VNAV spirals out of control....!!

[Guest Muppet22]

HiSee screenshot.I seem to be having some problems with the VNAV function of late.As you can see I have selected 310 kts in the MCP window for the initial climb until about FL270'ish to avoid an overspeed.It is my understanding of the VNAV logic that it should try to maintain the speed selected on the MCP FIRST, and then the VS will obviously allow to make the aircraft fly at this commanded speed.Most of the time this works, although on this occasion VNAV just wanted to climb with no regard to maintaining the speed. Speed when the screenie was taken was about 295 IAS and decaying rather rapidly due to the +4000fpm climb rate.Can't figure this one out, it seems to happen randomly and for no apparent reason.A few pointers:ZFW was 110.4 (about mid range)Flight was from LHR - BUD and I took off with 17tonnes of fuel.CLB2 was used until about FL070, then CLB1 till about FL200, then CLB aS PIC doesn't wash out the de-rate automatically.Only a light headwind, no sig turbulence and no sudden changes in SAT which happens to me sometimes and causes the TAS to fluctuate significantly.Original PIC airfile, with Lee H's merge used.And to end I'm pretty sure it would have ended up stalling if I didn't intervene by selecting a more appropriate VS.Any thoughts/comments are of course welcome!edited for poor spelling

[crvm]

A couple of things here. Although you can set the active speed mode through the MCP speed window, it really is not a VNAV commanded speed unless you change it through the FMC. Here's how it works in this case.1] If you wish to climb at a restricted speed or modified climb speed or any other reason you might have through the use of the MCP, first select the WHITE ROUND button on the MCP and NOT the SPEED BUTTON. Once you select the round button, it will unblank the spd window and show current speed, not the commanded speed, and on the FMC's VNAV pg it will show a changed speed mode thus "ACT MCP SPD CLB", now just set the speed you desire on the MCP speed window. You don't have to hit the speed button to command the newly set speed, it will follow whatever you input in the spd window. To return to VNAV commanded speed just reselect the WHITE ROUND BUTTON, the SPD window will blank once again and on the FMC's VNAV pg it will show " ACT ECON CLB" etc.2] If you choose the SPEED button, it will also unblank the MCP speed window showing current speed and ENGAGE V/S MODE, and on the FMC's VNAV page it will also show "ACT MCP SPD CLB". So really you would no longer be using VNAV whatsoever here. You could also select FL CH if you desire for climb.3] It's better to use the FMC to modify you climb speed although in my first example you can set a speed temporarily and then return VNAV command but pushing the round button. Addition- 1] Remember that it does not "hold" the mcp set speed once the window "Blanks" but always returns to the commanded speed set on the FMC. VNAV does not "work" with any set window speed unlike the ALT window, which does restrict the altitude even though the FMC is set higher or lower than what's in there.2] In your screenshot, it appears to me that you are going to slow, so that's why the nose is raised at such an angle whereas it would be around 2000FPM or so, but it's trying to keep the mcp set speed. It's possible that you are a bit heavy for your climb or some other reasons effecting this, so what I would have done was to select V/S to lower the angle of climb so it could catch the selected speed without such a high angle of attack.[h5]Best Wishes,Randy J. SmithSan Jose Ca[/h5][h3]" A little learning is a dangerous thing"[/h3]AMD [pink]XP[/pink] 2200 |MUNCHKIN 512 DDR RAM |ECS[/b ][i] K7S5A MB[/i] |GF3 64 MEG |WIN XP PRO |MITSUBISHI DIAMOND PLUS 91 19"

[Guest ilh]

Randy, but he isn't in SPD mode. VNAV is still illuminated on the MCP, and the fact that the speed window is not blank means he has used speed intervention. The real question is why is the A/P not reducing pitch to gain speed? It should be commanding pitch to hit speed targets.Lee Hetherington (KBED)

[crvm]

