Gimballed Motor Mount / Gyroscopic Stabilization

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

chronister

Member
Joined
May 27, 2009
Messages
16
Reaction score
0
Hi folks. I know that in NASA they use gimballed motor mounts, controlled by gyroscopes, to direct the thrust of the engine, for stability and steering.

For model rockets, it seems to work OK just to use fins, but for many scale models this is not very realistic. (I don't like clear plastic fins.)

Has anyone tried using a gyroscope (like the ones they use in RC helicopters) and hobby servos to control a gimballed motor mount? Or gyroscopically controlled vanes to direct the thrust, like in the V2? I'd love to see what projects others have done along these lines before I try one.
 
I seem to recall seeing some posts where someone had done it. Have you tried searching here and at YORF for gimbal control?

Anyway, you run into a number of significant problems with gimbal control on model rockets.

One, the burn time on most motors is too short to make gimbal control a viable solution. There is little time to react before the motor burns out. Control of the fins would be better, but then you come up to, or cross the line, of actively controlling the rocket in flight.

Second. finding inexpensive, lightweight and fast gyroscopes would probably be very difficult.

I am sure there are other problems too.
 
One, the burn time on most motors is too short to make gimbal control a viable solution. .

Its not about it being viable or not, its about the challenge of building a working system. one can choose to use slow motors or make their own?

And yes there is also a build thread on either RocketryPlanet/TRF or TRF Archive but the search engines wont find it for me...
https://www.ukrocketman.com/rocketry/gimbal.shtml <- there is more links at the bottom of that page.
 
Thanks for the links. I might have to check out that book. At first I thought, no, the short burn time has no bearing on anything. Then I realized, yes, it does, because once the motor burns out, there is nothing to stabilize the rocket!
 
Its not about it being viable or not, its about the challenge of building a working system. one can choose to use slow motors or make their own?

Um... viable means something that works.
 
Keep in mind the E6 and F10 that burn for 7-8 seconds and the G12 reload which also burns about 8 seconds.

Those are perfect for a gimbaled engine project.

I did a gimbaled engine project in 1989, using F10 power. The drawback to the system was partly the use of Sunguidance, which often caused liftoff over-reaction. One flight literally took off, veered one way to try to point at the sun, overshot the sun, and tried to correct back the other way, but the ground got in the way before it could complete it&#8217;s attempt to pitch back the other way. If it had used a gyro system or horizon sensor system, so it could have lifted off vertically and tried to keep that vertical flight attitude, it would have been a lot better off.

Two other problems I had. One, the servos turned out to be too slow. They were Canon super-micro servos. If they had been faster servos such as today, it would have worked better.

And the final problem was it had too much movement of the motor mount gimbal. It was over-controlled. It had at least twice as much angle movement than it should have. But the project ended before I tried to make adjustments, as again the sunguidance system just made the whole thing too flaky.

One issue that I considered was that as the rocket gets moving faster, the stability provided by the fins would cause the gimbaled thrust to have less and less of an effect, possibly to the point the rocket would go off course. The ideal way to solve that would be to have the guidance system set in a manner to provide more and more corrective angle (as needed) to the gimbaled mount as the model flew faster (probably just by programming it to do so over time, though using an air data probe for real-world feedback of airspeed would have been interesting. Or, making use of acceleration sensors in the pitch and yaw axes).

I tested out a far cruder method. I arranged for some fins to be glued to the engine mount itself, so the moving mount also produced aerodynamic control that became more effective with more airspeed, to balance out the reduced effectiveness of the thrust vectoring. An idea sort of &#8220;borrowed&#8221; from VonBraun, the Redstone Missile had air rudders that were driven by the same mechanisms that controlled the steering vanes in the rocket exhaust.

BTW - it is not a good idea to try to force an unstable rocket to fly straight by using vectored thrust. A model has such a small moment of inertia that it can respond way too fast, compared to the real things. Think for example of how easy it can be to balance a broom in the palm of your hand. Then think of trying to do that using a pencil. Then think of trying to do that with a toothpick. Our models are like trying to balance a toothpick in your palm compared to a broom stick (real launch vehicles), as far as moments of inertia are concerned. I won&#8217;t say &#8220;cannot be done&#8221;, but I am saying it is a very very difficult task that requires great precision and design to achieve compared to &#8220;just&#8221; trying to make a stable rocket fly slowly on gimbaled thrust. So a better goal would be to do it first with a stable finned rocket, work out the bugs, then maybe move on to finless.

John Pursley worked out a great merger of model plane gyros and model plane horizon sensors, for gimbaled engine scale models. He arranged it so that gyros kept it straight at lift-off and then at maybe 1 second in to the flight the horizon sensors took over (or took priority). The horizon sensors do not work too well for the first 50-100 feet.

I have to cut this short, too much other stuff I need to get to. At any rate, look at the F10 expendable and the G12 (32mm &#8220;R/C&#8221; reload).

