Revived "passengers vomiting in VNAV mode" thread

[martin-w 3,235]

Yes, I think some of you are experiencing the sudden wind shift issue. Which is an FSX bug, nothing to do with the NGX.The registered version of FSUIPC fixes the issue. Wind smoothing set, and turbulence suppression on.This isn't the issue I was talking about.

[shawkins 0]

Sorry, I'm still blinded by how marvellous this product is, and that's in light of the fantastic other products PMDG has produced before it!!! It is after all a non-professional, personal PC based application, not a $25M tool for airline pilot training, not to degrade the superior attention to detail PMDG has obviously invested. I've spent over 25,000hrs on FS (often to my chagrin) and the NGX takes the cake thus far!

[bpcw001 22]

Yes, I think some of you are experiencing the sudden wind shift issue. Which is an FSX bug, nothing to do with the NGX.The registered version of FSUIPC fixes the issue. Wind smoothing set, and turbulence suppression on.This isn't the issue I was talking about.

Well ... not quite.I'm running ASE and registered FSUIPC with wind smoothing and turbulence suppression. If you don't disable turbulence in FSX, you'll still get turbulence effects which will cause the NGX to pitch up and down occasionally, producing vertical speed changes of 2000-3000 fpm.It is probably necessary to disable turbulence in FSX too as it was written by some guy on these forums.Bad thing is: your flight will then be boring since there's not going to be *any* turbulence effect.So: either put up with the pitching up/down (which can be minimized, but not totally removed), or face an overly smooth flight at all times.

[martin-w 3,235]

Well ... not quite.I'm running ASE and registered FSUIPC with wind smoothing and turbulence suppression. If you don't disable turbulence in FSX, you'll still get turbulence effects which will cause the NGX to pitch up and down occasionally, producing vertical speed changes of 2000-3000 fpm.

It works with REX weather Dave.FSUIPC works perfectly for me in combination with REX weather.I haven't disabled turbulence in FSX.I never get the 2000-3000 fpm issue you do. However, despite having turbulence suppression on in FSUIPC... I still get some turbulence.Have you looked at the thread on the forum, regarding FSUIPC and ASE settings? It's obviously a lot more complex with ASE than REX weather.I do have ASE but not installed, i think I'll stick to REX weather then if ASE is so naughty.

So: either put up with the pitching up/down (which can be minimised, but not totally removed), or face an overly smooth flight at all times.

Nope, not for me with REX and FSUIPC.I have wind smoothing at something like 8 seconds per knot as I recall, and turbulence, variance and gusts suppressed.

[kabekew 0]

Here's a screenshot I made that illustrates this bug:Note a 4200 fpm climb at FL320! This is on autopilot in VNAV, plane loaded with passengers and fuel. It's trying to bleed off a mere .03 mach and it pitches way up from about 1300fpm to 4200 (actually I saw it go to 4800 briefly until it "stabilized" at 4200).This just isn't realistic. Yes, it can certainly "physically" achieve that by yanking back on the yoke, but the r/w 738 autopilots just don't behave that way. Please, PMDG, show this to a r/w pilot and he'll verify they never see 4200fpm climb rates at FL320.(To reproduce this:1. Plan any flight, say 75% pax and cargo loading, 20K lbs fuel, and climb to FL390 at normal FMC speeds (probably achieving mach .78 once transitioning to mach)2. Around FL320 hit SPD INTV and dial in .75 mach to simulate the very realistic ATC command to "reduce speed to mach .75 for spacing."3. Note the autopilot pitch up...way up...in a high g climb (I calculated on the order of 1.5-2g's just counting seconds and change from 1300fpm to 4800fpm over about 5 seconds).4. Note the unrealistic climb rate while it attempts to IMMEDIATELY bleed off all of .03 mach!The correct behavior should be that it uses more throttle than it currently does, instead of pitch, to achieve the desired airspeed. It's mainly to keep g forces on passengers (and stress on the airframe, I'm sure) to a minimum.)

