Jump to content

Sign in to follow this  
robert young

Latest Asobo update has radically changed the a/p

Recommended Posts

13 hours ago, marsman2020 said:

Well, spent most of my afternoon on this, now to attach it to a ZenDesk ticket so Asobo can ignore it:

 

 

Good work on documenting this. They really should be pulling the trigger on the opt-in beta ASAP as there's obviously a lot of excellent talent to create very clear use cases for bug fixing... 

Share this post


Link to post
Share on other sites

Why does everyone keep harping about this?  The turn rate is slowed down to enhance sightseeing.... LOL!

Very effective presentation by the way.

 

Edited by fppilot

Frank Patton
MSI Z490 WiFi MOB;  i7 10700k Comet Lake 3.8 Ghz CPU; Ripjaws 32 gb DDR4 3600; ASUS GTX 1070 TI Turbo 8GB; MasterCase H500M; Corsair H100i Pro cooler; Corsair RMX850X PSU; ASUS VG289 4K 27"; Honeycomb Alpha Yoke A+.  Former USAF meteorologist & ground weather school instructor; AOPA Member #07379126 
                        There is no limit to what a man can do or where he can go if he doesn’t mind who gets the credit! - Benjamin Jowett

 

Share this post


Link to post
Share on other sites

Keep the discussion on topic please. Those off topic have been hidden.

  • Like 1
  • Upvote 1

Ray (Cheshire, England).
System: P3D v4.5, Intel i7-8086K o/c to 4.6Ghz, Nvidia GTX 1080 Ti 11Gb, Samsung 970 EVO M.2 SSD, 1Tb Samsung 860 EVO SSD, Asus Prime Z370-A mobo, 32Gb G.Skill DDR4 3000Mhz RAM, Win 10 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke.
Cheadle Hulme Weather

Share this post


Link to post
Share on other sites
9 minutes ago, Ray Proudfoot said:

Keep the discussion on topic please. Those off topic have been hidden.

Thanks a lot Ray--this thread is CHOCK FULL of off topic replies.


Noel

System:  9900K@4.9Ghz@1.21v all cores w/ HT enabled, MSI MPG Z390M GAMING EDGE AC, Noctua NH-D15S, Corsair Vengeance 32Gb LPX 3200mHz DDR4, Sabrent NVme 2Tb x 2, RTX 2070 Super FE, Corsair RM 850W PSU, Win10 Pro, Dell curved 3440x1440, Saitek Yoke, TQ & Cessna Trim Wheel, UNLIMITED frames Vsync to 30Hz.

 

Share this post


Link to post
Share on other sites

TTDR (Too Technical, Didn’t Read)

In large, formal software development, the coders are never in charge and usually only understand their part of the project.  Yes, they are the ones that actually make the game in the end, but there are others with essential roles such as Architects, Designers, Domain Experts (pilots, aircraft manufacturers, GPS / autopilot manufacturers, etc.), and QA.  We also have a saying, “Architects architect, Designers design, coders code, and QA tests the result.  QA tests against the design – they don’t design or develop!

As to the lateral nav bug, Robert and others are right, but their analysis and proposed solutions are incomplete:

1.      The AP death dive, the wing rocking, the lack of a stable, reasonable bank angle, and the slow acquisition of a new heading are all symptoms of the same problem.

2.      The solution cannot be achieved through PID (proportional error) control only.

3.      The solution cannot be achieved through a fixed bank angle and some one-size-fits-all roll-out algorithm.

4.      The solution requires a combination of both, and that can only be enabled by Asobo.

TECHNICAL PART

Some have stated that PID control is inappropriate for heading control.  They are both wrong and right.

They are right in that PID control is absolutely wrong choice for achieving a constant bank angle during the first 80% of the turn from the current heading to the desired heading.  As a domain expert (inactive private pilot) working with a Designer, I would have a requirement for a definable, constant bank angle during this initial part of the turn, with a definable transition point (degrees left) where the PID logic takes over.

WHAT IS PID CONTROL?

What is PID control, and why must we use it for the final part of the turn?

Simplistically, PID control is a feedback system that gradually – and proportionally -- reduces the rate of turn (bank angle) as the heading error (difference between actual heading and desired heading) approach 0.  In other words, it is EXACTLY what we need to roll out on the correct heading (the PID “set point”).  The greater the difference (or “error”), the greater the bank angle. 

Tuned correctly (voodoo) in the AI.CFG file (by us, modders, or Asobo), PID control prevents overshooting the heading.  And it eliminates wing rocking.  And it accounts for varying speeds, wind gusts, and other adverse factors.