- George Gassaway
 
Last edited:
The Apogee F10 is a nice motor for gimballed rockets. It has a 7-second burn time, which allows attitude control for the whole 7 seconds. In the Quad Pod II the motor is mounted above the center of mass. This makes the motor control easier -- it just has to keep the motor mount vertical. I did have to add some damping to the control loop to reduce oscillations.

I have thought about moving the motor to the bottom below the center of mass. In theory, I should just have to flip the sign bit in the control loop and it should work just as well.

George is correct about the speed of the servos. I use fast micro-servos in the Quad Pod II. A rocket that is only a few feet long has a very short moment arm, and a very short time constant.

I think a really nice gimballed rocket project would be a rocket that takes off with a very low acceleration -- somewhere around 1 G. After a few seconds it would it would travel fast enough for the fins to keep the rocket stable. It wouldn't even need a launch rod. However, I would use caution in launching a rocket like this, and make sure it is located far away from spectators.

Dave
 
The Apogee F10 is a nice motor for gimballed rockets. It has a 7-second burn time, which allows attitude control for the whole 7 seconds. In the Quad Pod II the motor is mounted above the center of mass. This makes the motor control easier -- it just has to keep the motor mount vertical. I did have to add some damping to the control loop to reduce oscillations.

I have thought about moving the motor to the bottom below the center of mass. In theory, I should just have to flip the sign bit in the control loop and it should work just as well.

George is correct about the speed of the servos. I use fast micro-servos in the Quad Pod II. A rocket that is only a few feet long has a very short moment arm, and a very short time constant.

I think a really nice gimballed rocket project would be a rocket that takes off with a very low acceleration -- somewhere around 1 G. After a few seconds it would it would travel fast enough for the fins to keep the rocket stable. It wouldn't even need a launch rod. However, I would use caution in launching a rocket like this, and make sure it is located far away from spectators.

Dave

I think you need to build it and bring to an Alamo Rocketeers launch!
 
I think you need to build it and bring to an Alamo Rocketeers launch!
John,

I was thinking the same thing. I just ran a simulation of an 800 gram rocket on an F10, and it has an interesting flight profile. It would accelerate to 36 MPH during the first 1.5 seconds, and then maintain that speed for the rest of the burn. It would reach an apogee of 333 feet after about 8 seconds.

The Quad Pod II motor mount platform weighs 430 grams, which would leave 370 grams for the rest of the rocket. I'll have to see if I can get some thin-walled 4" tubing like Estes used to use. I don't know if I can get it together by the launch on the 15th, but I'll give it a try.

Dave
 
John,

I was thinking the same thing. I just ran a simulation of an 800 gram rocket on an F10, and it has an interesting flight profile. It would accelerate to 36 MPH during the first 1.5 seconds, and then maintain that speed for the rest of the burn. It would reach an apogee of 333 feet after about 8 seconds.

The Quad Pod II motor mount platform weighs 430 grams, which would leave 370 grams for the rest of the rocket. I'll have to see if I can get some thin-walled 4" tubing like Estes used to use. I don't know if I can get it together by the launch on the 15th, but I'll give it a try.

Dave

Dave,

Take your time and do it right. Our launch is this Saturday, the 2nd Saturday of the month.

We'd still like to have you there with whatever else you have up your sleeve and we'll treat it as a preview for Sept!
 
Dave,

Take your time and do it right. Our launch is this Saturday, the 2nd Saturday of the month.

We'd still like to have you there with whatever else you have up your sleeve and we'll treat it as a preview for Sept!

I guess I was thinking of the date of the AARG launch. Yes, the September Alamo launch date will give me a littler more time to work on it.

Dave
 
I guess I was thinking of the date of the AARG launch. Yes, the September Alamo launch date will give me a littler more time to work on it.

Dave

That sounds good. You can still show up this Saturday, though. We've gotten rid of the carnivorous cows.
 
Control of the fins would be better, but then you come up to, or cross the line, of actively controlling the rocket in flight.

What is this "line" you refer to? Is there some law actually prohibiting active guidance of a model rocket during flight? If so, do you have a link to it?

I cannot see any problem with active guidance as far as the NAR/TRA safety codes go, for either LPR or HPR rockets. We have prohibitions against launching at a target, or launching explosive warheads, but nothing regarding active guidance.
 
the biggest hurdle is a servo(s) that would react quickly enough to guide a rocket to that degree of accuracy.

unless you have a military budget and a big rocket with serious burn time ,,the benefits of a guided model rocket are seriously minimal..

it has been done but ,,actually hitting a target is just as viable with a non guided rocket.

it boils down to a servo system that is lightening fast (or faster)
 
I'm just about to start such a project and have done EXTENSIVE reading.

A few points to consider - You can *probably* use the ardupilot software. https://diydrones.com/profiles/blog/show?id=705844:BlogPost:44814

It has a 'wings level mode' which is designed to keep an airplane straight an level (pitch and roll axis)

This is all you need for a rocket although worth note is that this will not stop the rocket from rotating, just make it fly straight up.

