Selection of the appropriate fin and control system to stabilize a rocket traveling at 2 mach

[alvise]

I've read the forum discussions but didn't get a clear answer.
So I have a few questions for rocket enthusiasts.
-How can a suitable fin be selected for this speed?
-What is the benefit of using the gyroscop information of the IMU to stabilize the rocket with fins. Can it be used to direct the motor directly to the fin motors by using the angles obtained directly from the IMU?
- Should a wind tunnel operating at 2 mach speed be used to understand whether the program created for roll stabilization using fins in a rocket traveling at 2 mach speed gives the desired result or not?

Thanks in advance to everyone who shared their thoughts and experiences on this matter.

 

[JimJarvis50]

I've read the forum discussions but didn't get a clear answer.
So I have a few questions for rocket enthusiasts.
-How can a suitable fin be selected for this speed?
-What is the benefit of using the gyroscop information of the IMU to stabilize the rocket with fins. Can it be used to direct the motor directly to the fin motors by using the angles obtained directly from the IMU?
- Should a wind tunnel operating at 2 mach speed be used to understand whether the program created for roll stabilization using fins in a rocket traveling at 2 mach speed gives the desired result or not?

Thanks in advance to everyone who shared their thoughts and experiences on this matter.

I have a system that I use for roll and yaw/pitch stabilization. It has four canards, which are driven directly by RC-type servos. It works quite well, although it has not been flown yet to Mach 2. Regarding suitable fins, if you're asking about the canards, our strategy is to reduce the size of the canards as the speed increases. For Mach 2, we estimate canards with a size on the order of a postage stamp (an area per canard face around 0.5 to 0.75 sq. in.

Using gyroscope information can help to make the rocket go "up" and also to avoid coning (if the roll rate would otherwise increase above a certain rate). However, the information needed for this type of control does not come "directly" from the IMU. Instead, there are calculations that track the orientation of the rocket, starting from launch, using the gyros only. This is not trivial.

I'm not sure if wind tunnel tests at Mach 2 are feasible. It would be interesting though. One issue to deal with is called "control reversal", where the rocket can turn in the opposite of the intended direction due to vorticies produced by the canards interacting with the main fins. Wind tunnels, and CFD, have been used to investigate this. In my system, I use a spin can to avoid this interaction.

You can check on the forum for the thead "I could use just a little guidance", or find YouTube videos searching on Jiminaus50.

Jim

 

[alvise]

I have a system that I use for roll and yaw/pitch stabilization. It has four canards, which are driven directly by RC-type servos. It works quite well, although it has not been flown yet to Mach 2. Regarding suitable fins, if you're asking about the canards, our strategy is to reduce the size of the canards as the speed increases. For Mach 2, we estimate canards with a size on the order of a postage stamp (an area per canard face around 0.5 to 0.75 sq. in.

Using gyroscope information can help to make the rocket go "up" and also to avoid coning (if the roll rate would otherwise increase above a certain rate). However, the information needed for this type of control does not come "directly" from the IMU. Instead, there are calculations that track the orientation of the rocket, starting from launch, using the gyros only. This is not trivial.

I'm not sure if wind tunnel tests at Mach 2 are feasible. It would be interesting though. One issue to deal with is called "control reversal", where the rocket can turn in the opposite of the intended direction due to vorticies produced by the canards interacting with the main fins. Wind tunnels, and CFD, have been used to investigate this. In my system, I use a spin can to avoid this interaction.

You can check on the forum for the thead "I could use just a little guidance", or find YouTube videos searching on Jiminaus50.

Jim

Thank you for your answer Jim;
I have read the topic (about 30 pages) you have opened before and I am still following it.
-For now I just want to do some tests to prevent the rocket from spinning around itself.
I plan to use only 2 canards near the nose of the rocket.
My goal is for the rocket to maintain its initial roll angle (on the pad), so not just to prevent the rocket from spinning around in any position. For example, this will not be the desired result if the rocket comes out of the pad and after it spins around a bit and stabilizes the angle at which it rotates.

How can I determine the correct choice of canards for a rocket going 2 mach. It is really difficult to guess how the canards selected with the created program control the rocket. In the test made on the table, the wings move as much as the rotation of the rocket around itself (wing position = desired angle-current angle). Actual requirement may be different.

