As to the comments on the display gap I'm sure we are all just about over it. I wonder if the desire to see an 'aesthetic' ie is perfect alignment on both sides,  is because PMDG delivered us an 'as new' aircraft and as such, we might expect the displays to be aligned. But, when the scans have been taken possibly from a well worked line aircraft then in reality they probably won't be aligned. In this case both arguments might be correct but need to be qualified with 'in a new or near new aircraft' vs 'in an older line aircraft'. So perhaps this statement might be better FWIW: 

"In a new aircraft which PMDG have delivered us, the displays should be aesthetically aligned as if the aircraft had just been delivered new from the factory but instead they show misalignment as if the aircraft has already had display rotations as might be expected in an older aircraft."

Just be sure of one thing. If you now have the aircraft in your ownership, it is no longer new and you have dust on the CDU's and MFD. Additionally, it seems like someone fumbled their coffee in some unexpected CAT and a display had to be swapped out for a precautionary servicing.

You can't have a new aircraft forever!! 

Since some guys who actually fly these things either haven't noticed or if they have couldn't attest to all their fleet being like that, it says that this query as to accuracy in the PMDG model it so far down the scale of things as to be entirely ignored rather than warrant so many pages of analysis.  

 

Edit:  Special note to @Bobsk8 and @JRBarrett who are both avionics technicians and who have both stated that misalignment in service does happen. And, if we learned anything from this otherwise fairly useless discussion it was from JR Barrett's description of exactly how these units are located and  how difficult they are to swap out for servicing.  👍

Cheers

Terry

 

Edited July 15, 20241 yr by Lord Farringdon

I am really blown away that there is a discussion about some mm gap between displays. Of all the things to talk about. 🙂

4 hours ago, LRBS said:

First, they introduced these RF legs, but they are still not entirely correct. In normal wx conditions, the airplanes are rock solid and have constant bank angles during turns. Not in PMDG, where there is a constant back and force left/right banking.  Before or after wpt the airplane will turn momentarily L/R of course (as a pilot, if I notice this kind of behavior, I will instantly disconnect the A/P, we cannot allow the airplane to wonder in turns like that), turns should be precise and smooth not for the pilot to wonder what is she gonna do next

I’ve spent about 50 hours now flying the 777 in a variety of weather conditions, and I have done most of it closely zoomed in on the PFD looking at its performance in turns under autopilot LNAV control, and I simply do not see many of the things you mentioned. I have never seen the aircraft start to roll into a turn in the opposite direction to the actual turn direction, and I have specifically looked for this. I definitely know what you mean, because I have seen it in the PMDG 737, but not in the 777.

As far as the bank angle varying in a turn specifically when following an RF leg at a constant airspeed, the bank angle has to vary through the turn (at least a little bit) if there is any amount of wind at any point during the turn.

No aircraft, large or small, can follow a fixed, constant radius path over the ground without varying the bank angle and rate of turn as the wind direction changes. It’s something every PPL student learns and practices during ground reference maneuvers such as turns about a point or S turns across a road. A heavy 777 traveling at 250 knots won’t be affected the same way as a C152 traveling at 70 knots, but the same laws of aerodynamics apply.

To the degree that the 777 bank angle varies on an RF leg, it seems to do it correctly - slowly shallowing the bank when turning upwind and steepening the bank when turning downwind. I have seen absolutely no “constant” left or right corrections at inappropriate times as I do on the 737. It may not be “perfect” but for a $77 add-on it’s darn good, and much better than the 737.

The only constant LNAV  “glitch” I see is what happens when the aircraft levels off at the very end of a turn. When the bank angle gets down to about 2 degrees “to go” it suddenly and abruptly “snaps” level. I’ve seen this on every iteration of the 777 since release.

3 hours ago, JRBarrett said:

I’ve spent about 50 hours now flying the 777 in a variety of weather conditions, and I have done most of it closely zoomed in on the PFD looking at its performance in turns under autopilot LNAV control, and I simply do not see many of the things you mentioned. I have never seen the aircraft start to roll into a turn in the opposite direction to the actual turn direction, and I have specifically looked for this. I definitely know what you mean, because I have seen it in the PMDG 737, but not in the 777.

As far as the bank angle varying in a turn specifically when following an RF leg at a constant airspeed, the bank angle has to vary through the turn (at least a little bit) if there is any amount of wind at any point during the turn.

