> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On > Behalf Of Martin Spott > Sent: Sunday, May 13, 2007 5:01 PM > To: flightgear-devel@lists.sourceforge.net > Subject: Re: [Flightgear-devel] More ideas on dogfighting > > Hi Bill, > > "Bill Galbraith" wrote: > > > "Relative responses of the motion system, visual system, > and cockpit > > instruments shall be coupled closely to provide integrated sensory > > cues 6 These systems shall respond to abrupt pitch, roll and yaw > > inputs at the pilot's position within 150/300 milliseconds of the > > time, but not before the time, when the airplane would > respond under > > the same conditions. [...]" > > Uh, 300 ms response time looks pretty bad to me at a first > glance. They already include the time which the 'real' > aircraft would need to respond aerodynamically - right ? > > Personally I'd go crazy in the real Cessna if it would take > me one third of a second until the beast starts !! responding > to a control movement - this would turn almost every landing > at gusty crosswind into a really difficult situation .... >
Sorry. I wasn't specific enough (and figured SOMEONE would question what I wrote). That delay is not counting the aircraft aerodynamic delay. In the simulation world, we set up special code for this testing, so that it takes the same path through normal code, it just bypasses the aerodynamic effects and recognizes the step control input, and generates a step output on the three output systems (instruments, visual, motion). Typically, one would also do a throughput analysis, if you have separate computers for various functions (you do on real sims). You look at the control input happened at t=0, but ooo, it just missed a data transfer from the control loading computer to the host, so you have to wait for the next one to come along. The host recognizes the input and generates the output, and sends the signal out to the I/O, visual, and motion computers. Those transfers are usually initiated by the host at the end of a frame, but if they aren't you have to play the game of ooo, I just missed that data transfer, I have to wait for the next one to happen, all the time adding up the delays. If the theorical delay is too great, you don't have a good architecure for something (data transfer, host execution, whatever) and may have to make some changes. Your theorical value should be less than the 150 or 300 msec, since this is the worst that it's allowed to be. You won't always hit the theorical number that you calculate, but hopefully you are less some of the time. Bill ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel