## Re: [Jsbsim-devel] Actuators

 Re: [Jsbsim-devel] Actuators From: - 2007-03-07 00:44:05 Attachments: text/plain ```Clipping will work (although it will be non-linear with speed) in the case of an actuator with position feedback. -- -- Tony -------------- Original message ---------------------- From: "Jon S. Berndt" > > Hi, Adam: > This is good to know. It would not be difficult to add a "gain" setting > for the actuator that could be set by another property. That property could, > in turn, be calculated as a function of qbar. So then, the actuator could > model the effects of qbar. > I wondered about that potential effect, but I was guessing that the > actuator would have some kind of position feedback so it would automatically > arrive at the commanded position. > Jon > > > Those actuators do have position feedback. But, by design, they have a > maximum force. So it is not really the "gain" that is limited. Because that > would affect it over the whole range of motion, as a function of qbar. > Instead the maximum force is limited. So, effectively, it is just a limit on > the range of motion, as a function of qbar. > So, to illustrate the point, using completely made up numbers. At 100 kts > pressing the rudder pedal 1 inch, might move the rudder by 10 degrees, and > moving it 3 inches, would move the rudder 30 degrees. While at 250 kts, > pressing the rudder pedal 1 inch, would still move the rudder by 10 degrees, > but moving it 3 inches, would only move it to the current limit of 20 > degrees. > > I see, now. OK, so the maximum range should be clipped as a function of > qbar? Actually, it doesn't seem that simple, either. As qbar increases, a > commanded motion will result in a lesser actual position across the entire > range of actuator motion. So, merely clipping the final position isn't the > answer. This is actually why I suggested a gain. As qbar increases, the gain > would reduce from 1.0 to, say, 0.95 or whatever. The slope (dOutput/dInput) > should change as well as the limit, no? > > Jon > ```

 Re: [Jsbsim-devel] Actuators From: Adam Dershowitz - 2007-03-07 00:08:02 Attachments: Message as HTML ```On Mar 6, 2007, at 1:34 PM, Berndt, Jon S wrote: > > I believe that they all matter, and in some conditions can all be > important. > And just to throw another one in there is "blow-down" or "force > limiting". For example the 737 rudder is designed so that the > maximum deflection is a function of airspeed (really dynamic > pressure). At higher airspeeds the rudder will blow down (be > pushed by the dynamic pressure, and in turn move, the actuator), > thus limiting the control authority. > > --Adam > > Hi, Adam: > > This is good to know. It would not be difficult to add a "gain" > setting for the actuator that could be set by another property. > That property could, in turn, be calculated as a function of qbar. > So then, the actuator could model the effects of qbar. > > I wondered about that potential effect, but I was guessing that the > actuator would have some kind of position feedback so it would > automatically arrive at the commanded position. > > Jon > Those actuators do have position feedback. But, by design, they have a maximum force. So it is not really the "gain" that is limited. Because that would affect it over the whole range of motion, as a function of qbar. Instead the maximum force is limited. So, effectively, it is just a limit on the range of motion, as a function of qbar. So, to illustrate the point, using completely made up numbers. At 100 kts pressing the rudder pedal 1 inch, might move the rudder by 10 degrees, and moving it 3 inches, would move the rudder 30 degrees. While at 250 kts, pressing the rudder pedal 1 inch, would still move the rudder by 10 degrees, but moving it 3 inches, would only move it to the current limit of 20 degrees. ```