Just one observation in this...the minimum rate of climb or descnt. Is my understanding that 500ft is the minimum rate that should always be commanded in order to comply with atc. And thus the autopilot should always command this even when it's pitching up or down to adjust speed. I too have noticed a levelling off on times.

In the U.S. for turbojets it's 1000 fpm. This is because around the world really, ATC uses what's called the "vacated altitude rule" that once an aircraft has reported or is observed out of an altitude, another aircraft below or above may be assigned that altitude. That falls apart if aircraft were allowed to level off anywhere (say, 100 feet above their previous altitude, like this autopilot sometimes does). That's why you're not allowed to level off in your climb, unless you get explicit permission (or it's something obscure like a cruise clearance or a block altitude).PMDG needs to fix the VNAV code to limit it to a minimum climb of 1000fpm (if aerodynamically able, of course), and in a non-crossing restriction descent (where you are in fact allowed to level off) there also needs to be a minimum descent rate of 1000fpm.Plus of course the original point of this thread, the change in climb/descent rate needs to be clamped to a maximum g force, whatever the real autopilot limits it to (1.1? 1.2?). Pulling 2g+ climbs like I've seen with this just isn't realistic.Just some constructive ideas for SP2 I hope!

[Guest]

There is no needing to reproduce it, it happens quite frequently also to me for example and I share your thought it is not real even if I am not a real pilot.I would add, I had few times a 6000 fpm under normal circumstances but just for few seconds.I ended up stating that that flight was a part of the Shuttle training program, no problem therefore.-

[JC75 22]

Are you totally sure about the 1000fpm rule? That's quite a rate to comply with for all stages of flight...in all jet aircraft.in the uk it's 500fpm

[kabekew 0]

Are you totally sure about the 1000fpm rule? That's quite a rate to comply with for all stages of flight...in all jet aircraft.in the uk it's 500fpm

Hmmm, no I'm not sure. That's just what I was trained (rw controller in US) but I looked into it: In our "airman's information manual" at http://www.faa.gov/air_traffic/publications/atpubs/aim/Chap4/aim0404.html section 4-4-10(d) it does say to advise ATC if unable to keep 500 fpm in the climb. That's not legally binding, but pilots here do tend to follow that. So it seems to be 500 fpm here too.I guess from an ATC perspective we just never expect or see 738's go below 1000fpm until they're just about to their cruising altitude. And I've never seen them level off like this sim does when assigned a faster speed, ever. And I've never seen any B738 go above 2,500 fpm in normal climb. The fastest climb rate I ever saw in a commercial jet was 4,000 fpm in an empty B747-400 in 1997. It was on a repositioning flight and I guess the pilots just wanted to try it out. I've never seen a B738 do anything like that above 10,000 feet, though, ever. And yet climb rates like that seem pretty common with this current version. When I'm online, I always click off the VNAV after 10,000 and dial in a more realistic V/S rate so the online controllers can expect something more real-world.

[pbagrat 1]

Flew from KRAP to KMSP today and had VNAV armed on takeoff. When I reached 1500 AGL (I had 10 degrees nose-up climbing at 2900 FPM) I engaged the Autopilot (CMD A) and the aircraft immediately pitched up another 10 degrees and ascended at 4600 FPM for about 10 seconds then settled back to around 3200 FPM. This happens on just about every flight unless I fly it up to 10000 by hand. I don't have a problem with steep descents - only climb. I am using ASE with registered FSUIPC and SP1.

[bpcw001 22]

It works with REX weather Dave.FSUIPC works perfectly for me in combination with REX weather.I haven't disabled turbulence in FSX.I never get the 2000-3000 fpm issue you do. However, despite having turbulence suppression on in FSUIPC... I still get some turbulence.Have you looked at the thread on the forum, regarding FSUIPC and ASE settings? It's obviously a lot more complex with ASE than REX weather.I do have ASE but not installed, i think I'll stick to REX weather then if ASE is so naughty.Nope, not for me with REX and FSUIPC.I have wind smoothing at something like 8 seconds per knot as I recall, and turbulence, variance and gusts suppressed.

