Jump to content

Archived

This topic is now archived and is closed to further replies.

matta757

Increased sim rate = erratic behavior?

Recommended Posts

Hey all,I was just wondering if maybe there is a remedy to this situation. When I do flights, especially long flights, I tend to use an increased rate of simulation. Anywhere from 2x-8x is usually what I like to do. Using 2x always works without any adverse effects. Using 4x usually works, but every once and a while the aircraft will randomly and erratically spin out of control until I return to normal sim, where I have to re-establish my routing. When I use 8x, it works for some aircraft but not for others. Examples; my PSS 777's work just fine at 8x (thankfully since transpacific flights take forever) but my PSS 757's do not. When I do 8x on the 757, it starts out fine, but then the altitude starts to fluctuate. It's mild at first; one hundred feet up and a hundred feet down from set cruise altitude. But it gets worse and soon the fluctuations can be a thousand feet above and below, at which point I clearly need to reduce the simulation rate.Does anyone know what is causing these fluctuations? I have adjusted some weather options using FSUIPC so that winds don't gust aloft and the rate of change isn't dramatic, thinking that maybe the winds were causing the issue, but it persists. Any help or insights would be greatly appreciated!Regards,Matt

Share this post


Link to post
Share on other sites

I am not an "expert" - but it stands to reason that when you increase the sim rate, you are asking the CPU and Graphics card to process data substantially faster than "normal". That can lead to all sorts of issues.We are talking about hundreds of thousands of mathmatical calculations per second just to create what you see on your screen. Tis pure "magic" to me.

Share this post


Link to post
Share on other sites
... starts out fine, but then the altitude starts to fluctuate. It's mild at first; one hundred feet up and a hundred feet down from set cruise altitude. But it gets worse and soon the fluctuations can be a thousand feet above and below, ... Does anyone know what is causing these fluctuations?
matt:simple answer -in FS 'normal' time is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 (etc)these 'points' are data and the FMS/FMC/MCDU is designed to react to the aircraft movement through these points in 'normal time'.you 'speed up time' and data is now 1, 4, 8, 12, 16 (4x), or worse 1, 16 (16x). the aircraft JUMPS THROUGH SPACE at a rate the MCDU/FMC/FMS wasn't designed to control.--

D. Scobie, feelThere support forum moderator: https://forum.simflight.com/forum/169-feelthere-support-forums/

Share this post


Link to post
Share on other sites

Thanks guys. That all makes perfect sense but I was still just a little confused why it works for some planes but not for others, so that's why I asked. Basically you are saying there is no cure (besides flying closer to the normal rate) for this issue?Regards,Matt

Share this post


Link to post
Share on other sites

Well, that certainly doesn't make sense to me. Not being privee to the actual code used in FS, I cant be sure of this, but in general this type of programming uses multiple calculations between each 'screen' generation any way. Increasing the speed merely makes the sim take bigger jumps ( as stated), but thats no reason why a plane should spin out of control. After all, it is only a sim - the planes don't actually fly. If taking bigger steps caused the phenomenon then so would a slow FPS. I've never seen this type of behaviour anyway, either in FS or any games I've written.Dave

Share this post


Link to post
Share on other sites

more detail, and assuming you have read my prior post -in each 'time point' the dynamic world of FS has mutliple datapoints of wind, pressure, temp, etc.the FMS/FMC/MCDU/etc gauge work 'outside' FS, meaning they are reading FS's data but not part of the calculations.the MCDU/FMS/FMC/etc makes adjustments based on the vector of the plane as it is moved through each FS datapoint.'speeding up' messing with these assumptions so the FMS/MCDU/FMC/etc begins to under/over react to the plane's movement away from the programmed LNAV and VNAV route. becuase of the MCDU/FMS/FMC/etc expects weather to move at a 1, 2, 3, 4 timeframe it is 'behind the curve' of FS's 1, 16, 32 jumps through 'accelerated space'.--