-I can read raw acc, gyro, magnetometer information as well as roll, pitch, yaw information from the IMU I use.
I send the angles I read from the IMU to the motors (by mapping the 0-90 degrees of the IMU to the 0-20 degree angle of the motor).
Do you mind if I want to use any of these angles read from the IMU to prevent the rocket from spinning?
Most forums mention gyroscope, is there any particular reason why gyroscope was chosen?

Isn't there a way to do some testing without such large wind tunnels?
I'm a little cautious. I don't want to run tests without getting advice from experienced rocketeers.

 

[OverTheTop]

Search for "vertical trajectory system" to see what I am doing in this space.

 

[alvise]

Search for "vertical trajectory system" to see what I am doing in this space.

Can you give a link because. I'm doing research as a "vertical orbital system". This site gives the error that the links cannot be found. If there is a VTS topic, can you provide a link?

 

[OverTheTop]

Can you give a link because. I'm doing research as a "vertical orbital system". This site gives the error that the links cannot be found. If there is a VTS topic, can you provide a link?

Links on the Aus Rocketry website are broken currently. If you search for "vertical trajectory system" or "VTS" on this site there are some comments and images. Not everything, but it is the best you will get currently with ausrocketry site being down.

 

[JimJarvis50]

Thank you for your answer Jim;
I have read the topic (about 30 pages) you have opened before and I am still following it.
-For now I just want to do some tests to prevent the rocket from spinning around itself.
I plan to use only 2 canards near the nose of the rocket.
My goal is for the rocket to maintain its initial roll angle (on the pad), so not just to prevent the rocket from spinning around in any position. For example, this will not be the desired result if the rocket comes out of the pad and after it spins around a bit and stabilizes the angle at which it rotates.

How can I determine the correct choice of canards for a rocket going 2 mach. It is really difficult to guess how the canards selected with the created program control the rocket. In the test made on the table, the wings move as much as the rotation of the rocket around itself (wing position = desired angle-current angle). Actual requirement may be different.

-I can read raw acc, gyro, magnetometer information as well as roll, pitch, yaw information from the IMU I use.
I send the angles I read from the IMU to the motors (by mapping the 0-90 degrees of the IMU to the 0-20 degree angle of the motor).
Do you mind if I want to use any of these angles read from the IMU to prevent the rocket from spinning?
Most forums mention gyroscope, is there any particular reason why gyroscope was chosen?

Isn't there a way to do some testing without such large wind tunnels?
I'm a little cautious. I don't want to run tests without getting advice from experienced rocketeers.

Two canards is fine. If you are doing only roll control and not yaw/pitch, canards near the CG might be better. That would produce less torque on the air frame and might help to avoid control reversal by having the canards closer to the fixed fins. I would arrange for the canards to be offset from the fixed fins. I haven't studied this though, or done any modeling.

If your IMU produces a roll angle, then you can use that. Mine doesn't, so gyro readings are integrated externally to produce the roll angle.

My system has two forms of roll control, and I think both are important. I'm currently trying to understand how to make them work together. One control is on the roll angle itself. We refer to this as heading hold. In the simplest sense, it takes the roll angle error and calculates a proportional canard angle to return the roll angle to the starting point. We are using a gain such that an error up to 360° produces a canard deflection of up to 7.5°. We also control the roll rate. The specific canard angle is calculated as roll rate divided by 1000°/s times 7.5°. The two approaches work together to control roll. If the rate rate term is too low, the heading hold control will cause the rocket to ocscilate around the desired roll angle. On the other hand, if the rate term is too high, the return to the zero point is too slow. So, there is an optimum, which I am currently testing for. I think the optimum gains might be 360 and 1600. I think the deflections that result from the above are much lower than what you are considering? I also think that if you want to control roll position near launch and at Mach 2, you will need gain scheduling.

I typically fly with a spin can. That means the response of the rocket body to canard changes is very fast (lower moment of inertia and no fins to dampen the roll rate). You can easily calculate these forces and acceleration rates, and for my system, the potential motion of the rocket is fast in comparison to the times associated with control updates and servo response. My system operates at 40Hz with RC-type servos, and I'm sure better equipment would be helpful. I suspect I would get better performance, to a point, if I reduced both of the above gains and/or the canard size. With fixed fins, the effectiveness of reducing the gains is more limited because there needs to be enough control authority to overcome any misalignment of the fins. I think very straight fins would be essential. I haven't thought as much about roll control with fixed fins, but I'm hoping for some data over the next year.

