To the OP, I do not think this is soley an NGX issue. I have noticed it in the NGX, 747-400X MD-11X and 767-Level-D as well. Example if you did not follow my tutorial for editing your FSX.CFG and FSUIPC.ini and Activesky and have just the basic weather, and you suddenly have a radical wind shift and overspeed the plane will radically pitch up to bleed off speed as fast as possible.Another thing is when you are climbing in hot temps down near the equator and suddenly ASE refreshes and the Temps suddenly go from say +15c to 1 or 2c the plane will start to climb a little radically because of the cooler air temperature. Again this is something I have seen in all 4 planes. NGX, 747, MD11 and the Level-D. This is what makes me think its more a weather related thing than an NGX issue.Lets not forget that the real world aircraft and the real world weather act totally different because FSX can never fully simulate real world weather conditions. Maybe some day a Simulator will come out that more accurately simulates real world weather, but in the meantime developers are stuck with living with the limitations of FSX's Crappy weather engine.There is only so much you can do and sometimes comprises and sacrifices have to be made for the sake of function over realism.Finally I have never seen the NGX go over 2200 feet per minute climb rate above 30,000 even when the aircraft was light. Most of the time my climbs average from 700 to 1000 fpm above 25,000 unless there is a radical change in weather which is rare for me with my current weather setup.

FWIW I work in the terminal environment and most turbojets average around 2000fpm up to 15,000-21,000 from what I can tell on the scope. The bizjets pump out more avg around 3000fpm climb. This is interesting discussion btw cause now I know my climb rates are waaaay too much (some user inflicted by carrying light loads and low fuel) but perhaps I should pay better attention especially when flying on vatsim or pilot edge hehe.

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.

Update: I am currently doing LPPD to LPMA and weigh in at 142,000 lbs currently at FL370. I did a test and set altitude to FL370 and before I started my climb from FL350 I set speed intervention to 30 knots slower than the current speed was and then it started to climb and the max rate of climb was 2000 fmp. That is considerably less than your 4200 fpm climb. I have never seen that at high alittudes but I have seen it on lower altitudes when very light, or on descent when your at 260 knots and you select level change and it dives as much as needed to quickly get from 260 to 280kts. In this sceneario I have seen up to 5000 fpm until it got up to speed and finally leveled to a reasonable descent rate. I have also seen the following behavior. Your late on your TD and say 100 nm out, you set VNAV descent and say your cost index is 36 which would be a 280 knot descent. At first it will dive extremely anywhere from 3500 to 5000 fpm until it gets to the altitude it should be at then it starts descending more realistically.Update 2: While descending to LPMA on VNAV from 37000 at 120 nm out it was an 1800 fpm descent which is pretty normal I think. At 25,000 I set speed intervention to 300 and I was at 273 knots. It dived at 5000 fpm until it go to 300 knots. So I guess I see what your saying now. Rather than use the engines it just uses the pitch to get to whatever speed you set.I guess I just never noticed this much cause 99 percent of the time I use VNAV and not level change.

Edited December 2, 201114 yr by UAL115

Gabe, et al: I opened a ticket some time ago on this issue...I am seeing similar results with somewhat wild pitch corrections when SPD or VNAV mode is altered.

• Ticket Number 1BD-165ED99B-FE68

Hopefull this is something that can be fine tuned...I will sympathise with PMDG in that this type of thing is hard to pin down and get to a point where everyone is happy. Personally, I would like to see a dampening of the pitch reaction / AP inputs to the elevator...make a SPD change in flight and watch the elevator position display on the lower DU if you are not sure what I am on about.Rob Hall.