D. Scobie, feelThere support forum moderator: https://forum.simflight.com/forum/169-feelthere-support-forums/

Share this post


Link to post
Share on other sites
Hey all,I was just wondering if maybe there is a remedy to this situation. When I do flights, especially long flights, I tend to use an increased rate of simulation. Anywhere from 2x-8x is usually what I like to do. Using 2x always works without any adverse effects. Using 4x usually works, but every once and a while the aircraft will randomly and erratically spin out of control until I return to normal sim, where I have to re-establish my routing. When I use 8x, it works for some aircraft but not for others. Examples; my PSS 777's work just fine at 8x (thankfully since transpacific flights take forever) but my PSS 757's do not. When I do 8x on the 757, it starts out fine, but then the altitude starts to fluctuate. It's mild at first; one hundred feet up and a hundred feet down from set cruise altitude. But it gets worse and soon the fluctuations can be a thousand feet above and below, at which point I clearly need to reduce the simulation rate.Does anyone know what is causing these fluctuations? I have adjusted some weather options using FSUIPC so that winds don't gust aloft and the rate of change isn't dramatic, thinking that maybe the winds were causing the issue, but it persists. Any help or insights would be greatly appreciated!Regards,Matt
Matt there's a simple answer to this that would make more since to you. This was discussed some time ago when using Flight Deck Companion and finding the autopilot callouts weren't functioning with the aircraft. Simple answer:Some aircraft add-ons us more of the default autopilot's attributes/code more than others (plain and simple)....Expanding on this I can tell you that all PMDG birds have specially designed autopilots that drift away from FS's SDK specifications to more accurately model the real autopilot. There's no PMDG bird that'll fly at 16X sim speed outside of their lite Beech 1900 series. So for long haul flights you may want to try CLS's 742 or SkySimulations MD11 if your short on time. I rarely fly the PMDG 744 for this very reason. I was able to get the 737NG to run at 8X once I ran defrag and had smooth file access on a flight with not that many turns (KMDW to KBUF).All Feelthere birds with the exception of their 737-300 series can be ran with maximum sim acceleration. Even their Airbus thankfully retained much of the default autopilot specifications to allow it to run with max acceleration. The only thing that'll bite you with the Airbus is wing icing that'll almost knock the Airbus (all Vol 1-2 variants) out the sky at high acceleration rates. You have to turn wing icing off and monitor the 'Direct To' option (located on the MCDU) when the route requires a course change to insure your not in constant wide turns when the Bus is trying to find the course. I usually set my waypoints far apart for long haul flights so excessive heading hunting isn't a problem.Next up is the Outstanding Leonardo MD80 that is very happy with sim acceleration. For a plane this complex it's refreshing this bird can be flown long distances when a simmer has little time (these guys thought of everything).PSS's 777 will bounce up and down at 8X acceleration but it's not a big deal. At least the bird is able to run at a higher sim rate to allow a flight from KJFK to SBGL in under an hour. Don't try anything above 8x or you'll have problems, also defrag your HD before a flight...Captainsim's 757 is not worth talking about. Sim acceleration won't work with this bird at all...I'm currently looking at what the Level-D 767 can do which opens the door to lite projects like CLS's new 767. For those that don't have time for an 8-10 hour flight an autopilot built closely to the default specifications allows one to accelerate sim time once at cruise altitude to cut the time down to a fraction of what it would take in the real world... For me I like departures and approaches. All that idle time in between leaves me feeling like an NWA pilot needing my laptop to break up the boredom. :( You have to play around with your various add-ons to see what different aircraft in your stable can do. You'd be surprised as to how many add-ons will accommodate at least 8x speed. Developers trying to achieve as much realism in FS as possible have actually done a few things wrong from a practical use standpoint and it's apparent once the add-ons are on simmers hard drives (that's why these forums are so important as to fleshing out what works and what doesn't). A big black mark I've seen that stands out is the various ways autopilots have been built. Autopilots are pretty straight forward on the front end and borrowing as much from the default code as possible is actually good in the grand scheme of things and by all means the best way to go. As long as the AP follows the FMC course that's all we would want (I know there's purist looking for RNAV/VNAV preciseness but that's not the majority. Some of these guys need to start learning how to hand land the aircraft anyway versus expecting the ILS to give them a perfect hand's off touch down every time. Most if not all airlines require pilots to manually land birds so why should it be any different in the sim world). If an FMC can call up the default flight plans a simmer has built in FS that's outstanding. In the case of PMDG where the FMC can't read default flight plans requiring 10 minutes of punching in waypoints for a flight from KDTW to RJAA, that's in essence madness (never how it's done in the real world). On top of that the thing not being able to handle accelerated time in the sim is a huge disability leaving a lovely Rose Royce sitting in the garage which is where my PMDG heavies have been sitting as of late. So it's all in how a developer built the autopilot code nothing more. Thankfully many developers have figured out piggy backing off the default autopilot is not a bad thing especially if your add-on is a loug haul flier. Everything else is overkill for most especially in these days where time is limited for many of us...