No aircraft, large or small, can follow a fixed, constant radius path over the ground without varying the bank angle and rate of turn as the wind direction changes. It’s something every PPL student learns and practices during ground reference maneuvers such as turns about a point or S turns across a road. A heavy 777 traveling at 250 knots won’t be affected the same way as a C152 traveling at 70 knots, but the same laws of aerodynamics apply.

To the degree that the 777 bank angle varies on an RF leg, it seems to do it correctly - slowly shallowing the bank when turning upwind and steepening the bank when turning downwind. I have seen absolutely no “constant” left or right corrections at inappropriate times as I do on the 737. It may not be “perfect” but for a $77 add-on it’s darn good, and much better than the 737.

The only constant LNAV  “glitch” I see is what happens when the aircraft levels off at the very end of a turn. When the bank angle gets down to about 2 degrees “to go” it suddenly and abruptly “snaps” level. I’ve seen this on every iteration of the 777 since release.

I do appreciate the lesson about how "every PPL student learns and practices during ground reference maneuvers such as turns about a point or S turns across a road. A heavy 777 traveling at 250 knots won’t be affected the same way as a C152 traveling at 70 knots, but the same laws of aerodynamics apply."

 

I also wish you had noticed when I mentioned" no wind conditions".

 

I would also like to share with you that even in a 737 with wind variations up to 10 KTS,  bank angles are constant and smooth in these kinds of winds, more so in a 777.  Above those approximate values in turbulent conditions, you will see some variations.

 

Something else worth mentioning is that those RF approaches start at FAP with the airplane fully configured at Vref between 160 to 130 KTS based on the flaps setting and weight. Usually, the bank angle is steady, with no variations, somewhere between 10 to 20 degrees, steady and dead on the magenta line with no errors of any kind.

 

It is really interesting to see these arguments from people (no disrespect intended) who do not fly these airplanes for a living.

 

It's like pilots giving instructions to avionics engineers on how to do their job.  

 

Anyhow, I'll be home in a few days and have access to my desktop. If you are interested, I can find some examples for you and perhaps discover stuff that you didn't notice.

 

I agree with everybody that it's outstanding software for $77, but the discussion is not about that, and PMDG could do a better job.

Edited July 15, 20241 yr by LRBS

32 minutes ago, LRBS said:

 

I would also like to share with you that even in a 737 with wind variations up to 10 KTS,  bank angles are constant and smooth in these kinds of winds, more so in a 777.  Above those approximate values in turbulent conditions, you will see some variations.

 

Sorry but I gotta ask you! You suggesting that if 737 (hypothetically) was doing ground reference maneuvers with variable wind speeds up to 10kts  applying constant angle of bank there was no drift? Or you means something else ?

2 hours ago, LRBS said:

I do appreciate the lesson about how "every PPL student learns and practices during ground reference maneuvers such as turns about a point or S turns across a road. A heavy 777 traveling at 250 knots won’t be affected the same way as a C152 traveling at 70 knots, but the same laws of aerodynamics apply."

 

I also wish you had noticed when I mentioned" no wind conditions".

 

I would also like to share with you that even in a 737 with wind variations up to 10 KTS,  bank angles are constant and smooth in these kinds of winds, more so in a 777.  Above those approximate values in turbulent conditions, you will see some variations.

 

Something else worth mentioning is that those RF approaches start at FAP with the airplane fully configured at Vref between 160 to 130 KTS based on the flaps setting and weight. Usually, the bank angle is steady, with no variations, somewhere between 10 to 20 degrees, steady and dead on the magenta line with no errors of any kind.

 

It is really interesting to see these arguments from people (no disrespect intended) who do not fly these airplanes for a living.

 

It's like pilots giving instructions to avionics engineers on how to do their job.  

 

Anyhow, I'll be home in a few days and have access to my desktop. If you are interested, I can find some examples for you and perhaps discover stuff that you didn't notice.

 

I agree with everybody that it's outstanding software for $77, but the discussion is not about that, and PMDG could do a better job.

I agree on this and share this opinion. 

I too see strange behavior. The 777 does not follow the magenta line properly and regular does leave the line left or right sight and struggle to come back with a unusual sharp bank.....

Then a left and right  (or VV) correction begins. 

This was not the case with PMDG at other simulator at least (if this is something worth to mention) .

I think, the 777 struggles with calculations (codewise) and "frequently updated positioning is not precise enough and therefore these "leaving the magenta line" happens. 