Ok,Here's some facts folks - there is a lot misconception about how the plane's systems and physics in general really work going on in this thread:1. VNAV SPD is a "pitch for speed" mode - it is essentially the same thing as level change. Thrust is NOT modulated to control V/S in a VNAV SPD climb - EVER! The A/T sets the thrust at the climb N1 limit and then varies pitch to control the airspeed. If you use SPD INTV to slow down in a VNAV SPD climb, you are asking the airplane to pitch up - period. It will not reduce thrust and pitch to a lower angle, it keeps the N1 and pitches to whatever angle results in the selected airspeed.Reasons you could see "high" V/S in a VNAV SPD climb:a. The plane is extremely light - I see tons of people doing extremely short routes with the default payload and 1/3 fuel all the time. The thing is going to be a rocket even in the derated thrust modes at these weights. In real life, a super light NG is lost revenue for the airline. Try it closer to MTOW and see if the "high" climb rates aren't reduced.b. Derated/reduced climb not being used - if you put the plane in the unmodified CLB mode, that means the N1 limit the A/T shoots for is higher, requiring a higher pitch angle for the speed.(I believe these things explain why the real life controllers in the thread report seeing NGs doing ~2000FPM - they're heavy and they're derated)2. Regarding V/S - a lot of simmers have been led to believe by certain "passenger simulator" utilities that high V/S in a climb or descent is somehow bad or unrealistic. This is 100% untrue. One of our tech team pilots who flies the NG every week just told me that he saw over 7000fpm while reducing airspeed in the climb while light in a 700 the other day. (he's going to try to get video for me on his next trip) Bit of science/physics here - people *do not sense speeds*, we sense acceleration. Think about when you're up at cruise as a passenger on a plane in real life - you feel completely normal, you can get up and walk around etc. If you closed your eyes and put perfect earplugs in that blocked all the sound, you'd have no way of discerning whether you were sitting in your room at home or in a plane doing 600 knots over the ground. In fact, even when you're sitting at home in your room, the Earth your house is attached to is moving at a constant 67,000 mph around the Sun and in turn the Sun and our solar system as a whole is moving at an even higher rate of around 515,000 mph due to the rotation of our Milky Way galaxy. (You can thank Issac Newton and Albert Einstein for these insights into reality, seriously!) Motion at constant speeds do not matter in the SLIGHTEST and this applies to climb rates too - climbing at 500fpm feels exactly the same as climbing at 4000fpm as long as you're maintaining that constant vertical speed.3. Feet per minute in a climb does not = G forces. A 3000fpm climb does not mean you are pulling 3Gs. This is 100% an incorrect understanding. If you are ascending at a constant rate of 3000fpm, you're actually at around 1G. G forces are involved in the CHANGE between no climb and getting to 3000fpm. If you extremely violently pull up (like yank the yoke back completely in an instant), you will get that kind of G force. Gs are a measure of acceleration force - acceleration is a rate of change (think of it as change in speed per unit of time.) "G force" specifically is a measure of acceleration with respect to the normal pull of Earth's gravity. Standing on the surface, you're "pulling" 1G. 2Gs is two times that value, 3Gs is 3 times it and so on. I think a lot of simmers believe that G forces in an airliner are a lot higher than they really are - for comparison, an F-16C doing a max rate turn is going to pull somewhere in the vicinity of 9Gs max. No airliner can do anything even close to that even if you were completely manhandling the controls. Even 2Gs is unbelievably uncommon - I've done a 2G or so steep turn in a GA aircraft and if you haven't ever done it before, you're going "Holy crap!" - it is a HUGE change vs. what you would ever feel on an airliner in almost any situation.Now - taking into account all of this knowledge - is there still an assertion that something is wrong with our autopilot? If so, what exactly is it that does not violate any of the understanding above?

Now - taking into account all of this knowledge - is there still an assertion that something is wrong with our autopilot? If so, what exactly is it that does not violate any of the understanding above?

Excellent explanation and debunking of some myths, Ryan. The only thing I'd say in criticism is that the autopilot can be a bit agressive correcting for temporary speed fluctuations due to turbulence, leading to unwanted oscillations in V/S. It should be controlling the speed trend, not short term oscillations which should be filtered out. If you make sure turbulence is completely disabled, especially at altitude, then VNAV works just fine. It would be better if the VNAV speed control law was refined so turbulence could be used though.Kevin Hall

That's a matter of how fsx simulates turbulence. It isn't just rough air flow, fsx also dampens the control surfaces, sometimes to an unbelievable degree. Now matter what they did to tune vnav, it wouldn't matter if fsx drops the elevator effectiveness from 100% to 10% or less. I'll wager this dampening effect is ever changing as well, which would make compensating for it extremely difficult, if not impossible.

If you closed your eyes and put perfect earplugs in that blocked all the sound, you'd have no way of discerning whether you were sitting in your room at home or in a plane doing 600 knots over the ground.

I'd know in an instant........my furniture is way more comfortable!

That's a matter of how fsx simulates turbulence. It isn't just rough air flow, fsx also dampens the control surfaces, sometimes to an unbelievable degree. Now matter what they did to tune vnav, it wouldn't matter if fsx drops the elevator effectiveness from 100% to 10% or less. I'll wager this dampening effect is ever changing as well, which would make compensating for it extremely difficult, if not impossible.

I can't say I've ever noticed FSX messing with control surface damping, maybe that's with the default autopilot. I agree FSX turbulence is very poorly modelled. The airspeed fluctuations are excessive and violent.What I'm talking about isn't damping the controls but the control law which drives them. If you filter out rapid airspeed changes using an inertial acceleration term the autopilot won't try and compensate for short term speed changes. It wouldn't matter how low FSX made the surface effectiveness in that case.Kevin Hall

Here we go. My memory wasn't perfect, but still fairly close to what was said.

No Tim. I think ASE presently triggers the turbulence intensity level inside FS (0 to 5 if I remember correctly). Pseudo - random wind data mimicking turbulence are still generated by FS.While testing the NGx I found an "interesting feature". The FS sim engine in order to simulate heavy-ish turbulence and aircraft "displacement" effects simply appears to cut off, internally, the effectiveness of control surfaces (rendering thus the airplane almost helpless). It is like flying in vacuum...Just get rid of that turbulence and thermal effects "feature". It is OK for small planes but highly inaccurate for airliners.