 Re: [Jsbsim-devel] Actuators From: Jon S. Berndt - 2007-03-07 00:34:49 Attachments: Message as HTML ``` Hi, Adam: This is good to know. It would not be difficult to add a "gain" setting for the actuator that could be set by another property. That property could, in turn, be calculated as a function of qbar. So then, the actuator could model the effects of qbar. I wondered about that potential effect, but I was guessing that the actuator would have some kind of position feedback so it would automatically arrive at the commanded position. Jon Those actuators do have position feedback. But, by design, they have a maximum force. So it is not really the "gain" that is limited. Because that would affect it over the whole range of motion, as a function of qbar. Instead the maximum force is limited. So, effectively, it is just a limit on the range of motion, as a function of qbar. So, to illustrate the point, using completely made up numbers. At 100 kts pressing the rudder pedal 1 inch, might move the rudder by 10 degrees, and moving it 3 inches, would move the rudder 30 degrees. While at 250 kts, pressing the rudder pedal 1 inch, would still move the rudder by 10 degrees, but moving it 3 inches, would only move it to the current limit of 20 degrees. I see, now. OK, so the maximum range should be clipped as a function of qbar? Actually, it doesn't seem that simple, either. As qbar increases, a commanded motion will result in a lesser actual position across the entire range of actuator motion. So, merely clipping the final position isn't the answer. This is actually why I suggested a gain. As qbar increases, the gain would reduce from 1.0 to, say, 0.95 or whatever. The slope (dOutput/dInput) should change as well as the limit, no? Jon ```
 Re: [Jsbsim-devel] Actuators From: Ron Freimuth - 2007-03-07 01:12:12 Attachments: Message as HTML ``` ----- Original Message -----=20 From: Jon S. Berndt=20 To: JSBSim development issues=20 Sent: Tuesday, March 06, 2007 6:34 PM Subject: Re: [Jsbsim-devel] Actuators Hi, Adam: This is good to know. It would not be difficult to add a "gain" = setting for the actuator that could be set by another property. That = property could, in turn, be calculated as a function of qbar. So then, = the actuator could model the effects of qbar. I wondered about that potential effect, but I was guessing that = the actuator would have some kind of position feedback so it would = automatically arrive at the commanded position. Jon Those actuators do have position feedback. But, by design, they have = a maximum force. So it is not really the "gain" that is limited. Because = that would affect it over the whole range of motion, as a function of = qbar. Instead the maximum force is limited. So, effectively, it is just = a limit on the range of motion, as a function of qbar.=20 So, to illustrate the point, using completely made up numbers. At = 100 kts pressing the rudder pedal 1 inch, might move the rudder by 10 = degrees, and moving it 3 inches, would move the rudder 30 degrees. While = at 250 kts, pressing the rudder pedal 1 inch, would still move the = rudder by 10 degrees, but moving it 3 inches, would only move it to the = current limit of 20 degrees.=20 I see, now. OK, so the maximum range should be clipped as a function = of qbar? Actually, it doesn't seem that simple, either. As qbar = increases, a commanded motion will result in a lesser actual position = across the entire range of actuator motion. So, merely clipping the = final position isn't the answer. This is actually why I suggested a = gain. As qbar increases, the gain would reduce from 1.0 to, say, 0.95 or = whatever. The slope (dOutput/dInput) should change as well as the limit, = no? Jon I've been modeling MSFS rudder deflection some time now. Typically = the rudder on a jet transport can only move +/- 5 deg above 160 kts or = so. =20 The FS AIR file has tables for "control deflection vs q". I think = it is actually a slope. I set the rudder so it drops about as 1/q_ = above 90 lb/ft^2. That way, the physical yaw moment is limited to a = constant value at higher speeds. This keeps the AC from yawing = violently with full rudder pedals. However, FS also has tables that allow one to set the Lift (slope) = and corresponding control moments vs Mach. I've also set them to limit = Yaw (etc.) moments. Otherwise, low q, but high Mach numbers (high = altitudes) would result in excessive moments (even though deflection is = limited to a certain value). Seems that if one has an accurate hinge moment he could control = deflection based on the limitation of actuator force. Spoilers also tend to blow down. One might expect q*S*Cd_sp to be = nearly constant at high speeds, rather than increasing with speed. I don't think Flaps blow down, but they lose effectiveness at = higher indicated speeds. Lift polars show the effect, by 250 kts CL vs = alpha at a given flaps deflection has flattened out quite a bit. Spoilers are typically controlled with a spoileron mixer. It = tries to move the spoilerons so a constant roll rate is possible, = regardless of IAS. =20 All the above is stuff I've found on the WEB, or in FM = descriptions. General descriptions are often given, it's up to we = engineers to figure out how to model them. =20 Ron ```