Lee, he is in "active MCP speed mode" during his climb and VNAV will stay illuminated. The problem is that the aircraft is sill trying to achieve the vnav climb but his set mcp speed is lower than his current speed, so the aircraft has to have a higher angle of attack to maintain the two things going on here. If he raised the mcp speed, the aircraft would have to lower the agle to speed up. Since he was using the non-Pegasus ADI without the current speed showing, it's hard to guess at what speed he was at. But I assume that it was faster than what was set on the MCP so the aircraft was pitching to slow to the commanded speed. In the case that he was slower than the set MCP speed, then the aircraft could not maintain the current climb mode ((VNAV)) with the set speed. So for this particular climb, I would have changed ((climb)) mode to V/S to control the FPM and then selected the desired speed, paying attention closely to the aircraft's preformance at this altitude. [h5]Best Wishes,Randy J. SmithSan Jose Ca[/h5][h3]" A little learning is a dangerous thing"[/h3]AMD [pink]XP[/pink] 2200 |MUNCHKIN 512 DDR RAM |ECS[/b ][i] K7S5A MB[/i] |GF3 64 MEG |WIN XP PRO |MITSUBISHI DIAMOND PLUS 91 19"

[Guest Muppet22]

Hi Randy!Thanks for your answer. I was in fact always at a lower speed than what I selected in the speed window. By the time the plane had reached about 305 IAS then VNAV started going crazy, increasing pitch with no intention of maintaining the speed I set. In the end (just after the screenshot was taken) I switched to VS mode to prevent the plane from losing to much speed.Weight was probably not a factor as I only had a TOW of about 127.0 t, which is quite some way off the MTOW.I'm off to do some more investigating, cya!

[Guest ilh]

Randy, his EADI, despite not having the speed tape, clearly shows he is below his target speed (speed indicator below center, down near S), thus I do not see why it is pitching up so much. Also, my understanding is VNAV makes no attempt to fly a particular path on a climb, just a speed, thus "still trying to achieve the VNAV climb" doesn't make sense to me. It is in VNAV, just with a different target speed, and it is clearly not trying to maintain that speed by dropping the nose a bit.Lee Hetherington (KBED)http://vatsim.pilotmedia.fi/statusindicato...tor=OD1&a=a.jpg

[crvm]

Yes, not the correct statement there, I agree. Vnav is very complicated as you are well aware. Even some real world pilots have trouble explaining in detail the inner workings of Vnav. I know that the FMC's set crz altitude is Vnav's target and the FMC uses ECON SPD as calculated by weight, tempt and C/I for a vertical path to the T/C. Why it would climb at 2000FPM rather than 2200FPM, i'm not 100% sure. I could guess and sometimes I do. This is just an example of course. Hell i'm just a sim pilot hehe. But back to the issue at hand. I cannot create a situation like his no matter what I do. So I am stumped. Well be assured, in the real world many pilots have said something like this about auto flight "Why did it do that!!??"[h5]Best Wishes,Randy J. SmithSan Jose Ca[/h5][h3]" A little learning is a dangerous thing"[/h3]AMD [pink]XP[/pink] 2200 |MUNCHKIN 512 DDR RAM |ECS[/b ][i] K7S5A MB[/i] |GF3 64 MEG |WIN XP PRO |MITSUBISHI DIAMOND PLUS 91 19"

[Peter Clark]

Greetings,Looking at the indicators I agree with Lee that this looks like a standard LNAV/VNAV CMD engaged climb with speed-intervention mode active (see VNVAV SPD mode annunciation active)..I can often reproduce similar behaviour when climbing into 'turbulence' - wind shifts for direction and speed showing on the ND, AIS needle fluctuating, etc. PIC doesn't seem to do a good job of analyzing the averages of the weather going on around it, so it seems to instead latch onto a climb rate (my experience is that it's whatever the climb rate was when the turbulent layer was encountered, I guess in feet/min) and sticks there, completely ignoring the ever decreasing airspeed. In my experience once you get through the 'turbulence' (and I use 'turbulence' in quotes because I don't think what Microsoft calls turbulence is anything like what real turbulence must be as viewd from the front of the aircraft, could however be wrong) and the airmass smoothes out it's almost like the aircraft goes "Oh, oops, look - I'm down to 190 knots at Fl200, perhaps I should pitch down" and promptly does and from there on out (until descent when encountering another turbulent layer) will be quite happy. Just set up a bunch of clouds, set turbulence sharp in FSUIPC and climb through them.Static screen shot might not be enough to convey this behaviour being the root cause here though. Easy enough test to reproduce (at least for me).Luck!P