PREPAR3D v4/FS9.75

Share this post


Link to post
Share on other sites
Well, that certainly doesn't make sense to me. Not being privee to the actual code used in FS, I cant be sure of this, but in general this type of programming uses multiple calculations between each 'screen' generation any way. Increasing the speed merely makes the sim take bigger jumps ( as stated), but thats no reason why a plane should spin out of control. After all, it is only a sim - the planes don't actually fly. If taking bigger steps caused the phenomenon then so would a slow FPS. I've never seen this type of behaviour anyway, either in FS or any games I've written.Dave
Simulations like Flight Simulator use numerical integration methods to advance the simulation in time, step-by-step. These methods themselves can become unstable if the time step length is increased yo much - which is what effectively happens when FS in run at higher speeds. The maximum step length where the solutions remains stable depends on the characteristic of the underlying differential equations that are being solved.Google numerical integration stability - I got more than half a million hits!

Gerry Howard

Share this post


Link to post
Share on other sites
Simulations like Flight Simulator use numerical integration methods to advance the simulation in time, step-by-step. These methods themselves can become unstable if the time step length is increased yo much - which is what effectively happens when FS in run at higher speeds. The maximum step length where the solutions remains stable depends on the characteristic of the underlying differential equations that are being solved.Google numerical integration stability - I got more than half a million hits!
If you read my post above I challenge this point. Your right in essence but you'll find no issue with accelerated time with any of the default airliners. If an autopilot was built around proper SDK specifications and no liberties taken the add-on will have no issue with in sim accelerated time. It's only when a developer reinvents the wheel with the autopilot code that a problem arises. A long hauler in FS isolated to real time is all but useless to many with limited time. With a GA bird and/or regional airliner this is perfectly fine but long haul operation should always accommodate an option for in sim accelerated time. Again Few of us these days have 10 hours free to leave our computers running idle to get to a virtual destination. This has nothing to do with in sim stability, calculations, etc. it's all about how the developer designed the autopilot...

PREPAR3D v4/FS9.75

Share this post


Link to post
Share on other sites

Since autopilot control law can use differential and integral inputs, which of course are time dependent, I can see how changing the sim rate can affect the AP if the time differences aren't accounted for. In particular I would think the differential input could go haywire, when sim rate is x16 and AP is computing error increasing at 16 times the actual rate (depends obviously on how AP derives its clock).scott s..

Share this post


Link to post
Share on other sites
If you read my post above I challenge this point. Your right in essence but you'll find no issue with accelerated time with any of the default airliners. If an autopilot was built around proper SDK specifications and no liberties taken the add-on will have no issue with in sim accelerated time. It's only when a developer reinvents the wheel with the autopilot code that a problem arises. A long hauler in FS isolated to real time is all but useless to many with limited time. With a GA bird and/or regional airliner this is perfectly fine but long haul operation should always accommodate an option for in sim accelerated time. Again Few of us these days have 10 hours free to leave our computers running idle to get to a virtual destination. This has nothing to do with in sim stability, calculations, etc. it's all about how the developer designed the autopilot...
I disagee. The Learning Centre says the autopilot doesn't work at more than 16x. That's the default autopilot.

