reported ILS and PRGM issues

[Newbie62]

I've seen the ILSN director cross behave really wacky.  In the track file I'll upload, timestamps refer to what is seen on the clock in the top right corner of the F10 map.  Alternate between F1 cockpit and F10 map to see that I was nailing the director cross (best I've done yet, actually) in cockpit view, but watch see how the ILSN guidance "guides" me off to the left of the runway and right into the bunkers.

In the beginning, I ran a user-created F-15C cannon tutorial.  Then on the F10 map I see friendly units positioned around Vaziani, so I head there (the mission doesn't have a return leg built in).  It's a long flight, ctrl+z and watch with F10 until:

12:28:49 - I have converged on the approach line, and given myself extra distance on account of the bad weather and two hills along the flight path

12:29:30 - I have engaged ILSN and located Vaziani toggling through waypoints using the F10 map distance to confirm

12:30:15 - the director cross is trying to take me down and to the right, which I'm ignoring for now, because I can see on F10 map that I am right on the approach line, and there are still a pair of hills coming up which I must clear, so I'm staying above the heavy cloud cover

12:31:47 - flaps and gear down

12:32:48 - cleared the first hilltop, one to go

12:34:50 - the director cross begins behaving, maybe I just acquired the ILS beam

12:35:04 - start following the director cross, descending into the clouds and zero visibility, 9.3 miles to go

12:35:44 - start picking up heavy crosswind, director cross and velocity vector are almost off the HUD

12:36:09 - I'm still nailing it, but at this point I am focused on following the director cross, and I wasn't keeping track of the F10 map.  In hindsight though, looking at the F10 map in track playback, it's starting to guide me off to the left of the approach line; at this point because the velocity vector and director cross are off the HUD, I have used Rctrl+Rshift+num6 to bring them back into view

12:36:39 - cleared to land by tower; while fiddling with the radio I got blown off a bit, so at this time, I have gotten the velocity vector back over the director cross

12:37:00 - speed brakes out to start bringing down the speed, still nailing it following the director cross

12:37:12 - speed looking good now, brakes in, still nailing it following the director cross, but in F10 map it is VERY apparent now that it's guiding me way off to the left.  Because of zero visibility and focus on instrument landing so not checking F10 map, I have no idea at this point.

12:37:30 - still nailing it following the director cross, which has by now taken me across the line of the taxiway parallel to the runway, heading for the bunkers

12:37:39 - first trees come into view, something is horribly wrong but it's too late to recover

12:37:40 - bunkers come into view, death by ILS

[Newbie62]

I don't see how to upload a track file into this forum topic, so I have uploaded it here: https://drive.google.com/file/d/12pYyJK_BcuJJKkfdTWle4WADvGeeCq2b/view?usp=sharing

[Ironhand]

When I watched your track, the ILS bar was indicating that you were to the left of the runway for most of the way down from at least 8 km out. Then it suddenly centered and red flagged as you flew completely out of the beam. It sounds like that's not what you experienced on your end but, for me, you ended up where the ILS bars said you would.

[Newbie62]

If you are referring to the yellow bars on the ADI, having watched the track again, I see what you're saying.  I was focused on keeping the velocity vector centered on the director cross in the HUD.  At 4.1 km out (12:37:01 on the F10 map clock) and shortly thereafter, I was perfectly on the director cross, but the ADI vertical yellow bar is off to the right.  With the plane being off to the left at that point but with the velocity vector being on the director cross, the yellow bar should have been heading back to center.  Instead, I was heading further off to the left, and yes, the vertical yellow bar in the ADI reflects that.

The HUD director cross is what I referred to.  It was giving bad guidance.  By the time the red flags popped out in the ADI, the director cross had already guided me to being mere moments from impact into the bunkers.

[Newbie62]

Okay here's another.  I see "drag files here to attach, or choose files" now, so attaching to this comment.

