> Ignore the air compressor behind the curtain! (Sorry we just watched the > wizard of oz recently).
He, he.. before deciding on the Jun-Air compressor wanted to get some test time and size the volume and flow rates. OTOH one could just stick the el-cheapo unit in a separate room and run a long supply line to say an intermediate storage tank. Looks great, and sounds like you are getting a > lot > of bang for the buck with this approach! > Also *discovering* some interesting design issues that need working; e.g. 1) a victim of my own success. the bellcrank, connections, and mount points all use bearings; friction is sooo low that have to account for the weight of the column. move it a few degress off vertical either way and it flops to the end stop. 2) the high res pot gearing and connection to the column results in a non-linear reporting of the position; i.e. from full forward to vertical is 10 degrees, the pot reports 250 to 1900, from vertical to full aft is 25 degrees, the pot reports from 1900 to 3800. at least the roll channel is linear; 1000 (full left) to 2000 (center) to 3000 (full right) 3) have to account for the difference in surface area of the linear actuator piston due to the actuator rod, at least that is a FIXED value/ratio ;-) the rotary rack and pinion actuator for the roll channel has equal surface area on both sides of the piston 4) there is a little hystersis in the system, when disturbed the column returns to around 1%, for a setpoint of 2000 that results in a value of 2020 to 1980 which is really not all that bad but it grows some the further off the vertical one moves. Test data shows the *error* is consistent. had to design a virtual bobweight that countered the off-center moment created by the column, so that needs some tuning. Probably also do some more work on the PD control algorithm. Response times are quite acceptable, but still have to get it wired into flightgear, work on a control loading algorithm in JSBSim and then some flight testing. But I'm very encouraged by the results so far 5) the really good news is the motion and action is very smooth and steady, force gradients in either direction from the setpoint increase monotonically and are precise, and no apparent glitch, jerk, or bump when moving through the neutral stick position( setpoint); change in force is VERY smooth. Currently, working at a very low psi level, in the order of the 5-12 psi range and the forces generated are quite substantial while requiring minimum CFM. Will probably move to a higher level to factor in such items as Q (dynamic pressure) and G-loading Hooking up an autopilot is going to be a real blast! again, really liked the low cost aspects. the electro-regulators cost $1200 and the actuators $330, the filtering system including the mist seperator for oil-lubed compressor around $100, miscellaneous tubing, connectors, and exhuast silencers another $100, the USB controller card and chips about $150. So for around $2K one can build a substantial control force loading system for both the pitch and roll channels. Still need to post those pics. Just haven't had the time to figure out how to handle inputting stuff to the wiki page and not sure it is a good fit. Thought about the company site, but ATM that site is down as we're moving it to a new ISP. Also taking the opportunity to rebuild portions of the sim while it is disassembled, adding panels to the center pedestal for nav/radio/iff controls and major rework of the auto-throttles. Redoing the electronics based in part on the design changes we made for the AeroNvironment GO program. Lot more capacity for additional switches, rotaries, and lights/LEDS, as well as all panels will be backlighted. Jack ------------------------------------------------------------------------------ _______________________________________________ Flightgear-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/flightgear-devel