Gerry Howard

Share this post


Link to post
Share on other sites
I disagee. The Learning Centre says the autopilot doesn't work at more than 16x. That's the default autopilot.
Give it a try yourself... Maybe the 'Max Acceleration Option' in the pull down menu in FS9 is 16x either way you'll find no problem with the default autopilot when the sim is accelerated (same goes for the add-ons I mentioned above that work flawlessly under accelerated FS9...).Again like I said don't believe me or what you read, try it for yourself. :(

PREPAR3D v4/FS9.75

Share this post


Link to post
Share on other sites
Guest over and out

I would say this probably makes the crew sick, so after a while they screw up :(

Hey all,I was just wondering if maybe there is a remedy to this situation. When I do flights, especially long flights, I tend to use an increased rate of simulation. Anywhere from 2x-8x is usually what I like to do. Using 2x always works without any adverse effects. Using 4x usually works, but every once and a while the aircraft will randomly and erratically spin out of control until I return to normal sim, where I have to re-establish my routing. When I use 8x, it works for some aircraft but not for others. Examples; my PSS 777's work just fine at 8x (thankfully since transpacific flights take forever) but my PSS 757's do not. When I do 8x on the 757, it starts out fine, but then the altitude starts to fluctuate. It's mild at first; one hundred feet up and a hundred feet down from set cruise altitude. But it gets worse and soon the fluctuations can be a thousand feet above and below, at which point I clearly need to reduce the simulation rate.Does anyone know what is causing these fluctuations? I have adjusted some weather options using FSUIPC so that winds don't gust aloft and the rate of change isn't dramatic, thinking that maybe the winds were causing the issue, but it persists. Any help or insights would be greatly appreciated!Regards,Matt

Share this post


Link to post
Share on other sites
Give it a try yourself... Maybe the 'Max Acceleration Option' in the pull down menu in FS9 is 16x either way you'll find no problem with the default autopilot when the sim is accelerated (same goes for the add-ons I mentioned above that work flawlessly under accelerated FS9...).Again like I said don't believe me or what you read, try it for yourself. :(
I have tried it. I can run the default Cessna 172 at x128. The speed appears on the left-hand of the screen. If I turn the autopilot on that drops to x16. It remains at thst figure when I turn the autopilot off - though I can then reset it to x128.S0 what's wrong with my statement that The Learning Centre says the autopilot doesn't work at more than 16x. That's the default autopilot?

Gerry Howard

Share this post


Link to post
Share on other sites
I have tried it. I can run the default Cessna 172 at x128. The speed appears on the left-hand of the screen. If I turn the autopilot on that drops to x16. It remains at thst figure when I turn the autopilot off - though I can then reset it to x128.S0 what's wrong with my statement that The Learning Centre says the autopilot doesn't work at more than 16x. That's the default autopilot?
O.K. either I just had a light bulb moment or you knew something I didn't... :( I think we both have stumbled across something here... I never once thought FS9 capped the speed of acceleration within the sim itself if it detected the default autopilot was on (as they say you learn something new everyday, wow). That more than explains why planes like PMDG's birds have problems because their autopilots function outside FS9's Autopilot specifications leaving the sim able to accelerate past the safety net of 16x because the sim can't detect any autopilot is active. That explains a couple of days ago how I was able to fly the 737NG without any issue at 8x when I once thought all acceleration was impossible with PMDG products. I now need to test these birds flying at 16x or 8x and see how they all perform. I know Captainsim 757 won't work but other models that had problems might not be as bad as I once thought.See I interpret what the Learning Center is saying here as 16x should be the max for any autopilot operation within FS9 (not just default birds). The Learning Center took no account of add-on aircraft and specially built autopilots. With specially built autopilots like PMDG the simmer has to cap their acceleration settings at 16x manually because the safety net doesn't lend itself to autopilot's built outside SDK specifications. Most of the Feelthere add-ons will self cap at 16x because something in their code retained default specifications to function and when their autopilots are active FS9 sees the AP's are 'ON' and won't let acceleration go past 16x. This is interesting... :(

PREPAR3D v4/FS9.75

Share this post


Link to post
Share on other sites

×
×
  • Create New...