This time I ignored the HUD director cross and as you implied that I should have done, and focused instead on keeping the vertical ADI yellow bar centered.  I was staying a little high to avoid the hills, which can't be seen in the zero visibility conditions leading into Vaziani with this mission's weather.  I caught glimpses of them on the way in, clearing them by a few hundred feet, so I think I did well there.  Things were going well, but when I got to 3 miles from the runway threshold, according to the ILS bars in the ADI I was way over glide slope, and I figured at 3 miles I had better start coming down to meet glide slope.  Maybe even a little late for that, but anyway I went to 6 degrees down on the velocity vector to pick up the glide path.  At 2.2 miles from the runway threshold, you can catch the first glimpse of trees rushing up.  If you pause in that instant you'll see that the ADI ILS glide slope bar STILL says I'm still way over glide slope, whereas the director cross says I'm under.  It seems like ISLN is behaving as follows:

ADI bars:

Horizontal - good guidance (indeed, in this track, looking at F10 map I am dead on the runway centerline the moment before impact)

Vertical - bad guidance (into the ground 2.2 miles short of the runway in this track)

HUD director cross:

Horizontal - bad guidance (it took me over into the bunkers way left of the runway in the previous track)

Vertical - good guidance (both in the previous track, and presumably in this track since just before impact the director cross is showing me as being low, while the ADI is showing me as being high)

 

 

F-15C ILSN malfunction 2.trk

[Newbie62]

Here's an interesting variation.  I replayed that last track until my approach was set up, then used "take control" to try again on final approach.  I had taken note of the runway altitude of 2050 while reviewing another failed attempt, so that I could use the altimeter to manage descent into the runway threshold on my own.  That worked beautifully.  Interestingly, the ADI above/below bar (vertical aspect, horizontal bar) actually comes alive 1.2 miles from the runway.  So when I noted ADI "Vertical - bad guidance" (meaning the horizontal bar for above/below glide slope) it is apparently useless for most of the approach, but starts behaving itself if you get close enough.  Which is really hard to even get to that point when flying blind and the instruments aren't helping to get there.  It was trial and error.

Is this normal?

I brought the plane smoothly over the runway in this final track, set it down gently, and it promptly flipped on its side, slid down the runway, and exploded.  Probably because I was crabbing 20 degrees on account of the 19 m/s crosswind!

F-15C ILSN malfunction 2b.trk

[Ironhand]

I haven’t spent any quality time in the F-15C’s cockpit in about 12 years. Most of my time is in the Russian birds. So I’m less familiar with the information presented on western HUDs which is why I was focusing on the ADI and HSI. Those are familiar in any cockpit.

I’ve downloaded the additional tracks but haven’t had a chance to view them. Hopefully I will by the end of the weekend.

BTW, a 19 m/sec crosswind component is well beyond the 15 m/sec allowable for the Su-27 (my usual ride). I’m guessing that the F-15 would be similar. But that shouldn’t effect the instruments negatively.

In the Russian FC3 aircraft, the ILS needles don’t go live until 8 km (4 plus a little nm) from touchdown which may or may not be correct. Something similar might be going on in the F-15C pit.

[draconus]

I thought I know The Eagle pretty well but I took 2 quick tries to land in this setting and failed. The instruments didn't behave how I remember they should. Too little time to jump into conclusions though. This place is also the navigation and beginner's nightmare since there are 3 airfields in close vicinity and since the systems are simplified you can easily lost SA and mistake the airfields.

Realistically you would have to be diverted for both crosswind and visibility in this case.

Edited May 14, 2023 by draconus

[Ironhand]

@Newbie62

I still haven't viewed your subsequent tracks but I did take your original track, edited out all of the units and the clouds but kept the winds. I then added an F-15 to fly. The HUD took me right in to the runway. The landing went just about as well as expected...unfortunately.

Track attached.

