738 VNAV and trim destroy NDB approach

[Guest X-ray]

Hello,Suppose a (wonderful) NDB approach on rwy 08 that involves flying to the outer marker of rwy 26 first (desc to 3000ft), then continue rwy heading 259 and overfly rwy 26 (desc to 2000ft), then when crossing the middle marker of rwy 08 make a right turn 28 degrees, and eventually after about 4nm make a large left turn 208 degrees (desc 1400ft) to get in heading 079 on a final path towards the NDB (or compass locator) and rwy 08 (see attached picture).I let the autopilot with VNAV and LNAV armed fly the first part of the approach (until I'm aligned with rwy 08). I disconnect the autopilot when I'm on final course with the runway in sight. The NDB('s) are tuned.Sometimes this goes all perfect. I disconnect the autopilot at about 1500 ft, runway in sight, visual landing. If wx is bad and I can't see the rwy at that point I hand-fly the aircraft by following the NDB needle and I descend according to the chart until I see the PAPI and rwy.However, most of the time I experience this very annoying issue that destroys the whole approach, makes me execute a missed approach and renders the autopilot in an uncontrolable state (I have to hand-fly the aircraft to the rwy from a large distance).During the phase where the LNAV makes the large left 208 degrees turn the VNAV has the destructing habit of suddenly changing the trim setting (nose-up, and a *LOT*). As long as I don't disconnect the autopilot all goes fine, but when I disconnect the 737 starts climbing (on approach so low speed, I can avoid a stall but it comes very close) and I can forget about that wonderful visual landing. I have absolutely no clue to the whyabouts of this trim change. It never does this during ILS approaches for example.I tried changing the trim again to a normal setting after the VNAV changed it but it only makes it worse (i.e. the autopilot disconnects immediately and the aircraft just ignores my vertical yoke movements or it starts diving or climbing). I think the descent (from 2000ft to 1400ft) during that 208-degr-turn and the deployment of gear and flaps during that short phase is contributing to this issue.Strange about this issue is: 1) I never experience this on ILS approaches, always on VOR or NDB approaches, also on other NDB approaches than the one I explained above. 2) It only happens 80% of the times, the other 20% are perfect VOR/NDB approaches and landings.When this problem does not occur I also haven't heard the VNAV changing the trim. VNAV didn't change the trim at all in that case.Together with the spoilers issue, this is the only problem I'm having with the 738. I couldn't find this issue documented anywhere. Any suggestions to fix this and put the fun back in NDB approaches???Thanks a lot!!!BartSorry about two errors in this pic: occurences of "140 knt" should be read as "Vref + 5 knt" (or other wind correction speed). "vertically controlled by rudder" should be read as "vertically controlled by yoke"http://forums.avsim.net/user_files/137949.jpg

[Guest X-ray]

I reproduced the problem today without any difficulties ;-(.Any hints? What can trigger VNAV to change the trim?Regards,Bart

[crvm 0]

Sounds like you are still in APP MODE? Turn off the A/P and F/D. What I believe that souns like is the BIAS TRIM being applied for autolanding at 400 RA.[h4]Randy J. Smith[/h4]AMD 64 4000+|ASUS K8V DELUXE|SAPPHIRE ATI X800XT PE|MUNCHKIN 3200|80 gig SATA|DELL 1905FP 19" LCD|TRACKir PRO|PFC JEPPESEN MOONEY YOKE|CH PRO PEDALS|

[Guest tmetzinger]

Shouldn't be in APP mode on an NDB approach.When you make a large turn like that the autopilot will automatically roll in a significant amount of up trim to maintain the programmed path, and your autothrottles should spool up to maintain airspeed.What I'd suggest is that you try leaving it in LNAV, but manage pitch with SPD/VS, at least until you are straight and level inbound to the NDB.Another option is to track outbound farther from the NDB and then performing a procedure turn. You can program this in the FMC (sort of) by placing multiple Place/Bearing/Distance waypoints in the legs page to lead you away from the NDB and then come back in. Try and make sure you have at least a three mile straight in segment to the NDB.If you can give me the airport and approach in question, I can try to see if I can give you better advice.

