hi everyonefirst of all thank you very much to the whole forum community for the great help you give me in starting the adventure of home cockpit building.As I sais in my previouse posts I need to build a desktop panel that allows me to replicate realistically airline operations (ifr - multicrew) with fs2002. for this reason I need all rotaries to be capable of very fast twistings without missing inputs, so that I can switch between headings, altitudes etc... with just a coupple of quick movements, like in the real thing!I've tried keyboard hacking and lpt-switch, both proved very effective for pushbuttons and switches but not for rotaries.Which device should I buy-build?thanks

Trallucio,While it is possible to get fairly reasonable results from it, Windows is not technically a "real-time system". One cannot guarantee when an event will be recognized and acted upon. The basic software timer under Win98 will go no smaller than 55 milliseconds. NT (and I think XP as well) do better with timer settings as small as 10 milliseconds. Even so, you have no way to determine when an event will actually get serviced. The OS will place a message in a process's cue on the clock interval if there are no higher priority messages pending, otherwise, it waits. Clock timer ticks don't get lost, but they may get combined.For a non real-time OS this isn't all that bad. However, when it comes to rapidly spinning rotaries and expectations of smoothly incrementing/decrementing displays, it's right on the ragged edge.A fix for this is to handle the time critcal transactions locally and communicate the results back to Windows.As an example, the VOR/GS CDI project on my site uses a PIC16F628 to scan the two outputs of an inexpensive rotary encoder. It handles contact debounce and decoding. Turning the encoder increments or decrements a course heading variable which in turn is used to drive the observed heading and back course displays. The same PIC drives the 6, 7-segment LED displays. The result is a display that responds instantly to changes via the rotary encoder.The OBS/BC variable is polled by the host PC a few times a second.You might consider using a similar approach in your project. True, there is a learning curve with micro controllers, but they are quite versatile little beasties. Once you come up to speed with them, you'll find uses for them in any number of applications.Mikewww.mikesflightdeck.com

The GoFlight MCP autopilot does not "miss clicks" - it's talking directly to FS via a DLL rather than using keystroke stuff, so no event is ever lost. Also, the Precision Flight Controls avionics panel, courtesy of Pete Dowson's PFC.DLL, never misses a click either.So what you might consider, if you want an inexpensive autopilot that looks a lot like the real thing, is to get the GoFlight one (US$270 or so) and move the pots and switches to a custom-built panel.Or, do what I do and just use it as-is. My interest is more in general-purpose controls, and also I don't fly anything bigger than the Lear most of the time, so the MCP works perfectly for my needs.Dave Blevins

mikethank you very much, you have been clear and usefull as usual. indeed what I care is not "real-time" but just not missing inputs, what I don't want is twisting the rotary and get only 5-6 degrees of headung change instead of the 24 that it sholud provide, I don't care if changes are not smooth and fast. I know the two things are connected in "polling". May be using a "real" rotary encoder can help? I know these devices produce a digital signal so I suppose their output can be buffered locally, so that when the os poll the device this will send all stored inputs instead of sending only those that are generated in that short polling time. Is this right?tanksblaveI saw goflight mcp on their website, and I really like it, it would provide exactly what I'm looking for, but I can't spend 300$ for that!

Trallucio,Yes, local buffering could be used.Mikewww.mikesflightdeck.com

mery crhistmas everyone!so what about fsbus? how does it work with rotaries? does it miss "clicks" or not?Is it worth a try?thanks

Maybe George can answer the question about missing detents with the 20 detent switches he uses? George uses the REDec circuit with FSBUS and says it works well.George,could you clear this up for us please?You also might want to read the post titled "Alternate to REDec09b" Which I believe its currently on this forums second page. Kind Regards,

bradthanks for your reply. I've posted my questions on the relevant topic.byep.s.(I would be interested in reciving more info about redec. can you send me pcb, drawings etc? [email protected])