 

The speed does not matter. Even at 180 knots or 160 knots this happens. 

Also on stars

Marcus

 

6 hours ago, JRBarrett said:

I’ve spent about 50 hours now flying the 777 in a variety of weather conditions, and I have done most of it closely zoomed in one PFD looking at its performance in turns under autopilot LNAV control, and I simply do not see many of the things you mentioned. I have never seen the aircraft start to roll into a turn in the opposite direction to the actual turn direction, and I have specifically looked for this. I definitely know what you mean, because I have seen it in the PMDG 737, but not in the 777.

As far as the bank angle varying in a turn specifically when following an RF leg at a constant airspeed, the bank angle has to vary through the turn (at least a little bit) if there is any amount of wind at any point during the turn.

No aircraft, large or small, can follow a fixed, constant radius path over the ground without varying the bank angle and rate of turn as the wind direction changes. It’s something every PPL student learns and practices during ground reference maneuvers such as turns about a point or S turns across a road. A heavy 777 traveling at 250 knots won’t be affected the same way as a C152 traveling at 70 knots, but the same laws of aerodynamics apply.

To the degree that the 777 bank angle varies on an RF leg, it seems to do it correctly - slowly shallowing the bank when turning upwind and steepening the bank when turning downwind. I have seen absolutely no “constant” left or right corrections at inappropriate times as I do on the 737. It may not be “perfect” but for a $77 add-on it’s darn good, and much better than the 737.

The only constant LNAV  “glitch” I see is what happens when the aircraft levels off at the very end of a turn. When the bank angle gets down to about 2 degrees “to go” it suddenly and abruptly “snaps” level. I’ve seen this on every iteration of the 777 since release.

I tested it on 2 Kathmandu RNP AR approaches and Madeira RNP AR 05 Z and it performed without glitches in regards to all parameters, including speed and altitude constraints, but there are probably still some glitches...

OTOH this 777 can't get free from the freak effect that turbulence and up / downdrafts have in all aircraft in MSFS 2020, but that's inherent to the sim and we can only wish it gets better modeled in 2024...

Edited July 15, 20241 yr by jcomm

Morning Gang!

With my game developer hat on (my day job), I wanted to quickly comment on the pathfinding algorithm.

There is always going to be a trade-off on a home simulation as to put it simply, perfect pathfinding is going to eat all your processor time. The more processor budget that is spent on this, the less time for anything else getting done. Perfection costs! So you have to balance CPU cycles across all the competing needs. The pathfinding at PMDG has come an awful long way over the years, both with the rise in computing power freeing up more cycles and Vangelis and the team really honing their craft (I started my relationship with PMDG in 2003 on the beta for their first 737 and beyond). We have come a long long way. There is still  more to be done but much of that will need greater computer power as we iterate through new generations of PC's. I have spent a large part of my life creating pathfinding algorithms for commercial products and I look back at my previous work on this (such as Alien Vs Predator - 30 years ago this month) and cringe, but its all the processor I had to play with.

Bottom line is that we always rob Peter to pay Paul when it comes to resources. Processor has to be balanced across so many competing needs in a sim. The guys have done amazingly well as have other developers with the level of fidelity we see today.

I have seen the "jiggling" over the waypoints, which is much less pronounced on the 777 than the 737. I am speculating, but this may be linked to waypoint precision. Code tries very hard to hit a waypoint exactly. Quite often (and we see this a lot in videogame pathfinding, be it Call of Duty or a racing game etc), the position slightly overshoots. Lots of reasons for this, which would eat up a whole post on the technicalities... The code then tries to bring the position back. So you get the little jiggling with the overcompensation. One trick to help alleviate this is to add a little damping as you approach close to a waypoint to avoid oscillation and over-correction trying to hit the waypoint on the nail. I am sure the guys will continue to tweak it. It is a very common issue in pathfinding algorithms when you are aiming to hit a waypoint in 3D space very exactly. Again, the smarter you get with this, the more processor it will cost.

Bottom line, until we are all running supercomputers (and I used to make a living writing games for computers with 1k of memory!) we are going to have to live with little trade-offs to manage performance.

Have a great day everyone,

Jane

9 hours ago, Lord Farringdon said:

"In a new aircraft which PMDG have delivered us, the displays should be aesthetically aligned as if the aircraft had just been delivered new from the factory but instead they show misalignment as if the aircraft has already had display rotations as might be expected in an older aircraft."

