|
Post by connieguy on Oct 17, 2019 9:52:08 GMT -5
Many thanks, Manfred.
|
|
|
Post by connieguy on Oct 18, 2019 8:59:14 GMT -5
Hello Manfred. Below is the trail of my second trial flight. As you can see this one was much more successful, and again it was flown manually with the gauge closed. I think the problem on the first one was that I simply didn't turn quickly enough, because I am fairly certain that I had the intercept distance set correctly. In flying manual approaches and landings it has always been my practice not to make rather tight turns, but this kind of procedure requires them and as you can see I was much more successful at the second attempt and PAR did then direct me to the centre line. I think on the first flight I was simply too far away from it for the gauge to be able to work properly, which is fair enough given that we are told that the aircraft needs to be lined up properly already. However there is an issue that I will mention when I was very near to landing. The following two screenshots are of the view out of the cockpit and the Plan G map of where the aircraft was at that point. Basically I was to the right of the glide path but PAR was still telling me to turn right even though I really needed to turn left to bring me back to the centre line. At this point I ignored PAR and turned left anyway, then straightened up and landed successfully. I guess that if I had carried on PAR might then have advised a more abrupt turn left at a later stage, but this would have been more difficult especially if there had been very poor visibility. It could be - you will know better than I - that the gauge needs a little more fine tuning at this stage.
And the third flight, from Orly to Heathrow via the NDB at Epsom, with the intercept distance set at 8nm because I forgot to change it to anything else.
In the final stages of the approach the same thing happened. PAR was encouraging me to turn right when I could see the runway and that this wasn't the right thing to do, so I kept left. That is why the final track looks so straight, but you can see from these screenshots that at one point I was to the right of the runway and glide path in just the same way I was at Orly. I might add that on a flight in good visibility I would normally switch off PAR for a visual landing once the runway was in sight, and therefore for most purposes the gauge is fine as it is. However, as I said above, I have doubts whether it would currently work for a manual landing in poor visibility, although of course I haven't tried it. If it was tried with the autopilot on and landed successfully it would interesting to look at the track of the final mile of the flight. Hope this helps.
|
|
|
Post by mjahn on Oct 18, 2019 12:28:55 GMT -5
OK, this looks a bit better, but it is still a far cry from what I expect it to do. Actually, if it was as unreliable as pictured, it would be of little help and an awful stress trying to land in zero visibility. But maybe you should try it with AILA/AP open because then the data readouts make it much easier to troubleshoot. Would wind force be a factor in the equation? Also please give the auto learn mode a chance, even if only to ensure that there isn't some odd scenery glitch, AILA somehow getting false data or something. Seemingly wrong right/left callouts can occur because the Controller must give his instruction relative to the current heading, and then this kind of thing can happen. I'll check, though...
|
|
|
Post by connieguy on Oct 18, 2019 12:57:28 GMT -5
Wind will not have been a factor and I used two different airports for landing, neither of which have had an issue in the past, while in both of them the final approach was offset to the right. Remember that these were manual landings and that in addition to paying attention to the heading I also have to reduce power and increase trim to slow the aircraft down and then lower flaps and gear, all of which requires further alterations to the trim, engine manifold pressure and rpm. However, on the last flight I had done most of this before PAR got in touch. The wobbly look of some of the flight paths is down to it being manual control rather than an autopilot landing, especially when it comes to turning 90 degrees. I still think the gauge is doing a pretty good job except within the last mile or so, but that is what would make it very problematic in poor visibilty.
|
|
|
Post by Tom/CalClassic on Oct 18, 2019 13:16:10 GMT -5
Hi, I made a downwind approach to KLAX in a CV-340 using ALT but not HDG. That allowed me to handle the heading myself. APC put me about 4 NM away from the runway centerline, about 1 NM closer than the previous approaches. I used the default 12 NM circuit. I followed the voice instructions except for the PAR controllers initial heading, which was 068 - I turned to 064 to get to the centerline faster, since I had rolled out to the right of the runway centerline. I think this one call could be improved a bit? I turned at 15-25 degrees bank, nothing too serious. Here is my track: As you can see it was pretty easy to make the turns in plenty of time for the controller to guide me to the runway. Hope this helps,
|
|
|
Post by Tom/CalClassic on Oct 18, 2019 13:40:50 GMT -5
I just tried it again, this time purposely turning too wide on final (but not as much as Ken did). Like in Ken's example, PAR then pointed me towards the runway, instead of turning me sharper to get to the centerline sooner. If I would have stayed with his instructions I would have intersected the runway centerline just before touchdown.
