[simpits-tech] Avionics knobs

[Matt Bailey]

Sounds interesting. Encoders seem like overkil for my project though,
unless they would be more accurate, ie letting you flip the knob very
quickly without skipping updates to the computer.

I can think of many uses for the second unit you describe...mag switch,
avionics input selectors, radar range, EFIS map settings, etc etc etc...

	-Matt Bailey

"Alan D. Mazurka" wrote:
> 
> hi,
> 
> this might be a tad premature, but....
> 
> many folks have this need, and i was very disappointed with the most
> popular offerings out there.
> 
> a friend of mine has worked with me to develop a PIC-based rotary
> which outputs a well-formed 20msec pulse suitable for Hagstrom (and
> probably Epic, not tested yet)
> 
> we use grayhill encoders, and the design accepts 12, 24, and 36 step
> devices. it has an excellent "hand feel", and all the components are
> mounted right behind the rotary, so there is no external card to
> connect with. we have several prototypes, and are close to etching the
> first batch of beta units. they are static protected,
> reverse-connection protected, and have caps across the processor for
> filtering. a nice, tight design.
> 
> the "other" project (much more mature software-wise) is based on a
> mechanical rotary switch. it's a 12-position switch which has an
> internal stop which lets you make it assume any number of positions. i
> needed this for throttles in a train simulator - one provides 8 steps
> throttle,  another provides 9 steps brake. it only uses 2 hagstrom
> lines: the processor makes the switch behave just like a rotary.
> 
> i have since discovered it could be useful for many applications where
> you do not want a knob to continuously spin. it has a "strong" force,
> with an assuring "click" - much stronger than the rotaries. it has
> been rock-solid in execution.
> 
> at some point, i would like to work with a few "beta" testers for
> their input. because it's processor-based, we can tailor the response.
> 
> i'm curious about how many people might be interested in this tidbit,
> and what they would want from an "ideal" encoder.
> 
>  - adm -