I'm not sure what to say about testing without a wind tunnel. I can say that quite a bit can be learned just doing the calculations associated with the system. But, I can see that solving the problem will require a careful balance of many variables.

Just some thoughts...

Jim

 

[alvise]

Thanks for the explanation.Every idea given, even if it's just an opinion, is important...

To determine the size of the fins to be used close to the CG, I think the factor to consider is weight, not just speed.
-I think by roll rate you mean the rate of rotation of the rocket around itself. To determine this, I guess at the beginning I need to do a wingless measurement to test the spin speed, right? Or create an algorithm that adjusts the PID according to the spin rate. I'm not exactly sure.Built-in PID values can be adjusted in this servo motor.
-The servos I intend to use (mx-106T) are expensive, but I trust the recovery system. They are sensitive enough, but there is only a backlash problem, I don't know how this will affect it. The idea of using small fins at the beginning is reasonable. The motor has an overload measure on itself and I am thinking of placing a mechanical limitation on the fins as an extra. On the other hand, I think I can gather a lot of information to analyze with these engines.
-I'm thinking of using four fixed fins at the back and 2 fins in the middle close to the CG.
-It seems to me that spin can can't maintain a clear angle and angle control seems more difficult.

Many factors make it really difficult to calculate.

 

[OverTheTop]

You can find the rotational moments of inertia of your rocket using Rocksim or OpenRocket. You can calculate the lift provided by canard angles, and thus calculate control moments that will correct the flight. Note that the moment of inertia about the central axis is really quite small, due to all the mass being close to the axis. In my case I think it was about 150 times smaller than the inertias about the long axis. Even taking into account the lever arms of my canards (closer to the roll, further away on the pitch/yaw) the roll axis was 30 times more sensitive to canard movement.

 

[cls]

Might help frame your thinking to use the right words.

In 3 dimensions, there are 3 translation axes, and 3 rotation axes.

"Spin" doesn't mean much by itself.

Usually we speak of pitch, yaw, and roll, for the rotations.

For a radially symmetrical rocket, pitch and yaw are similar, but orthogonal. Most folks choose a preferred orientation, like along one fin.

Hope that helps.

 

[alvise]

You can find the rotational moments of inertia of your rocket using Rocksim or OpenRocket. You can calculate the lift provided by canard angles, and thus calculate control moments that will correct the flight. Note that the moment of inertia about the central axis is really quite small, due to all the mass being close to the axis. In my case I think it was about 150 times smaller than the inertias about the long axis. Even taking into account the lever arms of my canards (closer to the roll, further away on the pitch/yaw) the roll axis was 30 times more sensitive to canard movement.

I can say that I have a general idea. Motors used for roll rotation control always require less torque (according to pitch and yaw movement).
I think the point I should pay attention to is PID. Because when testing with IMU on the ground using only P, everything seems to be working fine. I or D seem to have no effect. Another thing is to adjust PID according to increasing and decreasing speed. The rocket is probably maxed out in 1.6 seconds. it will pick up speed and then it will start to lose speed continuously but still climbing. During this time the torque required to control the roll rotation will be different. I don't know how to set the appropriate PID for this.

Doubt and questions.

For a radially symmetrical rocket, pitch and yaw are similar, but orthogonal. Most folks choose a preferred orientation, like along one fin.

I think you're talking about yaw/heading, in general here.

 

[OverTheTop]

You can simplify your control system by using only PI. Forget the D to start with. P will get you down to a very low roll rate. Adding the I term will allow that roll to be stopped completely. That is what I is about, cancelling any steady-state error.

Yes, to do things properly and get the most out of the system there needs to be gain scheduling, according to speed, as the fins have more lift at higher velocities.

Start simple. Maybe P and a little I. Set gains low enough that the rocket doesn't oscillate a peak velocity, or maybe just a little. That will give you a good start to tune the loops.