SIMPLE HEADING VERSUS TRACKING A VOR RADIAL OR GPS / FMS COURSE

One of the reasons I say that the arguments have been incomplete is in the case of intercepting a desired course and tracking it laterally.  This is more complicated than rolling out on a manually set (bugged) heading because the autopilot must also account for:

1.      Correcting for wind blowing you off course – causes an adjustment in heading.

2.      Speed of the aircraft and max bank angle allowed – used to calculate when to make the turn and how much initial bank is required.

3.      Switching from coarse tracking to fine (LOC) tracking – makes the course tracking more sensitive and commands more frequent and finer heading changes.

This is why the AP screens show you a desired track heading and an actual track heading.  Depending on wind, the actual heading is going to be off a few degrees from the desired track heading, but the aircraft path will follow the desired track.

PROBLEM DESCRIPTION (AS OF 12/03/20 PATCH 7)

Asobo apparently – at least in 1.11 (patch 7) – decided to change the heading code to use ONLY PID.  Doing so resulted in a large bank angle only when the heading difference was high, and gradually (proportionally) reducing the bank angle as the difference got smaller.  The result is that we can’t get a constant bank angle at first, and the bank angle gets too small too fast at the end, causing the very long roll-outs.  If we tune the PID parameters to achieve a bigger bank angle and faster roll-out, we get the old wing-rocking.  If we leave PID as it was before the patch, then in the absence of the initial, constant bank angle (that they removed), we get the large bank angle, spiral of death that was reported.

SUMMARY

The final solution must address three things:

1.      Provide an initial, constant rate turn angle phase – not PID controlled.

2.      Roll-out to the commanded (manually set or AP calculated) heading using PID and limited by a definable max bank angle.

3.      Maintain that commanded heading using PID to react to wind gusts and other inputs, while eliminating unnecessary wing rocking due to overreacting to very small heading changes (less than 0.5 degrees).

While modders can “work around” the current PID-only setup, the workarounds are unsatisfactory for the long term.  Only Asobo can give us the initial constant bank / final PID control that we – and plane developers – must have in the end.

QA, alpha and bet testers, third party developers, and the customer base can – and I think have – correctly identified these adverse behaviors we’re seeing.  I think that Asobo knows all about it, but business decisions are currently driving priorities.

WHAT TO EXPECT

I feel very positive about this.  Notice my lack of emotion in my analysis.  I think Asobo has been learning what they don’t know and is looking to hire people with more knowledge in those areas (like PID control).  It’s just a matter of time.

For now, can you fly low and slow VFR without the need to turn sharply to an ILS course or localizer?  Yes.  Are they working on changes to enable realistic IFR?  Yes.  For now, rather than getting frustrated and quitting, fly VFR.

In my experience, the worst thing is to have a developer that never changes anything.  At this point in the life cycle of this sim, I would rather Asobo flail about and make mistakes than do nothing.  And even though they break IFR stuff (and the heading bug) with every release, I don’t see any big showstoppers with the IFR stuff on the software side.  Again, it’s just a matter of time.

  • Like 3
  • Upvote 3

Share this post


Link to post
Share on other sites

After the FS2020 mini patch released this morning (Dec 3, 2020) I decided to take the Bonanza G36. I climbed to 5,000 ft and did a 180 turn and it took 65 seconds, this was much faster than without the update. I continued on to CYYZ and did a 45 deg ILS intercept of runway 05 without any problems. Robert, do you think the problems with the autopilot have been corrected with the latest FS patch?

John

Share this post


Link to post
Share on other sites
1 hour ago, bdwhittle64 said:

Asobo apparently – at least in 1.11 (patch 7) – decided to change the heading code to use ONLY PID. 

I see no sign of this in the C172 or Bonanza... the initial turn is at a fixed bank angle and only the last 30 degrees appear to be PID controlled.  After the latest patch still a bit slow to finish the turn, but acceptable..


Bert

Share this post


Link to post
Share on other sites
2 hours ago, bdwhittle64 said:

TTDR (Too Technical, Didn’t Read)

In large, formal software development,

 

2 hours ago, bdwhittle64 said:

And even though they break IFR stuff (and the heading bug) with every release, I don’t see any big showstoppers with the IFR stuff on the software side.  Again, it’s just a matter of time.

Thank you for a well reasoned and easy to understand technical response that outlines the problem specifically and is void of some of the histrionics and mud slinging that others have been ascribing to this behavior.

 

Share this post


Link to post
Share on other sites
2 hours ago, bdwhittle64 said:

And even though they break IFR stuff (and the heading bug) with every release, I don’t see any big showstoppers with the IFR stuff on the software side.  Again, it’s just a matter of time.

There are numerous issues with IFR beyond just the autopilots..... 

