SMALL - espritmodel.com SMALL - Telemetry SMALL - Radio
Reply
Thread Tools
Old Dec 16, 2011, 08:40 PM
Registered User
Joined Jul 2011
721 Posts
Discussion
An Obvious Flaw in Phoenix Flight Modeling

I knew I didn't like the way Phoenix models 3d foam models. But I came across a REALLY extreme example of it being messed up today that, at least for me, is very reproducible.

Take the full size 4-site model and use the scaling wizard to scale it up 200%. Physics speed should be 100%.

Now fly it at full throttle. You may need to wait until you get in the air to apply full throttle.

Fly around a bit and watch the fun occur. It won't be subtle.
Madratter is offline Find More Posts by Madratter
Reply With Quote
Sign up now
to remove ads between posts
Old Dec 17, 2011, 07:44 AM
Bit-Twiddler and Flyer
USA, VA, Chantilly
Joined Apr 2002
1,702 Posts
Physics after scaling is not a physics flaw. You need to manually tweak the flight parameters after any scaling over 30-50 percent or so.
ggunners is offline Find More Posts by ggunners
RCG Plus Member
Old Dec 17, 2011, 07:55 AM
I don't want to "Switch Now"
pmackenzie's Avatar
Toronto (Don Mills), Canada
Joined Dec 2002
16,966 Posts
I tried it and first thought was - "seems fine, whats he talking about?". Then all broke loose

That is just wacky, and looks like some sort of bug to me.
pmackenzie is online now Find More Posts by pmackenzie
Reply With Quote
Old Dec 17, 2011, 08:24 AM
Registered User
Joined Jul 2011
721 Posts
Quote:
Originally Posted by ggunners View Post
Physics after scaling is not a physics flaw. You need to manually tweak the flight parameters after any scaling over 30-50 percent or so.
Do it and get back to me on this. This is not just a matter of it being a little heavy or the drag being slightly wrong. This is REALLY messed up. And by the way. I have now duplicated this bug on a non foamy airplane as well. And on that one, I didn't even need to go full throttle.
Madratter is offline Find More Posts by Madratter
Reply With Quote
Old Dec 17, 2011, 08:28 AM
Registered User
Joined Jul 2011
721 Posts
Quote:
Originally Posted by pmackenzie View Post
I tried it and first thought was - "seems fine, whats he talking about?". Then all broke loose