The 777 PMDG used as a survey aircraft is an older aircraft that has been in service many years, according to them. So it makes sense it has slightly misaligned displays. The 737 PMDG used meanwhile came straight out the factory and was brand new. This is also why in the 777 cockpit there is visible wear that has been modelled whereas the 737 has almost completely clean textures.

1 hour ago, JaneRachel said:

There is always going to be a trade-off on a home simulation as to put it simply, perfect pathfinding is going to eat all your processor time.

Is this comparable to path finding? Each leg and the subsequent legs to transition to contain, a good amount of information on how to sequence between them (like whether to make a left turn, how big of an arc to fly etc. etc. path finding needs to figure this all out). Also the FMS uses a specific set of math formulas to draw the magenta line (as-well as fly it) based on this information and data from the ADIRS. This is all then done on an avionics suite with CPUs less powerful than consumer hardware due to hardening and longevity requirements.

Now I don't work with PMDG, nor do I know it's internals. Though the code should be replicating or if possible (depending on data availability) just reusing these math formulas to replicate the same way the real aircraft would. If that were the case, I'm not see how a more powerful PC here matters. You're not trying to be more accurate with the computations, you need to compute it the same way the real aircraft would.

I could be way off here as this is just my perspective on it but with you're relationship with PMDG may be useful to give us an insight on areas I may be wrong.

23 minutes ago, Lucky38i said:

Is this comparable to path finding? Each leg and the subsequent legs to transition to contain, a good amount of information on how to sequence between them (like whether to make a left turn, how big of an arc to fly etc. etc. path finding needs to figure this all out). Also the FMS uses a specific set of math formulas to draw the magenta line (as-well as fly it) based on this information and data from the ADIRS. This is all then done on an avionics suite with CPUs less powerful than consumer hardware due to hardening and longevity requirements.

Now I don't work with PMDG, nor do I know it's internals. Though the code should be replicating or if possible (depending on data availability) just reusing these math formulas to replicate the same way the real aircraft would. If that were the case, I'm not see how a more powerful PC here matters. You're not trying to be more accurate with the computations, you need to compute it the same way the real aircraft would.

I could be way off here as this is just my perspective on it but with you're relationship with PMDG may be useful to give us an insight on areas I may be wrong.

Morning,

To be clear, I don't have or ever had access to PMDG code in any way, so I cannot speak for their internal code.

You are, of course, quite right about the CPU in avionics kit being less powerful than consumer hardware, but the difference is that those CPU's are dedicated to avionics, whereas our CPU has to render the aircraft, the world and do everything else also. I also agree on replicating maths formula. That formula though can get expensive and I would argue that a level of accuracy is required, in interpretation of ADIRS data etc.

I defer to your knowledge though in the sim (as I have never written this code for an aircraft, only in "mainstream" games as a games AI and pathing developer).

11 hours ago, Stearmandriver said:

 

I fully agree with LRBS that the LNAV still needs work.  Is it improved from the 73?  Absolutely.  Is it realistic yet?  Nope.  Plane should fly these procedures, to use a popular simmer slur, like it is on rails.  Literally.

I’m glad other Boeing pilots have noticed this!

Forums/discord/Reddit etc all seem full of people saying how great it is now.

The LNAV definitely definitely isn’t correct although it’s far better than the 737s 

I don’t think the FLCH works correctly either especially with the A/T but I got bored on the forum etc and gave up on giving feedback. 

How the hell are people expecting PMDG to be fully replicating a 777 1:1? I am not sure how many million manhours Boeing have spent on developing this aircraft, but it must be many. 

Maybe PMDG can spend 1000 manhours on improving the pathfinding a few percent, but I am not sure if it can be defended financially.

 

I spent two days touring the CAE Factory in Montreal Canada, years ago, when they were making the 767 Level D Sim that at the time cost around $25 million dollars. I remember them detailing how they got the flight info from Boeing Test Flights of the same aircraft, which cost them a couple of million dollars. Boeing even provided data from a few tested aircraft, so they could pick the average of the tested ones. 

Now, here some  are on Avsim, with a piece of $78 software on a PC that maybe is worth a bit over $1,000, expecting this software to fly exactly like the real aircraft. Talk about crazy......🤨

56 minutes ago, g-liner said:

I don’t think the FLCH works correctly either especially with the A/T but I got bored on the forum etc and gave up on giving feedback. 

What's incorrect about it?