Ardupilot works with thermopile sensors. A quick write up on how those works is here -> https://paparazzi.enac.fr/wiki/Theory_of_Operation#Infrared_Sensors

To make up for the somewhat slower thermopiles you'll need fast servos. Keep in mind thermopiles are not subject to gyro drift which is a really good thing.

https://www.mks-servo.com.tw/servo-05.htm

Finally, although I have not done all the complicated math you can improve stability by moving most of your mass to the top of the rocket.

I have bought the thermopile sensors but not yet bought the other hardware.

If this is a project you'd like to collaborate on that would be awesome.

I'll send you a PM with my email.
 
Hi folks. I know that in NASA they use gimballed motor mounts, controlled by gyroscopes, to direct the thrust of the engine, for stability and steering.

For model rockets, it seems to work OK just to use fins, but for many scale models this is not very realistic. (I don't like clear plastic fins.)

Has anyone tried using a gyroscope (like the ones they use in RC helicopters) and hobby servos to control a gimballed motor mount? Or gyroscopically controlled vanes to direct the thrust, like in the V2? I'd love to see what projects others have done along these lines before I try one.


Yes, John Pursley (former editor of American Spacemodelling-- the magazine now known as Sport Rocketry) did this and showed us his setup. He has a Vanguard rocket in about a 4 inch diameter and also did a superdetailed Mercury Redstone that he flew at NARAM years ago using a gimballed motor mount and a horizon sensor system (a modified model airplane system). He told some of the club guys and I all about it last year at a get-together we had at his house.

I asked much the same question you did-- what about gyroscopic stabilization, or using off-the-shelf RC heli gyros for inputs to the servos to control the motor gimbal. Problem is, gyros 'drift' and lose their bearings, so the rocket gradually drifts more and more off course throughout the flight, and the more time that passes from the gyro system being 'locked and started up' until actual flight the worse the problem gets.

The system he used was a slightly modified model airplane 'autopilot' system which was basically off-the-shelf. The four horizon sensors were placed in the nose cone on the Vanguard, looking out 90 degrees apart around the rocket toward the horizon. These sensors are CO2 sensors that 'see' the horizon by the opacity of the atmosphere because of CO2 all the way through the atmosphere above the horizon, compared with the limited amount of CO2 between the sensor and the ground nearby (and out to the horizon below the ground line). The sensors feed this data to the autocontroller, which then signals the gimbal sensors to feed in corrective thrust to keep the rocket flying vertically. If the sensors detect a yaw motion, say the rocket tipping at an angle to the horizon, the sensors on one side of the rocket will see the horizon line 'drop' as the angle increases and will see more air (white), where the opposite sensor 180 degrees on the other side of the rocket will see a lot more ground (black) and the controller will gimbal the motor until the rocket turns enough so that both sensors are seeing equal amounts of air and ground (black and white) again. If the rocket tips the other way, it feeds opposite control to the servos and gimbals the motor for correction until the sensors are balanced again. Same thing in pitch attitude. I asked about roll control, but he said it never was a problem and his system isn't set up to address roll control.

On his Mercury Redstone, the rocket placed the sensors under the capsule, the simplify recovery and wiring of the sensors. It worked on the same principle.

If you want more information, you can contact him directly at his website: https://www.accur8.com/

Good luck and hope this helps! OL JR :)
 
...
I asked much the same question you did-- what about gyroscopic stabilization, or using off-the-shelf RC heli gyros for inputs to the servos to control the motor gimbal. Problem is, gyros 'drift' and lose their bearings, so the rocket gradually drifts more and more off course throughout the flight, and the more time that passes from the gyro system being 'locked and started up' until actual flight the worse the problem gets.
...
Gyro chips can easily handle the duration of a rocket launch. The trick is to not start the integration until the moment the launch occurs. I launched the Quad Pod II at LDRS last year, and it sat on the pad for at least 15 minutes before it was launched. I couldn't use a G-switch to detect launch since the accelleration is fairly low on the Quad Pod. Instead, I used a magnetic relay switch like the ones used to detect open windows and doors in home security systems.

While setting on the pad, the Quad Pod would monitor the average values coming out of the gyro chips. Since the Quad Pod wasn't moving, these values represented zero angular velocity. The moment the magnetic relay opened the Quad Pod would use the most recent average values from the gyros as the DC offest values.

Dave
 
Hi to all from Italy. I have Casimiro, the first model rocket with active stabilization ever built in Italy. It flew quite perfectly, after a year of development (and after two years and a half from the start of the project) in august 2005. Casimiro has a perfect cardanic suspension (two axis) for a little 24 mm. motor, obviously an Estes E9 or an AT F12 RMS. Two gyros, a PIC controller, an home-grown inertial sensor for deploy are the minimum of electronics needed. I am currently working onto two other little rockets, one (Gyro Gearloose) with a simpler gimbal and another which employes four little long burning engines (SF Held 1000) as "verniers".

My Casimiro went on Sport Rocketry two years ago.

Please visit my website www.criscaso.com

ciao
Cristiano
 
Back
Top