Hope this helps,
|
|
|
Post by Tom/CalClassic on Oct 18, 2019 14:01:16 GMT -5
BTW, when the PAR controller starts saying "on centerline, on glidepath" (which he says very rarely when the AP HDG mode is active but says a lot when it's not), he repeats it almost constantly. Those could certainly be spread out a bit.
|
|
|
Post by Tom/CalClassic on Oct 18, 2019 14:08:05 GMT -5
Hi,
I tried the approach again with a larger aircraft (DC-7C at 107,000 lbs at 150 kts, flaps 2) and again didn't have a problem making the turns to base and final, using a 15 deg. bank (they looked wider on the map, but still much the same as they did above). The PAR controller still wanted me pointed directly at the runway, even though I was left of the centerline this time.
|
|
|
Post by connieguy on Oct 19, 2019 4:05:38 GMT -5
Thanks Tom. I agree on the frequency of some of the calls. Alice's heading calls I think come at exactly the right frequency, but the PAR glidepath calls when in manual mode come more or less one after another and I think could be spaced out more.
|
|
|
Post by mjahn on Oct 19, 2019 6:07:40 GMT -5
Okay, thanks very much for the feedback. The frequency of the calls I don't mind that much, it's always good for the nerves when you are told you are properly lined up, isn't it. The calls are officially supposed to come in 5 secs intervals, I don't even think our PAR is far off there. Anyway, a minor point I believe. To put PAR under extreme stress I redid an approach to LFPO runway 02, setting only the default 8 mile intercept, and straying three (!!) nmiles to the right even after passing the intercept (directly over LFPY), a condition that no PAR would accept under normal cicumstances (I think). And yet, as you can see even with only 6 miles to go, the PAR instructions definitely got me back to the centerline (not the runway), and was was even good enough to hit the centerline about 1.5 miles out, which just made a rough landing possible. In fact, knowing, as a pilot, that the rwy heading is 019 and that I am given correcting courses to something like 345, I could know that I was miles off (a look at the AP would confirm) and make a much tighter turn to the runway myself. PAR could be made to issue a similar instruction, but I think the far better solution is to make sure to proceed from a better alignment to the intercept, and maybe advise a goaround, when for some reason or other, the offset from the centerline at the intercept exceeds a critical distance. Par's initial course corrections may be a bit odd, because he first sets us a course to the intercept, but then tries to move to the centerline. I'll look into that and see what gives. Opting for a high realism challenge from the beginning means setting oneself a near impossible task, being on your own without an AP or even a co-pilot. In real life, I well remember, Controllers asked everything with wings and wheels doing an approach if they would please agree to do a GCA for practice purposes, mainly for Controllers' practice, they said, but quite possibly they meant pilots, too. EDIT: hang on, I, too, am getting some odd results now. So back to the drawing board ... will keep you posted, hopefully with an update.
|
|
|
Post by mjahn on Oct 20, 2019 11:11:37 GMT -5
Found one nasty and critical bug, now hopefully corrected, thanks to the pointers given here. Also increased the callout intervals a bit. New version here: www.dropbox.com/s/eo586yl0j486x11/Generic%20GCA_V2.1_03.zip?dl=0And here is a runup to LFPO 020 in semi-auto ('green') mode, standard 8-mile intercept, which looks more or less as it should:
|
|
|
Post by Tom/CalClassic on Oct 20, 2019 12:19:57 GMT -5
Success! The same approach as I did above brought me right to the runway centerline soon after turning final. Nice work.
|
|
|
Post by Tom/CalClassic on Oct 20, 2019 13:06:53 GMT -5
Hi again,
I would like to set my intercept distance to 5.5 NM, since this gives me a more typical 1900 ft AGL pattern. The problem I'm having is how to do this. Currently I set up my runway, then enter 0.55 into the number boxes and press INT. Then I click the blue SET word. This sets the approach at 5.5 NM. But when I then click the APC button the distance changes back to the default 8.0 and if I don't click SET again before Alice starts speaking I get the default 8 NM intercept. There is a relatively short window between clicking APC and clicking the SET again to get my desired intercept distance. Is this the intended behavior?
Thanks,
|
|
|
Post by Tom/CalClassic on Oct 20, 2019 13:46:29 GMT -5
I figured it out. In the apsmall code you limit the intercept distance to 6 NM, and if set to less than that it is reset to 8 NM. I just changed that lower limit to 5 and it works fine. But wouldn't GA aircraft have shorter intercept distances than that, typically at pattern altitudes (1000-1200 ft AGL)? Anyway, setting it to 5.5 NM seems to work fine so I think the limit can be smaller than 6?
|
|
|
Post by Tom/CalClassic on Oct 20, 2019 13:59:31 GMT -5
I just tried a 3.5 NM intercept distance (1150 ft AGL) flying a CV-340 and it worked fine. Is there a problem with shorter intercepts?
Edit: tried it again with a 3.0 NM/~1000 ft AGL intercept, which also worked fine on AP. Much less than that and I'm not sure my CV-340 would be able to deal with it, but a GA aircraft should be fine using this setting.
And I tried a 5.5 NM intercept flying the DC-7C and that worked fine too.
|
|