Replies to This Discussion

Permalink Reply by Scouser on June 6, 2009 at 3:15pm

Hi Jonny,

Lots of high wing RC planes have plain ailerons and turn nicely without using the rudder at bank angles up to 45 degrees. If you want to fly very steep turns then a rudder is required. Very few RC pilots co-ordinate the rudder in turns. Sometimes it is required ( for example on my Piper Cub ) then, you can either program a mix on the transmitter or use a "Y" lead to couple the rudder and aileron together. You need experience to adjust the mechanical linkage to the rudder if you use a "Y" lead, but that would be possible with this board as it is.

Here is a link to my project where this board flies my high wing Twinstar in steep, well co-ordinated turns on just ailerons and elevator. The turns were not intentional, but the board did an excellent job of keeping the plane level. Only after lots of "dog fighting" did the speed decrease.

http://diydrones.com/forum/topics/my-uav-devboard-project

regards,

Scouser

• ▶ Reply

Permalink Reply by jonny on June 7, 2009 at 4:28am

Scouser

Yes, I saw the video shortly after you posted it, Very impressed! I find it interesting that you mention the piper cub as an example that requires rudder input as this is the intended platform for my DevBoard (when I get round to it), and was the initial reason for my question.
You mention that I could program a mix on the transmitter, I can not see how the autopilot would recognise this? Also you mention using a "Y" lead, I assume like one would couple aileron to rudder with picopilot? But I am unsure what you mean by adjusting the rudder linkage?
I do not need to fly steep bank angles, a gentle slow turn would be preferable, however i do fly quite often in strong wind
Your input is geatly appreciated

Kind regards
Jonny

• ▶ Reply

Permalink Reply by Scouser on June 7, 2009 at 2:28pm

Jonny,

I'm sorry if I was unclear. The autopilot wouldn't recognize a transmitter-mixed rudder input. I fly my cub (which doesn't have an autopilot) with transmitter mixing. The Y lead does a similar thing between the Rx and the servos, or between the autopilot and the servos. I found that to control adverse yaw the rudder movement is smaller than the aileron movement, but with a Y lead both servos receive the same pulses. You need a long rudder horn and a short servo arm, and once it is set you can't spin or stall turn the cub.

Bill's solution is obviously the best way to go.

best regards,

Scouser

• ▶ Reply

Permalink Reply by William Premerlani on June 7, 2009 at 2:53pm

Jonny and Scouser,

I will develop firmware for rudder-aileron control....its going to need 5 PWM channels:

throttle, rudder, aileron, elevator, and manual/stabilized/RTL mode control.

So, I have a couple of questions:

1. What radio transmitter would you like use? (So I can figure out how to best set up the mode channel). You will need at least 5 channels.

2. What radio channels do you normally use for elevator, rudder, and ailerons?

3. I can assign the rudder, aileron, and elevator to the UAV DevBoard PWM channels 1, 2, and 3 in any order that you prefer. Do you have a preference?

Kindest regards,
Bill

• ▶ Reply

Permalink Reply by jonny on June 7, 2009 at 3:31pm

Bill

I use a 7channel jr unit, this has both toggle and proportional switches that are spare/aux.
I use a number of different recievers, all of which have different locations for the relevant control surfaces so I dont suppose it makes much difference what order they are as long as I can get them to comunicate with the relevant surface.

Once again Bill Many thanks
Jonny

• ▶ Reply

Permalink Reply by Scouser on June 7, 2009 at 4:02pm

Bill,

I'm quite happy with the present channel set-up. My only plane that needs co-ordinated rudder is the Clipped Wing Cub and I don't intend to put the autopilot in it. It flick rolls, tip stalls and generally misbehaves!

I use a Spektrum Dx7 which has throttle on channel 1, aileron 2, elevator 3, rudder 4. The only 3 position switch is on channel 6, but I can mix it onto where I want it.

If you are extending into aerobatics you might want the rudder yourself. Imagine a slow roll where the autopilot feeds in exactly the right amount of top rudder as you go round. I'm not sure how many mode switches you will need...........

Oh, the stabilization mode won't let my twinstar roll or loop :-(

Best regards,

Scouser

• ▶ Reply

Permalink Reply by William Premerlani on June 7, 2009 at 6:33pm

Hi Scouser,

Stabilization mode was intentionally set up to prevent rolls and loops, to keep me from crashing my Endurance, which has a tendency to go into a death spiral when it rolls. Stabilization mode was not intended for aerobatics.