 

[alvise]

You can simplify your control system by using only PI. Forget the D to start with. P will get you down to a very low roll rate. Adding the I term will allow that roll to be stopped completely. That is what I is about, cancelling any steady-state error.

Yes, to do things properly and get the most out of the system there needs to be gain scheduling, according to speed, as the fins have more lift at higher velocities.

Start simple. Maybe P and a little I. Set gains low enough that the rocket doesn't oscillate a peak velocity, or maybe just a little. That will give you a good start to tune the loops.

-The servos I use for roll stabilization are positioning servos. It can go to the specified angles (even if the P value is 20 I and D 0) without oscillation. But what is the equivalent of this in rocket stabilization?
-In my current setting, when the rotation of the rocket is detected and the degree of rotation is detected, the more the fin moves in scaled degrees.
How much can this stabilize the rocket. What kind of problems can occur?

georgegassaway say here.
I know many in this thread know this, but I’ll mention that aerodynamic force is squared by the increase in velocity. So let’s say the roll control system is tuned to work perfectly at 100 mph, no overshooting. Now let’s say the rocket will have a maximum velocity of 600 mph (I choose 600 to avoid messiness with Transonic issues). The aerodynamic forces are not 6 times higher than at 100 mph. They are 36 times higher than at 100 mph ( Velocity squared, so 6x6=36). So if the roll control response was ideal at 100 mph, then the servos need move 1/36th as much at 600 mph than at 100 mph.

Of course I’ve just picked a convenient number by saying 100 mph, let’s say the roll control was tuned to work ideally at 200 mph, in which case for 600 mph the servos should move 1/6th as much (or for 800 mph, ignoring sonic issues, 1/16th as much).

The thing that puzzles me is how to determine the appropriate PID values without doing a flight test.
If it is approached with the said logic, if a PID is determined by making a calculation in this way, what is the probability of responding at a speed like 2 mach.


I'm extremely skeptical about this.

 

[cls]

I'm extremely skeptical about this

Yes, you should be sceptical about the entire idea.

Go do some research, figure out how much power is required to operate aircraft control surfaces at high subsonic speeds. Look at the slew rate of the controls - both linear, and rotary servos. How are they powered? Model aircraft servos are designed for 100 MPH, some for 200 mph. Are they really going to work at 600 mph? 1200 mph? 1800 mph?

 

[alvise]

Yes, you should be sceptical about the entire idea.

Go do some research, figure out how much power is required to operate aircraft control surfaces at high subsonic speeds. Look at the slew rate of the controls - both linear, and rotary servos. How are they powered? Model aircraft servos are designed for 100 MPH, some for 200 mph. Are they really going to work at 600 mph? 1200 mph? 1800 mph?

Thanks for your suggestion. I've been doing research and still trying to run some tests. By the way, the servos I will use are not the ones used in normal RC airplanes. I mentioned it in my previous posts.
My purpose here is to test the reliability of the P I values for roll stabilization.
-I wanted to do some tests in OpenRocket, but I got some results as follows. My question is, is the pitch rate information here really correct, does the rocket swing that much in pitch angles while climbing? So is this normal?


-For testing purposes, I put 2 fins on the nose of the rocket with an angle (fin cant 10 degrees) facing each other in the same direction. But this does not seem to have an effect on the rocket because it does not change the distance or direction of the rocket in any way. Do you think this is normal?

Anyone have any comments on this?

Best regards

 

[alvise]

Mr.Jim although I have read your topic from beginning to end, I could not clearly understand how the wing servos are programmed for active stabilization with 4 canards.

When the rocket is seated on the pad, all 4 canards are at 45 degree angles with respect to the pad.
How exactly should the servos be programmed to control the roll, pitch, yaw angles when using 4 canards. For example, if two opposing motors (motor 1-2) are selected to control the roll angle of the rocket, and if these two motors are determined to control the roll angle, they will be reversed. 2 motors (motor 3-4) remain, if these motors are adjusted independently of roll stabilization to control pitch and yaw, they cannot perform active stabilization of the rocket.

Can you explain what kind of confugration you used for this situation?It might be really helpful if you could expand on this topic a bit.