I have pretty much the same settings and oriented myself along one long thread in this forum talking about ASE/FSUIPC weather settings.It works pretty well, but I do get these oscillating decents/climbs which the OP talks about. Even when there's no wind indicated.I suspect it has to do with the speed bug shifting around at certain speeds, adapting to pressure, temperature or whatever conditions that get injected by ASE, and the NGX autopilot reacts violently towards these changes, chasing the speed bug by altering pitch.I'd definitely like to see a resolution for this issue.

[kevinh 409]

Here's a screenshot I made that illustrates this bug:

Bug is a bit strong. VNAV works fine in most cases. It's more a lack of refinement in turbulence.

The correct behavior should be that it uses more throttle than it currently does, instead of pitch, to achieve the desired airspeed. It's mainly to keep g forces on passengers (and stress on the airframe, I'm sure) to a minimum.)

The correct behaviour should be how the real world NG operates. You don't reduce thrust to control airspeed in a climb and I've not seen any autopilot do this in practice. If turbulence put you under the VNAV speed thrust could not be increased beyond CLB N1 to increase speed so any attempt to use thrust in this way would result in an asymmetric control law, not likely to be stable. Some autopilots don't go straight for full CLB thrust in Level Change mode for small altitude changes but in a steady VNAV climb thrust will usually be set to the rated CLB thrust and stay there.Perhaps the NGX is a bit agressive in its speed control in Mach mode. The autopilot shouldn't try to follow rapid speed fluctuations due to turbulence. Control loops usually use acceleration terms to damp that out. I've only seen bad climb behaviour in the NGX if turbulence is active and PMDG recommend disabling that off in their manuals.I already had disabled turbulence before the NGX arrived because FSX turbulence is so unrealistic. The only time I see it is if I use real world weather, which seems to introduce turbulence of its own, ignoring the turbulence selection. A simple way round this is to select real world (static), then after the weather is set, switch to customized weather but leave the settings as is. That way you get real world weather and no turbulence.Kevin Hall

[RYR738 9]

VNAV works a treat in VNAV PATH descents. I am very happy with this.What I am not happy with are the climbs in VNAV Speed.They are far too aggressive.Fred.

[RYR738 9]

After lots more testing, especially with takeoff derates and being more careful with my fuel and payload weights,I would now like to retract my above statement that VNAV climbs are " far too aggressive "Maybe a little bit high at times but overall very good.I would not like to mislead anyone especially when in my case it was user error.Fred.

[bpcw001 22]

After lots more testing, especially with takeoff derates and being more careful with my fuel and payload weights,I would now like to retract my above statement that VNAV climbs are " far too aggressive "Maybe a little bit high at times but overall very good.I would not like to mislead anyone especially when in my case it was user error.Fred.

The derated takeoff gives me what I consider to be "realistic" performance up to 250kts. When reaching 250kts as speed limit below FL100, VNAV makes the NG rocket up at 3.000-4.000fpm. It flattens out starting from FL100 since the FMC-commanded speed will increase, but it still reaches 2.500-3.000fpm at times.That's for a fully loaded (pax+cargo) -600 or -700 with approx. 6.000-7.000kgs fuel.Is that a realistic performance? What is done against these "spikes" of climbing at 3.000 - 4.000fpm in real life? Do you change from VNAV to V/S?

[RYR738 9]

The derated takeoff gives me what I consider to be "realistic" performance up to 250kts. When reaching 250kts as speed limit below FL100, VNAV makes the NG rocket up at 3.000-4.000fpm. It flattens out starting from FL100 since the FMC-commanded speed will increase, but it still reaches 2.500-3.000fpm at times.That's for a fully loaded (pax+cargo) -600 or -700 with approx. 6.000-7.000kgs fuel.Is that a realistic performance? What is done against these "spikes" of climbing at 3.000 - 4.000fpm in real life? Do you change from VNAV to V/S?

Hi Dave.The derates give you a thrust reduction for takeoff but you might also need to choose CLIMB 1 or CLIMB 2.These give you a thrust reduction for climbs.Maybe try it and let me know how you got on.Fred.