However, I would consider implementing an aerobatic stabilization mode, if you could describe what you would like it to do. Though you would have to test it, I am not very good at flying aerobatics, I need the stabilization to work just the way that it does now: it makes a neutrally stable plane behave like a Gentle Lady sailplane.

Best regards,

Bill

Permalink Reply by Scouser on June 8, 2009 at 1:12am

Bill,

I had misunderstood your previous posts about developing aerobatic capability.

I don't want my autopilot to fly aerobatics. Each aerobatic might need different mode selections:
♦ pitch up while maintaining yaw and roll to fly loops.
♦ roll while maintaining alt and track to fly rolls.
♦ pitch to vertical then hold attitude and close throttle to tailslide!

It might be an interesting project but it's not my cup of tea.

I had forgotten that the stabilization mode was in when I tried to loop and roll the Twinstar!

Best regards,

Scouser

Permalink Reply by Chris Anderson on June 6, 2009 at 3:27pm

Bill,

I slightly disagree with you on an airspeed sensor. We've got a lot of experience with autpilots without and with airspeed sensors, and the bottom line is that on an average slow speed aircraft, like the powered gliders and high-wing trainers that most of use, you can't reliably handle winds of more than about 10-15mph without an airspeed sensor.

With constant throttle, you get too much variation between upwind and downwind (including difficult in making headway upwind and unreliable GPS course over ground vectors due to slow speed). To maintain constant speed, you need to adjust throttle control to take into consideration both ground speed and airspeed, lowering throttle downwind and raising it upwind.

I don't see how you can do that with IMU and GPS alone.

• ▶ Reply

Permalink Reply by Matt on June 6, 2009 at 3:59pm

Why would you adjust throttle to lower the airspeed downwind and raise it upwind? Sounds like you are trying to maintain a constant groundspeed to help out your Navigation algorithms. Flying with respect to the ground will end up throwing you into the ground oneday. I think you should be flying with respect to the air you are in. In reality you should be setting a constant airspeed and have the AP account for all the wind, bearing and heading variables.

• ▶ Reply

Permalink Reply by Chris Anderson on June 6, 2009 at 5:10pm

Not raising and lowering airspeed, but raising raising and lowering ground speed to take into consideration airspeed, for exactly the reason you state. We fly primarily by airspeed, which is why you need an airspeed sensors (which was my point in the response to Bill). But heading into a strong wind, you've got to consider ground speed too or you won't make any headway.

Going downwind, you don't want to go too fast or it makes for sloppy waypoint hits on the downwind leg. That's a function of airframes, not autopilots.

• ▶ Reply

Permalink Reply by William Premerlani on June 6, 2009 at 7:43pm

Chris,

Although I agree with much of what you said, I make the following observations...

With a good enough IMU and a GPS, it is theoretically possible to compute the wind vector, as long as the plane makes a few turns. There are several ways that you can do it, but basically they all amount to comparing the change in direction reported by the IMU with the change in direction reported by the GPS, during turns. Computing the wind vector is on my list of things to do, and I think that Louis LeGrand is working on a technique. We both have worked out the math for a couple ways that it might be done.

Once you know the wind vector, you can use the vector relationship that the ground velocity vector is equal to the airspeed vector plus the wind vector to back out airspeed vector from the IMU information and the GPS ground velocity vector.

That being said, there is a long path from theory to implementation, and I will not be trying to implement the computation of the wind vector any time soon.

In the meantime, my situation is generally what you describe, but even so, I think that an IMU gives you a leg up in windy conditions.

What I do for control and navigation is to use the IMU to tightly control the bank and pitch angles to desired values, and ignore the airspeed. My observations with my Gentle Lady (a slow airframe) is that I can easily fly in turbulent conditions, up to where the wind is about 50% of the glide speed of the plane. The IMU pitch control indirectly controls the airspeed in such a way that the plane pretty much flies at constant airspeed, with a close to optimum angle of attack. As you point out, the ground speed has large variations, but I have not had any problem with that.

I have not tried waypoints yet, but during an RTL situation, the aircraft stays rather close to the RTL point once it gets there. I think the reason is that an IMU lets you use much higher turning feedback gains without going unstable than you could use without an IMU.

My preference is to maintain constant airspeed, not constant ground speed. It seems to work for me. I agree with Matt.

Best regards,
Bill

• ▶ Reply

• ‹ Previous
• 1
• …
• 12
• 13
• 14
• 15
• 16
• …
• 56
• Next ›
• Page