Best regards

 

[OverTheTop]

How exactly should the servos be programmed to control the roll, pitch, yaw angles when using 4 canards. For example, if two opposing motors (motor 1-2) are selected to control the roll angle of the rocket, and if these two motors are determined to control the roll angle, they will be reversed. 2 motors (motor 3-4) remain, if these motors are adjusted independently of roll stabilization to control pitch and yaw, they cannot perform active stabilization of the rocket.

Yes, for roll control opposite fins need to operate in opposite directions. For pitch and yaw the opposite fin needs to work in the same direction (if looking at the front of the rocket anyway). To control both roll and pitch (or roll and yaw) the relevant fins need to move to the algebraic sum of the roll and (pitch or yaw) control loop signals.

 

[alvise]

Yes, for roll control opposite fins need to operate in opposite directions. For pitch and yaw the opposite fin needs to work in the same direction (if looking at the front of the rocket anyway). To control both roll and pitch (or roll and yaw) the relevant fins need to move to the algebraic sum of the roll and (pitch or yaw) control loop signals.

thansk for answer ;

I give an example from the picture. Let's say we separate these two motors for roll control in order for Motor 1 and Motor 2 to move in opposite directions to stabilize the roll.
So while motors 1 and 2 are trying to control yaw, which motors should move if the angle of inclination is to be controlled?
Will only 3 and 4 motors move? Or will they be positioned relative to the pitch movement on the 1 and 2 motors?
In V-tail airplanes, two ailerons can control both pitch and yaw. But here as an extra to the V-tail aileron logic.
There are 2 more fins at the bottom.

Is there a resource that explains a little about this?

 

[alvise]

Mr. Jim I want to ask you something(Actually, I'm asking this to everyone).
What is the stability caliber in your model?
I guess they are very close to each other.
If the caliber is too large, it will require more force to control, which means larger fins.
Unless, of course, I'm wrong in this equation.

 

[alvise]

Hello everyone after a long time.
I decided to do some simulation tests before doing some real ones.
I decided that the X-plane is the best choice for this. I can design any rocket and simulate it with an X-plane. Control canards and the like can be added and controlled. So I can test an algorithm that will generate through this simulation.
But I'm still chasing how to properly build a model.
Has anyone dealt with 3D simulation for rocket before?

 

[JimJarvis50]

Mr.Jim although I have read your topic from beginning to end, I could not clearly understand how the wing servos are programmed for active stabilization with 4 canards.

When the rocket is seated on the pad, all 4 canards are at 45 degree angles with respect to the pad.View attachment 581890
How exactly should the servos be programmed to control the roll, pitch, yaw angles when using 4 canards. For example, if two opposing motors (motor 1-2) are selected to control the roll angle of the rocket, and if these two motors are determined to control the roll angle, they will be reversed. 2 motors (motor 3-4) remain, if these motors are adjusted independently of roll stabilization to control pitch and yaw, they cannot perform active stabilization of the rocket.

Can you explain what kind of confugration you used for this situation?It might be really helpful if you could expand on this topic a bit.

Best regards

The orientation of the rocket is determined with respect to yaw and pitch. Two opposing canards deal with yaw and the other two deal with pitch. Obvoiusly, the sets of opposing canards move in opposite directions. Any roll correction is superimposed on all four canards, moving in the same direction. Generally, the four canards will have different deflections.

Jim

 

[JimJarvis50]

Mr. Jim I want to ask you something(Actually, I'm asking this to everyone).
What is the stability caliber in your model?
I guess they are very close to each other.
If the caliber is too large, it will require more force to control, which means larger fins.
Unless, of course, I'm wrong in this equation.

The rocket I fly is typically set up for a stability of 1.3-1.5 calibers, including the effect of the canards on stability. For the canards I use (a root and span of 2" for example), a canard deflection of 6° might result in a turn rate of 25°/s for yaw or pitch. This is much greater than would be required, for example, for vertical control.

Jim

 

[alvise]

Jim thanks for good answer.

There is open-source openrocket where we can test things before we build the rocket, and that's really good. But is there an open source or sold (other than Xplane) application that can be used in flight simulation at supersonic speeds to be able to test rocket's active stabilization and roll stabilization ?

I guess professional hobbyists use this kind of app.
In my case I believe it is necessary to do a good simulation before doing the actual flight test.