  • The 'Crazy Ivan' turn when activating an approach / Garmins adding USR/USER points in ways that are not useful/realistic to what an actual Garmin does
  • Numerous approaches missing from the Navblue data
  • All localizers are locked to the runway headings.  Any LDA approach (offset angles >3-5 deg, FAA vs ICAO have different cutoffs) or any regular LOC or ILS approach with an offset is wrong.  Even some of the Asobo "hand crafted" airports have this problem.  They have overwritten the localizer orientation values in the Navblue data in the .bgl files with the runway headings, so there isn't even good nav data to work off of in the database unless we pay for Navigraph
  • Inability to select and load an RNAV approach for an airport in-flight.  It has to be a 'destination' in the world map, or it doesn't work
  • LPV and LNAV/VNAV approaches that don't display vertical guidance properly on instruments (the AP will follow the correct glide path but the 'pip' stays pegged high or low). 
  • The ATC doesn't give vectors to final for approaches that require them to intercept a localizer anymore, it did in FSX
  • The 'distance per pip' doesn't scale the way it should in some instances as you get closer to the threshold

AMD 3950X | 64GB RAM | AMD 5700XT | CH Fighterstick / Pro Throttle / Pro Pedals

Share this post


Link to post
Share on other sites
1 hour ago, Bert Pieke said:

I see no sign of this in the C172 or Bonanza... the initial turn is at a fixed bank angle and only the last 30 degrees appear to be PID controlled.  After the latest patch still a bit slow to finish the turn, but acceptable..

Interesting.  I've only personally tried the Caravan and TBM.  I've been a long time "lurker" on the Avsim forums, but until MSFS came out, didn't feel compelled to post.  Prior to my long post, I did read all the threads here and on the official forum pertaining to this lateral nav problem, and I downloaded the Working Title mods and analyzed the PID changes they made for the A320 and the CRJ.  I experimented myself with my own parameters on the Caravan and TBM.

Since Robert -- the author of a Bonanza mod, I believe -- passionately decried the broken lateral nav in AP mode, I assumed that the flaw affects all aircraft APs and is thus a flaw in the general code.  When I get time, I'll try the  vanilla 172 myself.  I've learned to never say never, and in software, as soon as you're 100% about something, a software update inevitably makes you wrong.

Good debugging should eliminate variables, not introduce them.  Our problem is that Asobo can slip in a patch anytime we start the simulator, and has no obligation to tell us.  Believe me, that's a good thing.  I've made some embarrassing mistakes in my long career and couldn't just quietly fix it.  We had to publicly issue an explanation and distribute a hotfix, or best case issue a quick "feature" update and hide the hotfix under the "numerous other issues addressed".

MSFS's way is better!

Share this post


Link to post
Share on other sites

Asobo only mentioned a navdata update, but I believe they snuck in some fixes at the same time..


Bert

Share this post


Link to post
Share on other sites
20 minutes ago, Bert Pieke said:

Asobo only mentioned a navdata update, but I believe they snuck in some fixes at the same time..

Those cheeky b*stards! 😄

  • Like 1

Share this post


Link to post
Share on other sites
22 minutes ago, Bert Pieke said:

Asobo only mentioned a navdata update, but I believe they snuck in some fixes at the same time..

I hope not.  That would be really annoying.  If they are changing stuff they should tell us.

I'll record a new video tonight.  Probably won't spend so long on the "production value".

Edited by marsman2020

AMD 3950X | 64GB RAM | AMD 5700XT | CH Fighterstick / Pro Throttle / Pro Pedals

Share this post


Link to post
Share on other sites
33 minutes ago, marsman2020 said:

There are numerous issues with IFR beyond just the autopilots..... 

What you list are absolutely correct and will need to eventually fixed before we have full-fledged, realistic, simulation of modern-day autopilots.  Many users on this thread have stated that they believe that lateral navigation is the most important and most urgent AP problem.  I completely agree and from a pilot's perspective, here's why.

Back when I got my private pilot license (1975 at Ft Lauderdale, FL -- FLL), ALL private pilots had to learn to navigate using radio navigation aids VORs and ADFs.  We didn't have GPSs at the time, and most GPS aircraft even today still have backup radio nav equipment.  Go to the FAA site and look at the flight training materials, and you'll see that you still have to learn that stuff today.

I learned -- like many of today's pilots -- on a Cessna 152.  Like in the simulator, it doesn't have an autopilot, not even a wing leveler.  If you wanted to take your hands off the controls, and / or stop fighting the yoke every time you changed pitch or power, you had better get good at trimming the aircraft.  You flew with one hand on the yoke, and the other on the throttle or pitch trim wheel.  You FREQUENTLY trimmed the aircraft!