Newbie62_F-15C ILSN_VFR.trk

Edited May 14, 2023 by Ironhand

[Newbie62]

The crosswind was obviously excessive.  I guess ATC isn't programmed to issue divert instructions in such a case.  So when draconus said "realistically" he must have meant real world.  As for three airfields in proximity, I got bitten by selecting the wrong airfield the first few times I tried it, until I figured out the trick of pausing, noting the distance to airfield indicated in the HUD, and switching to F10 map to see which airfield is at that distance.  In all the tracks I presented here, I had used that trick to get the right airfield.

@Ironhand, I tried watching your track, and there seems to be a glaring inconsistency in coordinates.  When I replay it on DCS world 2.8.4.39731, it has you descending into a field about 12 miles southwest of Vaziani.  So it's impossible to make out how the HUD director cross and ADI ILS needles behaved for you...

At a guess, when you added an F-15 to fly, the one you added was about 12 miles northeast of the F-15 in the original track, but your actions from that point were applied in replay as if you had taken control from where I was, so everything was offset.  I rewatched yours fast forwarding with F10 map zoomed out quite a bit to see if your added F-15 shows up, but no such luck.  It seems like track files and/or the replay engine can't handle adding and flying another aircraft.

[Ironhand]

@Newbie62

My apologies for wasting your time. I have been severely under the weather since Friday and I have no idea what I did in saving the track. I reviewed what I uploaded and that isn’t the one. Again my apologies. I’ll try and locate the right one and, if not, try flying it again.

[Newbie62]

Oh no worries.  I appreciate your attempt at confirming or disproving that there is a real issue here.  If there is a real issue, maybe it'll get the devs' attention.  If you do give it a go, might I suggest escape -> take control after fast forward playback to about 15 miles out?

 

To reiterate my observations:

ADI bars:

Horizontal (horizontal guidance, vertical bar) - good guidance (indeed, the F-15C ILSN malfunction 2 track, looking at F10 map I was dead on the runway centerline the moment before impact)

Vertical (vertical guidance, horizontal bar) - bad guidance (into the ground 2.2 miles short of the runway in that track, but as observed subsequently in the 2b track, it came alive at 1.2 miles)

HUD director cross:

Horizontal - bad guidance (it took me over into the bunkers way left of the runway in the first track shared via google drive)

Vertical - good guidance throughout approach, well beyond 1.2 miles

 

Edit: on another try just now, I managed to pick up the ILS beam (both needles alive in ADI) at about 2.5 miles, double what I saw before.  I wonder if the horizontal guidance farther out (10 miles or so) is coming from TACAN, with the ILS beam becoming available when you get closer, but ONLY if you get into the beam both horizontally and vertically.  If that's all normal then the only real issue what happened n the original track, when the HUD director cross guided me off course and down into the bunkers...  Maybe the trick is to trust the director cross until both ADI bars come alive, at which point, trust them and not the director cross.  If that's expected and by design then, well it's awfully complicated but that's part of the charm of DCS!

Edited May 15, 2023 by Newbie62

[Ironhand]

4 hours ago, Newbie62 said:

Oh no worries.  I appreciate your attempt at confirming or disproving that there is a real issue here.  If there is a real issue, maybe it'll get the devs' attention.  If you do give it a go, might I suggest escape -> take control after fast forward playback to about 15 miles out?…

Oh, there’s a problem. Just don’t know what combination of factors causes it. It’s not the wind alone. Unfortunately, every time I fast forward your track, it breaks in some way. So that’s not a viable option.

I did find the track I thought I was uploading the other day. It’s attached below (F-15C HUD Preselected ILSN Vaziani). Like I said in an earlier post, I removed everything except the wind and, then, added an F-15 starting closer to Vaziani to fly. It may be important that I selected Vaziani as the landing runway in the Editor because ILSN took me in perfectly. My landing/crashing at the end, though, was anything but perfect. I did find the runway, however.

I set up another mission with the same winds without a landing runway selected and flew again (using an approach position closer to your own)—F-15C ILSN Non-Preselected Bug. This time, I ended up to the right of the runway as opposed to ending up to the left as you did.