 

[alvise]

Hi,everyone.
I would like to ask this question to anyone who has an idea. Before starting the active stabilization tests, what methods can be used to find the torque needed for the servo that will move the canards?
-The simulation tool I am currently using does not seem to be capable of doing this. If anyone knows a different simulation application and method, please let me know. I'm sure it will work for everyone.

 

[OverTheTop]

Mathematically. You know velocity and angle of attack. You just need to work out where the center of lift is. On a regular wing I seem to remember 25% of chord, but you are supersonic which complicates matters. If you use a double-wedge airfoil the center of lift is well known and moves around less and you can work out the forces, turning moments, and torques on your wing, based on that pivot point.

I use a double-wedge airfoil because it is easier to calculate everything. Finding the relevant equations on the www or in books might take a little hunting though.

Limit the angle of throw for the fins mechanically to simplify the maths further, and also make for a safer flight if your control system malfunctions.

 

[Steve Shannon]

Perhaps don’t try to conquer two problems at once. If you haven’t already, I suggest making and flying rockets that exceed M2 with fixed fins. Once you can repeatedly do that, then start working on movable fins.

 

[alvise]

Mathematically. You know velocity and angle of attack. You just need to work out where the center of lift is. On a regular wing I seem to remember 25% of chord, but you are supersonic which complicates matters. If you use a double-wedge airfoil the center of lift is well known and moves around less and you can work out the forces, turning moments, and torques on your wing, based on that pivot point.

I use a double-wedge airfoil because it is easier to calculate everything. Finding the relevant equations on the www or in books might take a little hunting though.

Limit the angle of throw for the fins mechanically to simplify the maths further, and also make for a safer flight if your control system malfunctions.

Yes, I actually created a design. If things go wrong I made a mechanism that mechanically limits the rotation of the canard.

I am planning to use double-wedge airfoil.
It will probably take time to dive into the math and physics books again. Aren't there any little apps that can do the job?

Perhaps don’t try to conquer two problems at once. If you haven’t already, I suggest making and flying rockets that exceed M2 with fixed fins. Once you can repeatedly do that, then start working on movable fins.

I'm simulating fixed blades up to mach 2 and there doesn't seem to be a problem, but I can't determine the appropriate dimensions for moving canards. I don't have a program to simulate this.
In order to start active stabilization with canards, the axial torque created by the dimensions of the canards to be used should not exceed the torque of the servo I have determined.

Not:I'm looking for a way to simulate or calculate accurately to provide an initial idea.

 

[OverTheTop]

It will probably take time to dive into the math and physics books again. Aren't there any little apps that can do the job?

I didn't find any apps or write any programs. Some basic mathematics with initial assumptions to simplify the task were used.

Regarding sizing of the canards: You need to determine the moments of inertia about the three axes. Once you do that you can determine a force to restore them in a reasonable time. Once you have the force needed you can determine the required area, based on maximum velocity. Be aware that the area needed for pitch/yaw is significantly larger than that required for roll. In my system it is a factor of 30 greater, based on the required forces and lever arm lengths.

 

[alvise]

I didn't find any apps or write any programs. Some basic mathematics with initial assumptions to simplify the task were used.

Regarding sizing of the canards: You need to determine the moments of inertia about the three axes. Once you do that you can determine a force to restore them in a reasonable time. Once you have the force needed you can determine the required area, based on maximum velocity. Be aware that the area needed for pitch/yaw is significantly larger than that required for roll. In my system it is a factor of 30 greater, based on the required forces and lever arm lengths.

Can't OpenRocket be used to calculate moments of inertia? Because in Openrocket pitch rate, pitch momentum coefficient etc. It has calculations.
-For example, I created 4 control fins and I am testing the angle of each of them by changing the "Fin cant" value. The pitch rate value varies.
Do you know how better tests can be done for using a method like this?

 

[OverTheTop]

Yes, OR will give you the moments of inertia. You can then make an estimate of how much acceleration you need to move the axis in a reasonable amount of time. You don't want neck-snapping speed and control. If you keep the controls gentle the control loop will be easier to dial in. Remember you need to accelerate the airframe and then decelerate it. Factor that in too.

Once you know desired accel you can work out the lift required by the fins at their distance from the center of the relevant axes.

Remember to use angles in radians .