Before you can get your license, you had to do a solo cross country, and also a long, three legged cross country using radio navigation and pilotage (like the 152 lessons in MSFS).  At a decent size flight school, it was usually only after you get your license that you moved up to a plane like a 172 that was IFR rated -- meaning it had an autopilot that could hold a heading or altitude, or track a VOR radial.

My point?  We didn't have GPS, FMS, or any other systems that would follow a flight plan or show you where you were.  It  was all done by flying from one VOR to another, or to an ADF, or hold a heading to an intersection of two VOR radials.  It we (private, commercial, and airline pilots) could fly IFR with just that basic equipment, so can the serious simmers with MSFS.

Of course, it would be nice if they don't break the heading mode on the AP.  For now, I'm manually flying from one heading to the next, then synching the heading bug when I'm on course and turning the AP back on.  I manually adjust the heading bug for small wind corrections.  It's real flying, but that's why a lot of us are here!  You ever wonder why there was so much hue and cry over the steam gauge 172?  That's why.

So while you're absolutely right that the AP needs a whole lot of work, for many of us, as long as we have workable heading and altitude hold logic, and two nav radios and DME, we can competently fly IFR without ever looking at a GPS.  And it's a whole lot of fun, too! 

It's just so cool to take off and 500 feet later you have zero visibility until you break out at 300 feet at the destination airport.  And the only thing you had to get you there were two nav radios, the artificial horizon, the DG, and your wits and skill as a pilot!

  • Like 3
  • Upvote 1

Share this post


Link to post
Share on other sites
28 minutes ago, Bert Pieke said:

Asobo only mentioned a navdata update, but I believe they snuck in some fixes at the same time..

It's very likely, and it's a very good thing if they do.  All software companies secretly try to fix significant mistakes before they are noticed, or like I said before, hide a large amount of bug fixes in any new release.  They usually find some very positive bug fixes to brag about, and hide the rest.

Keep in mind that this is a very savvy parent company taking a flyer on a very creative division to produce what?  A video game!  They have done that in the past, and many people are puzzled by that.  But remember that old saying, "Follow the money!"

They are all in this for the money in the end, although I see a lot  of passion for the graphics and flying -- from Asobo.  From the Microsoft motive, why a whole world flight simulator?  Bing, Bing Maps, and Microsoft Cloud (Azure) services, anyone?

Google eats Bing's lunch.  Google Maps eats Bing Map's lunch.  Amazon AWS is a dominant competitor to Azure cloud services.  MSFS is right now a killer VFR flight simulator and very cool XBox flight game.  With a little more time and SDK and basic AP development, third parties will make it a killer IFR simulator, including for the big iron.

If I'm Microsoft, I'm pleased that I already funded the next few years of development on the sim.  I'm also excited that it is funding a lot of Bing maps, AI research, and demonstration of Azure through an impressive, hosted flight simulator as a service.  But I'm absolutely STOKED at the amount of people spending a large amount of time talking about, making videos about, flying around with, posting pictures of, MICROSOFT FLIGHT SIMULATOR, USING WORLDWIDE MICROSOFT BING MAPS FOR SCENERY, HOSTED ON MICROSOFT AZURE AS A CLOUD SERVICE, AND BEING RELEASED ON MICROSOFT XBOX (ALTHOUGH YOU WILL HAVE TO BUY A NEW ONE).

That is all very, very good publicity and business for Microsoft.  Companies pay millions of dollars a year for that kind of advertisement.  The old saying is, "Even bad publicity is still publicity."  So even all these horrible posts in the forums shows Microsoft how much interest there is in the sim, and how passionate that interest is.  It's just great business moves on Microsoft's part.

I want you to know that I believe in aliens, but I'm not a conspiracy theorist.  I do find it highly coincidental, however, that Microsoft releases this game right in the middle of the covid pandemic -- right when everyone's stuck at home with lots of time on their hands.  And lots of forums to fill up with all manner of messages, opinions, and emotion!

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

  • Tom Allensworth,
    Founder of AVSIM Online


  • Flight Simulation's Premier Resource!

    AVSIM is a free service to the flight simulation community. AVSIM is staffed completely by volunteers and all funds donated to AVSIM go directly back to supporting the community. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. Thank you for your support!

    Click here for more information and to see all donations year to date.
  • Donation Goals

    AVSIM's 2020 Fundraising Goal

    Donate to our annual general fundraising goal. This donation keeps our doors open and providing you service 24 x 7 x 365. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. We reset this goal every new year for the following year's goal.


    40%
    $10,150.00 of $25,000.00 Donate Now
×
×
  • Create New...