I just tried a bunch of tests in the 700WL - around 114.0 ZFW, 12,000 lbs of fuel (this is extremely light) no wind, standard ISA.I do not see this "high G" pitch up thing that's being talked about. I had the Shift+Z G meter open the entire time - when I went from Mach .78 to .75 at FL320 I saw a change from around 1700fpm up to about 3000 or so, and the highest G load was 1.3, up from 1. This is exactly in line with the explanations I gave earlier tonight... I think people are assuming large V/S value = high G. It does not. I paid close attention to the speed trend vector too during this, it is never very large at all, which is what it would have to be in a situation where the plane was pitching up at some absurd rate. Even in the screenshot earlier in this thread with the 4200FPM, look at the trend vector - it's not huge at all. You have to realize that even a very small change in pitch attitude up at altitude at those speeds is going to result in a rather large V/S change.

Tabs,Here is an example of what people are talking about (its not my video):http://www.youtube.com/watch?feature=player_detailpage&v=W1DdQhWdXwwThe climb issue is on 11.00The descent issue is on 21.20I flew many times in cockpit of 737. I am not a pilot yet, but know all things you told about in previous posts.I fully agree with you and I nearly agree with people, who says about fsx weather problem. BUT, If you look at the wind and temperature indicators on this video, you will see that from 0fpm to 5000fpm the wind was the same 280/15 and -3C.The change of vert speed is predicted of course! Because basically lvl change uses that 10knots overspeed to boost rate of climb. But the reason, everyone posting here is actully the acceleration between 0fpm to 5000fpm. I really jumps up! I know 737 not badly, and I can say that this deffenetly not how vert speed changes in 737s.

I'm no confusing VS with g. In my case I wasn't seeing excessive VS in the climb, rather not enough. I have long since disabled turbulence and thermal effects as recommended, but was using real world weather (static) and was getting a lot of turbulence from that (it seems to add it even if you have turbulence disabled). As the aircraft climbed the turbulence caused it to lose airspeed, pitch reduced to recover speed, then it pitched up again to continue climbing, lost speed, pitched level to recover, and so on in a fairly long period oscillation. It was taking an age to get near the cruise flight level. I paused the sim and changed the weather selection to "customized" (without changing anything else). When I unpaused the turbulence was gone, the pitch oscillation disappeared immediately and the climb continued correctly. I assumed this was because of the turbulence stopping but it could have been due to any effect introduced by the real world weather.It seems to me that the VNAV speed control law works well as long as there are no wind or airspeed fluctuations, such as from turbulence. It's not a big loss to disable turbulence to avoid this, but it does mean using real world weather is an issue.Kevin Hall

Now - taking into account all of this knowledge - is there still an assertion that something is wrong with our autopilot? If so, what exactly is it that does not violate any of the understanding above?

I do agree with you posting and if I may, I would like to make the following suggestion:We know there are some outside factors from the weather changes simulated in FSX. The changes of wind and turbulence are more severe that you typically find in real life - except really bad weather weather at which the real AP also has his limits.But for the typical small changes a few knots in speed and a few degrees in direction makes the airspeed sensed and displayed moving around a few knots. While certainly an AP in initial climb or on final needs to control this tightly, I thing the AP at climb and cruise speed is too much chasing the target speed. This resulting is stronger pitch movements or throttly controls than necessary as the aircraft is flying fast and response on trim movements are much stronger that during the low speed phase. So my suggestion to overcome those observations: I would soften the AP control sensitivity at higher speeds so that the ride becomes smoother. I have seen in real live and you find it also on videos such deviation of 3-5 kts in cruise and the AP slowly compensates them. The current one in the NGX is just too firm at high speeds.The same is also valid for the bank control.The banking into the curve is just smoother in real life.

After going through this thread it seems people are confusing their real-world 37 passenger experiences with no way to really correlate what the see/feel to what's actually happening up front.Tab's hit in on the head: you're not playing with thrust in the climb - this applies to your bare-bones 150 just as much as your big 747. Pitch is adjusted for airspeed. That's how it works.Yes, autopilots can be aggressive some times. This goes for the RW too.... If you don't feel 'comfortable'...engage TCS and fly with that. There's all sorts of tips-and-tricks RW pilots use as well with the autopilot; it isn't just a VNAV/LNAV engage and let the thing do it. Playing around with heading select, bank angles, LVL CHG (which is used more then people think I believe) are all ways to help mitigate wonky autopilot behaviour. That's why they are there.

The NG's - the smaller variants at least - are rockets. That's how they are. These aren't huge airplanes gang.You got a -600 with only 75 people onboard and a light 7-8000 lbs fuel load for a rip up to Yellowknife and you'll scream up to altitude in no time. Couple WJA flight's I've been on we've haved no problem getting up to 410 on the southbound 1.5 hour leg. Fuel's not even worth tankering around any more nowadays

It is a bummer that even with funky weather programs we all have the occasional temperature inversion or phantom thing that throws things out of wack. If you gave a RW autopilots a 5 degree temperature change and the sudden 50 or so knot IAS change it would probably freak out too....in fact it would probably just disconnect.The Dr.'s put in some rediculous math into the autoflight system in this thing and I think it's all we can expect. Asking them to smoothen it out would only result in us whining later about unresponsive autopilot commands and such.