That is just wacky, and looks like some sort of bug to me.
Pretty crazy isn't it .
Madratter is offline Find More Posts by Madratter
Reply With Quote
Old Dec 18, 2011, 12:16 AM
FS One
MSelig's Avatar
United States, IL, Champaign
Joined Sep 2005
501 Posts
Quote:
Originally Posted by ggunners View Post
Physics after scaling is not a physics flaw. You need to manually tweak the flight parameters after any scaling over 30-50 percent or so.
Tweak?
MSelig is offline Find More Posts by MSelig
Reply With Quote
Old Dec 18, 2011, 02:54 AM
Registered Addict
Kris...'s Avatar
United Kingdom, Stoke-on-Trent
Joined Jul 2010
5,247 Posts
Quote:
Originally Posted by MSelig View Post
Tweak?
Change
Kris... is offline Find More Posts by Kris...
Reply With Quote
Old Dec 18, 2011, 06:03 PM
Registered User
Joined Jul 2011
721 Posts
If anyone wants to see weird things with a non foam based aircraft in Phoenix, scale the Edge 540 (Red Bull) up 200% and start flying. As a teaser, let's just say that EDFs are tame in comparison. Fly full throttle. Things will be normal for a while, and then they won't. This particular problem will manifest itself well below full throttle, but do it at full throttle first.
Madratter is offline Find More Posts by Madratter
Reply With Quote
Old Dec 19, 2011, 01:19 AM
Registered Addict
Kris...'s Avatar
United Kingdom, Stoke-on-Trent
Joined Jul 2010
5,247 Posts
The Scale wizard is notorious for being buggy you still have to tweak/change parameters to make it fly well.
Kris... is offline Find More Posts by Kris...
Reply With Quote
Old Dec 19, 2011, 07:33 AM
Registered User
Joined Jul 2011
721 Posts
Quote:
Originally Posted by Kris... View Post
The Scale wizard is notorious for being buggy you still have to tweak/change parameters to make it fly well.
Again, this isn't a matter of flying well vs not well. This is a matter of flying well for a time, and then things change drastically all of the sudden. If it was a simple parameter that needed to be tweaked problem, then the problem I am seeing would manifest itself right from the start.
Madratter is offline Find More Posts by Madratter
Reply With Quote
Old Dec 19, 2011, 09:20 AM
Registered Addict
Kris...'s Avatar
United Kingdom, Stoke-on-Trent
Joined Jul 2010
5,247 Posts
Quote:
Originally Posted by Madratter View Post
This is a matter of flying well for a time, and then things change drastically all of the sudden.
Thats my point what you say here is a classic bug and why the scale wizard feature is unstable.
Kris... is offline Find More Posts by Kris...
Reply With Quote
Old Dec 19, 2011, 01:03 PM
Registered User
Joined Jul 2011
721 Posts
I'm going to go out on a limb here but I have a background in physics and I am a software engineer. My guess is that the time step they are using in their physics model is not sufficiently small when dealing with the forces generated by larger airplanes. That is only one possibility but that is my best guess based on what I am seeing. Things remain fine when flying with the large model until some large force is generated by a maneuver that overwhelms the time step they are using. On top of that, based on what I am seeing, I believe their drag model has issues as well. At any rate, whether those things are bugs or problems with the physics engine is a matter of semantics. What is absolutely clear and demonstrable is that if you scale planes up, bad stuff can happen.
Madratter is offline Find More Posts by Madratter
Reply With Quote
Old Dec 19, 2011, 02:37 PM
Registered Addict
Kris...'s Avatar
United Kingdom, Stoke-on-Trent
Joined Jul 2010
5,247 Posts
Well your the software engineer and you have answered your own question
Kris... is offline Find More Posts by Kris...
Reply With Quote
Old Dec 19, 2011, 07:44 PM
FS One
MSelig's Avatar
United States, IL, Champaign
Joined Sep 2005
501 Posts
Quote:
Originally Posted by Madratter View Post
I'm going to go out on a limb here but I have a background in physics and I am a software engineer. My guess is that the time step they are using in their physics model is not sufficiently small when dealing with the forces generated by larger airplanes. That is only one possibility but that is my best guess based on what I am seeing. Things remain fine when flying with the large model until some large force is generated by a maneuver that overwhelms the time step they are using. On top of that, based on what I am seeing, I believe their drag model has issues as well. At any rate, whether those things are bugs or problems with the physics engine is a matter of semantics. What is absolutely clear and demonstrable is that if you scale planes up, bad stuff can happen.
Time Step: It actually goes the other way. With larger planes, the rates are slower (not as nimble) and you can run at slower physics update rates, e.g. ~30 Hz might work for a jet transport in Microsoft Flight Sim. But when going smaller, the rates are higher and things can blow up unless the rate is increased. We run FS One at 300 Hz on physics side. So when we scale up to larger size, this 300 Hz is really overkill, i.e. things *actually* get easier.

MSFS ~30 Hz Ref: Michael Zyskowski, AIAA Paper 2003-5818 (August 2003)
download.microsoft.com/download/1/7/8/17808bed-9b1e-4542-90a6-4c1d8af27fae/Aircraft_Sim_Tech_Zyskowski.pdf
MSelig is offline Find More Posts by MSelig
Reply With Quote
Old Dec 19, 2011, 08:51 PM
Registered User
Joined Jul 2011
721 Posts
Quote:
Originally Posted by MSelig View Post
Time Step: It actually goes the other way. With larger planes, the rates are slower (not as nimble) and you can run at slower physics update rates, e.g. ~30 Hz might work for a jet transport in Microsoft Flight Sim. But when going smaller, the rates are higher and things can blow up unless the rate is increased. We run FS One at 300 Hz on physics side. So when we scale up to larger size, this 300 Hz is really overkill, i.e. things *actually* get easier.

MSFS ~30 Hz Ref: Michael Zyskowski, AIAA Paper 2003-5818 (August 2003)
download.microsoft.com/download/1/7/8/17808bed-9b1e-4542-90a6-4c1d8af27fae/Aircraft_Sim_Tech_Zyskowski.pdf
Hmm. That makes sense, but on the other hand, it really does look like a time step kind of problem. I wonder if they are doing something sneaky like reducing the time step for larger aircraft, and they are overshooting what they can get away with. Oh, well, at the end of the day it really doesn't matter. Things are messed up, one way or another.
Madratter is offline Find More Posts by Madratter
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Need suggestions to hide flaw in mahogany deck RichMPL Scale Boats 8 Sep 16, 2011 03:07 AM
Discussion fundamental flaw in a two-party government? lymon Life, The Universe, and Politics 38 Jul 26, 2011 06:34 PM
Alert Stolen Planes/Kits in Phoenix - Please keep an eye out crxmanpat Gilbert Quiet Flyers (Arizona) 11 Apr 22, 2011 08:23 AM
Alert Bad CS I can handle, but an obvious attempt at deceit? eeephreakeee Banana Hobby 12 Mar 23, 2011 10:59 PM
Discussion Modeling this would be an exercise in SLENDER! crashsiegel Electric Plane Talk 3 Sep 10, 2007 07:55 PM