 Re: [Jsbsim-devel] Actuators From: Adam Dershowitz - 2007-03-07 16:28:27 Attachments: Message as HTML ```On Mar 6, 2007, at 4:34 PM, Jon S. Berndt wrote: > >> Hi, Adam: >> This is good to know. It would not be difficult to add a "gain" >> setting for the actuator that could be set by another property. >> That property could, in turn, be calculated as a function of qbar. >> So then, the actuator could model the effects of qbar. >> I wondered about that potential effect, but I was guessing that >> the actuator would have some kind of position feedback so it would >> automatically arrive at the commanded position. >> Jon > > Those actuators do have position feedback. But, by design, they > have a maximum force. So it is not really the "gain" that is > limited. Because that would affect it over the whole range of > motion, as a function of qbar. Instead the maximum force is > limited. So, effectively, it is just a limit on the range of > motion, as a function of qbar. > So, to illustrate the point, using completely made up numbers. At > 100 kts pressing the rudder pedal 1 inch, might move the rudder by > 10 degrees, and moving it 3 inches, would move the rudder 30 > degrees. While at 250 kts, pressing the rudder pedal 1 inch, would > still move the rudder by 10 degrees, but moving it 3 inches, would > only move it to the current limit of 20 degrees. > > I see, now. OK, so the maximum range should be clipped as a > function of qbar? Actually, it doesn't seem that simple, either. As > qbar increases, a commanded motion will result in a lesser actual > position across the entire range of actuator motion. So, merely > clipping the final position isn't the answer. This is actually why > I suggested a gain. As qbar increases, the gain would reduce from > 1.0 to, say, 0.95 or whatever. The slope (dOutput/dInput) should > change as well as the limit, no? > > Jon > Actually it depends on the airplane. Some stay linear to a maximum force then just clip. Others use different gains at different airspeeds. ```
 Re: [Jsbsim-devel] Actuators From: - 2007-03-07 00:44:05 Attachments: text/plain ```Clipping will work (although it will be non-linear with speed) in the case of an actuator with position feedback. -- -- Tony -------------- Original message ---------------------- From: "Jon S. Berndt" > > Hi, Adam: > This is good to know. It would not be difficult to add a "gain" setting > for the actuator that could be set by another property. That property could, > in turn, be calculated as a function of qbar. So then, the actuator could > model the effects of qbar. > I wondered about that potential effect, but I was guessing that the > actuator would have some kind of position feedback so it would automatically > arrive at the commanded position. > Jon > > > Those actuators do have position feedback. But, by design, they have a > maximum force. So it is not really the "gain" that is limited. Because that > would affect it over the whole range of motion, as a function of qbar. > Instead the maximum force is limited. So, effectively, it is just a limit on > the range of motion, as a function of qbar. > So, to illustrate the point, using completely made up numbers. At 100 kts > pressing the rudder pedal 1 inch, might move the rudder by 10 degrees, and > moving it 3 inches, would move the rudder 30 degrees. While at 250 kts, > pressing the rudder pedal 1 inch, would still move the rudder by 10 degrees, > but moving it 3 inches, would only move it to the current limit of 20 > degrees. > > I see, now. OK, so the maximum range should be clipped as a function of > qbar? Actually, it doesn't seem that simple, either. As qbar increases, a > commanded motion will result in a lesser actual position across the entire > range of actuator motion. So, merely clipping the final position isn't the > answer. This is actually why I suggested a gain. As qbar increases, the gain > would reduce from 1.0 to, say, 0.95 or whatever. The slope (dOutput/dInput) > should change as well as the limit, no? > > Jon > ```