So something is wrong somewhere. I made a quick baseline flight into Butami. There was no wind and Butami was not selected for landing in the Editor. It worked perfectly. So…

1) It’s not wind alone since both flights into Vaziani had wind. One was successful. One was not.

2) It’s not preselection of the runway alone. Both Vaziani and Butami had success.

3) It’s not Vaziani alone simply because there was success at Vaziani.

EDIT: Just noticed your edit. No, nothing that complex should be necessary. While you should always be in the habit of cross checking your instruments, you should be able to follow the HUD all the way down.

 

F-15C HUD Preselected ILSN Vaziani.trk F-15C ILSN Non-Preselected Bug.trk

Edited May 15, 2023 by Ironhand

[Newbie62]

@Ironhand "Unfortunately, every time I fast forward your track, it breaks in some way" - I read elsewhere about that as a bug, if the ff speed is too great, it breaks.  I have been getting away with about 16x, more if I'm in F10 map view and nothing is going on (the long leg of the flight back to Vaziani), when I try a "do over" of the approach and landing attempt.

An aside on terminology: what I've been calling director cross is actually called the landing steering cross (I just looked it up in the manual).  I'll just abbreviate it as LSC for accuracy.

Looking at your tracks where you tried to reproduce the issue in isolation without the cannon training and long flight to Vaziani, there are a few things that I noticed.

 

F-15C HUD Preselected ILSN Vaziani.trk:

I suppose because Vaziani is preselected, while you are in nav mode, you are following a bank steering index (the vertical bar, not a LSC yet) in the HUD.  When it transitions automatically to ILSN after crossing waypoint 1, it remains a BSI and ILSN guides you to another waypoint.  When you cross that next and last waypoint, then the HUD switches from BSI to LSC, and the ILSN NAV info block shows range to the runway threshold, starting at 9.4 miles.

At that point, as you're straightening out and stabilizing your approach, between 9 and 6 miles out, notice that the ADI vertical guidance (horizontal bar) is telling you to go down.  If you had followed it, you would have gone right into the ground.  But the ADI lateral guidance is good - inertial or TACAN, who knows?  Also in this time, you're climbing to glideslope because the LSC is good.  For now.

Then about 8 miles out, you've come up to meet glide slope and you begin to descend, following the LSC.  So far so good.

At 5 miles out, you're following the LSC perfectly, but it is starting to take you left.  This is apparent both in F10 map, and the ADI lateral guidance (vertical yellow bar).  ADI vertical guidance (horizontal yellow bar) remains bad.  It's still telling you to nose down into the ground.

At 2 miles out, check out the F10 map.  Following the director cross is taking you on the same path that I originally experienced - off to the left of the runway, towards the bunkers.  ADI vertical guidance (horizontal yellow bar) remains bad.  It's still telling you to nose down into the ground.

At 0.1 miles out, you're actually past the runway threshold.  0.1 miles is how far to the left of the runway that the LSC brought you in to.

At that point you apparently realized that you were left of the runway, and you demonstrated amazing altitude control as you held just off the ground, veered right, straightened out over the runway, and touched down.  You almost stuck that landing...

I disagree that ILSN took you in perfectly.  Had that been the case, you would have been settling in over the runway threshold, not 0.1 mile to the left of it.

 

F-15C ILSN Non-Preselected Bug.trk:

Once you had Vaziani selected in ILSN mode, the LSC actually gave you good route guidance to get on the approach, both in altitude to meet glide slope and and azimuth to meet the approach path.  At about 4 miles out, ADI started giving vertical guidance.  At 2.0 miles out, you were exactly on glide slope, both horizontally and vertically, as can be seen in the ADI needles, but your heading was off to the right, because you were following the LSC, and it was continuing to take you off to the right of the runway.  You wouldn't know this in zero visibility conditions like your "preselected" track or my runs, but with no clouds you can see it happen very clearly.  Why it took you off to the right rather than off to the left - or for that matter, why it doesn't take us into the runway as it's supposed to - remains unknown.

