- Joined
- Jan 17, 2009
- Messages
- 5,204
- Reaction score
- 1,547
For those who have not heard, there is finally an off-the-shelf R/C plane "Stabilizer" that can make a rocket fly vertically.
There is more than one, but the one I got to try, and have also seen fly, is the Eagle Tree Guardian (It seems to be the best one). It is an autopilot, or gyro, that has two modes. The 2D mode is the one that is great for rockets, because it can keep the rocket pointing STRAIGHT UP. More tech info here:
https://www.eagletreesystems.com/index.php?route=product/product&product_id=50
When I got to try one, it was in an electric R/C plane. It is set up by plugging servos into the Guardian. Then plugging the Guardian into the proper servo outputs from the R/C receiver. Yes, you do need to have an R/C system in the loop, but for a rocket, mainly for the pre-flight set-up. For the actual rocket flight, the transmitter is on but nothing needs to be done but sit back and watch the Guardian do its thing.
Anyway, for the plane, the Guardian was mounted horizontally , lengthwise along the line of flight. It was plugged into the elevator (pitch) and aileron (roll) servo channels. To initialize it, the plane is put onto the ground, level in pitch and level in roll. Turn on the power, and the Guardian starts to initlialize, "learning" what wings level and pitch level is (It is learning what it then assumes to be a "virtual horizon". This is why it needs to be level, not angled). After 15 seconds or so, it is ready. Although, if there are any obvious small trim errors, the transmitter trim levers can be used to adjust those before takeoff (that is important for rocket flight).
When I used it with a plane, the transmitter can still be used to fly the plane, of course. The magic happens when in flight you let the sticks go to neutral, because when in 2D mode, the Guardian will level the wings and level the pitch. Put the plane into a dive and let go, it pulls out. Put the plane into a 90 degree roll, let go it levels the wings. Have it inverted and let go, it levels out. Just, incredible. Fly the plane for 15 minutes, get it into trouble and let go, and it levels the plane exactly like it did 15 minutes before, there is NO DRIFT! Also, on windy days, it holds the plane pretty steady, not nearly as much bouncing around from gusts (this is the second biggest reason for using it, to handle winds a lot better).
And so this is how and why it can make a rocket fly straight up. Use it exactly the same way, with the Guardian mounted LEVEL as it is for a plane in 2D, NOT mounted end-on at 90 degrees. The Guardian is about 2" long, so to mount it level into a rocket, the smallest practical diameter is an Estes BT-70, about 2.2" diameter.
The same exact servo outputs are used. Pitch, yeah, of course. But also, to use the aileron servo output for "roll". It might sound confusing, but when the Guardian is used that way, mounted horizontally in a rocket pointed vertically, it ends up that the "roll" axis of the Guardian is what is needed to control the Yaw axis of a vertical rocket. And in the same vein, the "yaw" axis of the Guardian, would be used to control roll in a rocket.
Now, the Guardian has another mode for controlling flight, 3D mode. In that mode, it does not know anything about what is horizontal. It simply tries to keep the plane pointing the same direction it was aimed at, when you went to neutral on the sticks. However, that can drift over time. I saw an R&D project attempting to use the 3D mode, of another brand of autopilot, and clearly it did not work as advertised, the R/C Rocket Boosted Glider pitched down about 30 degrees on a B powered flight (a heck of a lot of "drift"), just as bad, if not worse, than old R/C gyros. For whatever reasons, they chose to do it in 3D mode (probably to be able to orient the autopilot board parallel to the fuselage). So, I really do not suggest using 3D mode for rocket boosts, unless you absolutely know what you are doing and do some VERY careful testing.
The 2D mode, again, is incredible. Another R&D project, by Alyssa Stenberg, used the Guardian in 2D mode, and it worked great for her. She had rockets with control surfaces, and even one with a gimbaled engine, which flew finless a couple of times.
At NARAM last summer, she did a demo flight of a model with control fins on the nose, that was angled 30 degrees off vertical (she initialized it vertically, then once running properly and trimmed she angled the rail 30 degrees. If she had tilted it 30 degrees, then initialized it, the Guardian's "virtual horizon" reference would have been tilted 30 degrees in relation to the real world horizon, and it would have wandered thru the sky trying to fly anywhere from 30 to 60 degrees off-vertical, depending on its roll orientation from where it was when initialized). It was flown 2-staged. After liftoff, it pitched down to about 45 degrees. After staging it had enough velocity to start to pitch up, and by burnout (I think it was an E9) it was pointed absolutely vertical, proving it worked. Three photos are attached, showing that flight.
