is it normal practice to increase speed even with AP selected in order to maintain constant altitude in a turn with the EJets?

[scoobflight]

Have you tried as I recommend?

[LecLightning56]

Lost me there. Please explain what asni recommend is.

[LecLightning56]

24 minutes ago, scoobflight said:

Have you tried as I recommend?

No. I am surprised that aircrew may have to be alert to this on a fully automated system particularly when their workload is excessive positioning the aircraft on the final stages of an approach. I am not aware of any other airliner which would require this degree of intervention in order to maintain constant altitude with an automated system. Is your recommendation in keeping with real world practice for this aircraft?

[LecLightning56]

Yes it does look as if increasing speed maintains constant altitude in a turn with AP and A/T on but I am not convinced that this is standard procedure and would be aghast if I had to be quite so alert with e.g. the 2000 feet turn for London City (the problem has nothing to do with steep approaches).

[scoobflight]

19 minutes ago, LecLightning56 said:

increasing speed maintains constant altitude in a turn with AP and A/T

Certified pilots have stated that the updated flight dynamics are closer to real then the prior release (which was also stated as being very close to real).  I'll wait for comments from them on real v game for any final analysis.

I recommend you set a v+10, or whatever worked, for full flaps, for turning maneuvers.

[LecLightning56]

Progress at last. It would appear that if I disarm A/T and control speed with the throttle lever, irrespective of considerations of v+10 knots for full flap (e.g. maintaining VAPP=115 knots), then altitude remains fairly constant in a turn or at least what could be determined as acceptable. So A/T seems to be the key player in the problem I have got, and is best disarmed when tight turns are flown at low speed and low altitude. This does seem a bug of sorts since I have proved that there is no necessity to increase speed with A/T off in order to maintain constant altitude in a turn. Puzzling why the A/T has this influence when it otherwise maintains constant airspeed as selected (as it should).

[scoobflight]

1 minute ago, LecLightning56 said:

So A/T seems to be the key player in the problem

Slow autothrottles is a known issue in the real world Ejet.  The response was moderated further in SP1 from original release after real-world pilot recommendation.

[LecLightning56]

11 minutes ago, scoobflight said:

Slow autothrottles is a known issue in the real world Ejet.  The response was moderated further in SP1 from original release after real-world pilot recommendation.

Interesting indeed. I did note that, although the FeelThere EJets are not configured for such procedures, for a steep approach to e.g. London City, the A/T manages to keep constant speed throughout the glideslope capture whereas previously the speed increased significantly during the descent.

It would however be useful if your real-world contacts could verify my findings that real EJets suffer such issues in reality, or if further tweaking may be required in the FeelThere flight dynamics to tighten up on the aircraft's ability to remain stable in a turning manoeuvre under automated control.

[Triholer]

Hey, 

even though I‘m not part of the feelThere-Team, I can assure you that the E-Jet Autothrottle is not always a real help on the real aircraft as well. 
 

The real A/T is a little bit slow on the compensation of changes (keep in mind, that even the 737 is not quite perfect there), but especially it has a tendency to rather go a little bit under the selected speed than above it - so for your calculations be rather pessimistic und choose speeds a little higher than you‘d usually choose. 

Therefore in our operations we stand to use our hand as a backward limit to the throttle, manually overriding it regularly. Especially in gusty conditions I would not trust the real A/T doing its business alone and by the calculated values. We do always override, but still keep it on as a safety net. 
 

By the way, what is your weight in the scenario? I do fly the E190 myself and know the E170 only from the Simulator, but 115knts - especially in a turn - sound really low to me. 
 

Best Regards
Björn

[Biggles2010]

I don't know what weight you were at but 115 kts seems quite slow to be doing any kind of sustained turn. A final approach speed is intended for small corrections during the approach. Remember that stalling speed increases progressively as bank angle increases, and so does drag, especially when the aircraft is 'dirty'. Any very low speed turn requires a significant thrust increase to maintain height. The auto throttles  always have a slight lag to avoid cycling, and often don't react fast enough. I cannot say how real Ejets perform at low speed, but jet experience years ago taught me to be very cautious about very low speed, high power turns at low altitudes. They easily lead to a fatal accident.

 

Edit: Triholer just beat me to it, and I think he confirms my view.

 

[Matty260191]

^^what he said.

[Biggles2010]

1 hour ago, Matty260191 said:

^^what he said.

Triholer and myself both posted at exactly the same time, hence my comment in the later edit.

[Matty260191]

OK THEN WHAT they said 🙂

[LecLightning56]

12 hours ago, Triholer said:

Hey, 

even though I‘m not part of the feelThere-Team, I can assure you that the E-Jet Autothrottle is not always a real help on the real aircraft as well. 
 

The real A/T is a little bit slow on the compensation of changes (keep in mind, that even the 737 is not quite perfect there), but especially it has a tendency to rather go a little bit under the selected speed than above it - so for your calculations be rather pessimistic und choose speeds a little higher than you‘d usually choose. 

Therefore in our operations we stand to use our hand as a backward limit to the throttle, manually overriding it regularly. Especially in gusty conditions I would not trust the real A/T doing its business alone and by the calculated values. We do always override, but still keep it on as a safety net. 
 

By the way, what is your weight in the scenario? I do fly the E190 myself and know the E170 only from the Simulator, but 115knts - especially in a turn - sound really low to me. 
 

Best Regards
Björn

Thanks Bjorn for the benefit of your own personal experience with the real aircraft, which is only for the good in terms of educating we earthbound mortals about how things really are.

The weight in my scenario is indeed low, with about 35% fuel and no passengers and cargo, (approximately 55,300lbs GW, sorry about imperial units). I am inclined to agree that for turning flight at such a low speed of 115 kts that the A/T may be endeavoring to maintain the speed but at the cost of not being able to provide the increased thrust to counter the increase in drag and hence the corresponding loss in altitude in order to maintain the speed itself, which I hope is correct A/T logic under such circumstances although certainly not to be entertained under real-world conditions. 

On the subject of the RNAV transition to London City EGLC RW09 (ODLEG 1G), I have repeated this with more realistic speeds for the turns associated with intercepting the localizer and things do play out very much better now. I need to practice this more carefully to get to grips better with the automation within the EJets but otherwise I am satisfied that I may have simply been putting impossible demands on the aircraft which probably even a real-world aircraft would struggle with to say the least (best not to live dangerously at low speed, low altitude when already at high power anyway). 

[Biggles2010]

1 hour ago, Matty260191 said:

OK THEN WHAT they said 🙂

Sorry.  I had misunderstood your previous comment.