 

I believe that you have reproduced exactly the issues which I raised here...

[Ironhand]

You're right. I forgot that, when I make an ILS landing with a strong crosswind component in the Su-27, the HUD indicators keep me looking out the side of the canopy straight down the runway's center line. It doesn't "circle" me in like what's happening in the F-15.

[Newbie62]

Thank you.  Now that it has been independently confirmed, hopefully this forum topic gets the devs' attention so that they might look into it.

[Ironhand]

@Newbie62

Just to add to the fun (evidence), here’s a comparison of the F-15’s and Su-27’s “ILS” ground track using the same mission. The F-15 is the aircraft that surges ahead at the start of the video and, then, stays to the left (above in the video) of the centerline while on the glideslope. (I was fast forwarding each track at different times during recording which is why an aircraft slows down/speeds up. Video is no real time.)

You may want to change this thread’s title to something like F-15’s ILSN does not keep aircraft on centerline in crosswind conditions.

To me it almost feels like it keeps you aimed at a point on the runway rather than points on the glideslope. Without a crosswind, you might not notice it.

Edited May 16, 2023 by Ironhand

[Newbie62]

Wow, you really went all out on this.  Thank you!  At this point, it's pretty much settled that there's something wrong, although we can only guess at the root cause.

The only question now is whether the devs will take note of this forum topic and do something about it...

For now, I won't be trying to land the F-15 if I can't see the runway.

[Ironhand]

PM Flappie and ask him to check this thread, when he has a chance. He’ll let us know, if there’s anything else he needs.

Creating the video only took a total of about 20 minutes. So I’d hardly call it going “all out”. Just thought a visual aid might have more impact.

Edited May 16, 2023 by Ironhand

[Newbie62]

I don't know Flappie.  Anyway thanks for confirming the issue and cooking up that visual aid.

[draconus]

3 hours ago, Newbie62 said:

I don't know Flappie.

You don't have to know him personally. Here we can ping him @Flappie so it can be looked at.

10 hours ago, Newbie62 said:

For now, I won't be trying to land the F-15 if I can't see the runway.

That's what I meant before by "realistically" if you want to stay safe and simulate RL The F-15C wasn't qualified for no-vis landings afaik.

Keep in mind there's also HSI which is actually the best in this case in keeping straight lineup with RWY - only with the right airfield selected and ILSN mode.

[Flappie]

Hi gents, I'm currently on vacation far way from my PC. I'll add this thread to my to-do list right away.

Thank you all for the detailed posts and the tracks.

[Ironhand]

Enjoy your vacation!

And, BTW, there are some Su-27 ILS issues, too. More later in a new thread.

[Newbie62]

9 hours ago, draconus said:

Keep in mind there's also HSI which is actually the best in this case in keeping straight lineup with RWY - only with the right airfield selected and ILSN mode.

Yes, HSI worked perfectly, every time, so it never came up for discussion.  For that matter, ADI lateral guidance worked perfectly every time.  The issues are ADI above/below guidance not coming alive consistently, and the HUD LSC's lateral guidance leading the player off to one side or the other by a fair distance, which seems to be caused by crosswind.  One would expect that as crosswind starts to blow the plane off course, the LSC would respond by guiding you into the wind and back towards the approach vector.

I'm not sure how far out the ADI above/below guidance should come alive but it was observed to vary between "never" (maybe if the player has already followed LSC outside the beam by the time one gets into range), or coming alive at an inconsistent range between 1 and 4 miles even if you are presumably in the beam.  On my later attempts when the nature of the issues became more clear, I got into the ILS beam by following ADI for lateral guidance and LSC for above/below guidance.  Even then, ADI vertical guidance was not consistent about the range at which it comes alive, if it does.

[Newbie62]

One final thought until Flappie is back from vacation, it seems like the LSC calculations are trying to account for crosswind, but perhaps doing so with the polarity reversed.  That's speculation on my part, of course.