Anyway, I wanted to let people know, that if you have wanted to add vertical guidance to a rocket, but felt the technology was way too complex, not reliable enough, and/or way too expensive, that is no longer a problem. The Guardian is $75. Almost any cheap R/C system which uses common plug-in type servos, will work with it (as opposed to planes with servos built into the receiver and very small planes with funky systems).
The transmitter..... it is turned on, and before liftoff maybe adjust the trims if a surface seems to be a bit off. But after that, hold it but do not do anything, let the Guardian do it all. Although if a person did move the elevator (or aileron) stick, the rocket would veer off (much as it does to let people fly planes), but if you let the stick go back to neutral, then the Guardian would steer it back towards vertical.
Now, I would not suggest just adding this to say any random L powered rocket. Anyone who did that, ought to work out a small BT-70 prototype first, to get accustomed to using it and to work out the mechanics of for example making moveable nose fins. Once that works out, then it would be time to scale it up, and make use of many lessons learned from the prototype. But also don't go nuts, for an HPR rocket it ought to fly slower, rather than faster. Actually the most impressive on a guided rocket is a long burn anyway. Due to potential extreme aerodynamic loads on the control surfaces, the bigger and faster the rocket, the more sturdy and powerful the servos should be. If the rocket is going to fly really fast, then the control fins either should be smaller, or set up to move fewer degrees, than for a rocket that would be flying slower.
Now, I am not entirely sure how well the overall R/C and Guardian work if the receiver loses signal, such as for example a HPR rocket flying so high that it loses signal from the transmitter. I think that if a person was using a receiver with "fail safe" technology, set up to stay at neutral if it loses signal, that the Guardian would keep steering the rocket vertically. The simple ground-test for that would be to see what happens, once the system is up and running, by turning off the transmitter to simulate loss of signal, causing the fail-safe to kick in.
The ultimate solution to that loss of signal problem would be to program a microcontroller (or use/build an old-school servo controller) to produce the same servo 1.50 millisecond control pulses (servo neutral) as an R/C receiver produces. And use that onboard, plugged directly into the Guardian rather than a receiver. If done that way, it would also be desirable to be able to adjust the pulses a little bit, like 1.2 to 1.8 milliseconds, to allow a decent amount of trim adjustment.
For those who worry, an onboard guidance system like this is totally legal to do. It is to make the rocket fly vertically, NOT make it go towards a target.
- George Gassaway
(flier of automated guided model rockets since 1988, with Sunguidance)
There is more than one, but the one I got to try, and have also seen fly, is the Eagle Tree Guardian (It seems to be the best one). It is an autopilot, or gyro, that has two modes. The 2D mode is the one that is great for rockets, because it can keep the rocket pointing STRAIGHT UP. More tech info here:
https://www.eagletreesystems.com/index.php?route=product/product&product_id=50
When I got to try one, it was in an electric R/C plane. It is set up by plugging servos into the Guardian. Then plugging the Guardian into the proper servo outputs from the R/C receiver. Yes, you do need to have an R/C system in the loop, but for a rocket, mainly for the pre-flight set-up. For the actual rocket flight, the transmitter is on but nothing needs to be done but sit back and watch the Guardian do its thing.
Anyway, for the plane, the Guardian was mounted horizontally , lengthwise along the line of flight. It was plugged into the elevator (pitch) and aileron (roll) servo channels. To initialize it, the plane is put onto the ground, level in pitch and level in roll. Turn on the power, and the Guardian starts to initlialize, "learning" what wings level and pitch level is (It is learning what it then assumes to be a "virtual horizon". This is why it needs to be level, not angled). After 15 seconds or so, it is ready. Although, if there are any obvious small trim errors, the transmitter trim levers can be used to adjust those before takeoff (that is important for rocket flight).
When I used it with a plane, the transmitter can still be used to fly the plane, of course. The magic happens when in flight you let the sticks go to neutral, because when in 2D mode, the Guardian will level the wings and level the pitch. Put the plane into a dive and let go, it pulls out. Put the plane into a 90 degree roll, let go it levels the wings. Have it inverted and let go, it levels out. Just, incredible. Fly the plane for 15 minutes, get it into trouble and let go, and it levels the plane exactly like it did 15 minutes before, there is NO DRIFT! Also, on windy days, it holds the plane pretty steady, not nearly as much bouncing around from gusts (this is the second biggest reason for using it, to handle winds a lot better).
And so this is how and why it can make a rocket fly straight up. Use it exactly the same way, with the Guardian mounted LEVEL as it is for a plane in 2D, NOT mounted end-on at 90 degrees. The Guardian is about 2" long, so to mount it level into a rocket, the smallest practical diameter is an Estes BT-70, about 2.2" diameter.