[Guest X-ray]

Hello Randy and Timothy,Thanks for your replies!APP is indeed not armed on this approach.Timothy, your suggestions make sense, I'm going to try them out and I will report here. Thanks!!This is the approach I'm referring to: http://www.ivao.be/files/charts/EBOS_IAC01.pdfIt also happens on other NDB approaches to other airports, but this one is a very good example.ONO is the IAF, on my original drawing I left out the other large left turn on the right side of the map because I do not have to make that one (I'm coming from the east to the airport, cfr altitude section of chart).Thanks for your help, regards,Bart

[Guest tmetzinger]

ok, it's bedtime for me now, but I'll look at this first thing in the morning.

[Guest tmetzinger]

ok got the plate up.since I don't see a SID/STAR file for EBOS, then I would program the approach into the FMC as follows:Starting from the IAF, ONO:ONO 200/2000DD 185/2000 (you're a CAT C or D aircraft in a 737)DD287/05 185/1400 (defines the start point of the turn)DD259/06 185/1400 (defines the end point of the turn)KOK035/6 160/1400 (defines the FAF)DD /500RW08Then, I'd fly it in LNAV and VNAV until I reached 1400 feet. At 1400 feet I'd switch to ALT HOLD/SPD/LNAV through the turn, and then reduce the speed, lower the gear, set approach flaps as you approach the FAF. Then I'd do a VS descent at 800-900 feet per minute to DD at 500 feet.Give that a try. You could try it in VNAV the whole way through, but be ready to toggle VNAV off and on again if the airplane doesn't descend the way you want - there's a bug that sometimes causes VNAV to not descend at the right vertical speed. Looking at the DES page as you go through the approach is a good way to check.Best wishes,

[Guest X-ray]

Hello Timothy :-),Thanks for your proposed approach. I'm going to try it out tonight.I tried your other suggestion yesterday. I turned off VNAV and I let spd and v/s control the vertical navigation but it triggered the same problem with the trim. However, this time the autopilot didn't change the trim at the beginning of the large turn but rather at the end (was almost aligned with the runway).There is a chart with the STARs (http://www.ivao.be/files/charts/EBOS_STAR01.pdf) but since the STARs do nothing more then guide the pilot to the IAF I assume it doesn't change anything to what you've written above. I'm coming from DENUT or FERDI. All charts are listed on http://www.ivao.be/charts.php?page=ebosThanks again, I will report how it went!Regards,Bart

[Guest X-ray]

Hey Timothy,I just did a perfect NDB approach on EBOS :-). Visibility was low and I couldn't see the runway at the FAF. I disconnected the autopilot and the aircraft was listening to my yoke movements just as it should be. I had no problems following the ADF needle to DD. I came in a bit too high but could correct it easily and touched down at perfect speed with two red and two white PAPI lights :). I programmed the FMC just as you proposed, it worked like a charm. I heard the autopilot changing the trim nose-up again when I switched from VNAV to ALTHOLD/SPD at the very beginning of the turn but he only did it once instead of many times, which didn't affect the controllability of the aircraft after disconnecting the a/p. The FMC also gave a STEEP DES AFTER DD message (when I was DD inbound after the FAF), but that turned out to be no problem either.Thanks a lot for helping me out! This will put the fun back in challenging NDB approaches!Until now I was relying on the approach procedures of Navdata. Since I also had this problem on other airports I will start to program my own approach in the FMC according to the appropriate charts instead of using Navdata's on the airports where this problem occurs.I have the impression that it's better to keep a constant altitude and speed during a very large and short turn in order to avoid stressing the autopilot too much. At the beginning of the turn, you should already maintain the speed and altitude that you desire at the end of the turn. Is this correct, or to put it in another way, could you maybe, to conclude this topic, tell to which items one should pay attention when programming a non-precision approach in the 737NG FMC to avoid problems like this?Again, thanks!Regards,Bart

[Guest X-ray]

*

[Guest tmetzinger]

The autopilot can certainly handle a turn and altitude change at the same time, but the 737 is a slippery aircraft and it's often hard to slow down and go down at the same time without putting out spoilers or gear.