The same exact servo outputs are used. Pitch, yeah, of course. But also, to use the aileron servo output for "roll". It might sound confusing, but when the Guardian is used that way, mounted horizontally in a rocket pointed vertically, it ends up that the "roll" axis of the Guardian is what is needed to control the Yaw axis of a vertical rocket. And in the same vein, the "yaw" axis of the Guardian, would be used to control roll in a rocket.
Now, the Guardian has another mode for controlling flight, 3D mode. In that mode, it does not know anything about what is horizontal. It simply tries to keep the plane pointing the same direction it was aimed at, when you went to neutral on the sticks. However, that can drift over time. I saw an R&D project attempting to use the 3D mode, of another brand of autopilot, and clearly it did not work as advertised, the R/C Rocket Boosted Glider pitched down about 30 degrees on a B powered flight (a heck of a lot of "drift"), just as bad, if not worse, than old R/C gyros. For whatever reasons, they chose to do it in 3D mode (probably to be able to orient the autopilot board parallel to the fuselage). So, I really do not suggest using 3D mode for rocket boosts, unless you absolutely know what you are doing and do some VERY careful testing.
The 2D mode, again, is incredible. Another R&D project, by Alyssa Stenberg, used the Guardian in 2D mode, and it worked great for her. She had rockets with control surfaces, and even one with a gimbaled engine, which flew finless a couple of times.
At NARAM last summer, she did a demo flight of a model with control fins on the nose, that was angled 30 degrees off vertical (she initialized it vertically, then once running properly and trimmed she angled the rail 30 degrees. If she had tilted it 30 degrees, then initialized it, the Guardian's "virtual horizon" reference would have been tilted 30 degrees in relation to the real world horizon, and it would have wandered thru the sky trying to fly anywhere from 30 to 60 degrees off-vertical, depending on its roll orientation from where it was when initialized). It was flown 2-staged. After liftoff, it pitched down to about 45 degrees. After staging it had enough velocity to start to pitch up, and by burnout (I think it was an E9) it was pointed absolutely vertical, proving it worked. Three photos are attached, showing that flight.
Anyway, I wanted to let people know, that if you have wanted to add vertical guidance to a rocket, but felt the technology was way too complex, not reliable enough, and/or way too expensive, that is no longer a problem. The Guardian is $75. Almost any cheap R/C system which uses common plug-in type servos, will work with it (as opposed to planes with servos built into the receiver and very small planes with funky systems).
The transmitter..... it is turned on, and before liftoff maybe adjust the trims if a surface seems to be a bit off. But after that, hold it but do not do anything, let the Guardian do it all. Although if a person did move the elevator (or aileron) stick, the rocket would veer off (much as it does to let people fly planes), but if you let the stick go back to neutral, then the Guardian would steer it back towards vertical.
Now, I would not suggest just adding this to say any random L powered rocket. Anyone who did that, ought to work out a small BT-70 prototype first, to get accustomed to using it and to work out the mechanics of for example making moveable nose fins. Once that works out, then it would be time to scale it up, and make use of many lessons learned from the prototype. But also don't go nuts, for an HPR rocket it ought to fly slower, rather than faster. Actually the most impressive on a guided rocket is a long burn anyway. Due to potential extreme aerodynamic loads on the control surfaces, the bigger and faster the rocket, the more sturdy and powerful the servos should be. If the rocket is going to fly really fast, then the control fins either should be smaller, or set up to move fewer degrees, than for a rocket that would be flying slower.
Now, I am not entirely sure how well the overall R/C and Guardian work if the receiver loses signal, such as for example a HPR rocket flying so high that it loses signal from the transmitter. I think that if a person was using a receiver with "fail safe" technology, set up to stay at neutral if it loses signal, that the Guardian would keep steering the rocket vertically. The simple ground-test for that would be to see what happens, once the system is up and running, by turning off the transmitter to simulate loss of signal, causing the fail-safe to kick in.
The ultimate solution to that loss of signal problem would be to program a microcontroller (or use/build an old-school servo controller) to produce the same servo 1.50 millisecond control pulses (servo neutral) as an R/C receiver produces. And use that onboard, plugged directly into the Guardian rather than a receiver. If done that way, it would also be desirable to be able to adjust the pulses a little bit, like 1.2 to 1.8 milliseconds, to allow a decent amount of trim adjustment.
For those who worry, an onboard guidance system like this is totally legal to do. It is to make the rocket fly vertically, NOT make it go towards a target.
- George Gassaway
(flier of automated guided model rockets since 1988, with Sunguidance)
Last edited:
