Gimbal mount

[dcshrum]

I am thinking about building a gimbal mount for a motor. I've got a pretty good background with computers and electronics and I'm anticipating this will be a fun project.

The intent is to produce a device that will result in the rocket flying perpendicular to the horizon. That is, its not really a guided rocket, its just going to fly straight up.

I'm planning on building a circuit board that uses thermopile sensors for horizon detection. You can read up on those here: https://paparazzi.enac.fr/wiki/Infrared_Sensors

Once the circuit board is built I would attach the servos to the motor mount.

I am wondering if it would be necessary to move the gimbal in proportion to the heat difference between the two sensors or if it would be sufficient to move the servo and gimbal a fixed amount in the case that the difference between the two sensors exceeds some percentage.

Thoughts?

 

[JAL3]

Go for it!

One of the coolest rockets I have seen is Dave Hein's Quad Pod.

The video below was its maiden flight on an E9 but he has since upgraded to to the long burning Apogees.


[YOUTUBE]<object width="425" height="344"><param name="movie" value="https://www.youtube.com/v/wUJx7LbRZn0&hl=en&fs=1&"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="https://www.youtube.com/v/wUJx7LbRZn0&hl=en&fs=1&" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" width="425" height="344"></embed></object>[/YOUTUBE]

 

[brianc]

The May/June 2007 issue of Sport Rocketry had an article by
Christiano Casonati...

Being from Milan Italy his web site is in Italian, but here's the
link- https://www.criscaso.com/ModSpaz/Criscaso su SR.htm

(scroll to near the bottom)

 

[dcshrum]

But the article is in English! Thanks for the link

 

[dave carver]

There's an artical in High Power Rocketry on a gimbaled motor with pictures. I can't tell you what issue it was in but it was early on in HPR's printing.

 

[DaveHein]

dcshrum,

It's hard to say how fast the thermalpiles will react, or how large the difference signal will be. It seems like you'll have to buy a pair of them and experiment with them. I would guess that there will be a lot of variation in performance depending on cloud cover and ground cover. It seems like the sun would also swamp any sensor that is pointed at it. However, it would be interesting to check it out.

I used a 2D gyro chip to control the motor direction on the Quad Pod II. The motor is mounted at the top of the rocket, so for the most part I just maintained the motor mount at a vertical angle. I did add a small amount of damping to reduce oscillations. In theory, a rear-mounted motor would work the same, but with the sign bits reversed in the control loop.

My webpage on the Quad Pod II is located at
https://home.swbell.net/davehein/ModX/quadpod_ii.html

Dave

 

[dcshrum]

dcshrum,

It's hard to say how fast the thermalpiles will react, or how large the difference signal will be. It seems like you'll have to buy a pair of them and experiment with them. I would guess that there will be a lot of variation in performance depending on cloud cover and ground cover. It seems like the sun would also swamp any sensor that is pointed at it. However, it would be interesting to check it out.

I used a 2D gyro chip to control the motor direction on the Quad Pod II. The motor is mounted at the top of the rocket, so for the most part I just maintained the motor mount at a vertical angle. I did add a small amount of damping to reduce oscillations. In theory, a rear-mounted motor would work the same, but with the sign bits reversed in the control loop.

My webpage on the Quad Pod II is located at
https://home.swbell.net/davehein/ModX/quadpod_ii.html

Dave

From the reading I have done -> here is a nice short paper https://www.ctie.monash.edu.au/hargrave/horizon_sensing_autopilot.pdf

Theropiles will sense faster, are not prone to dirft and work just dandy in the clouds. Its an infrared signal after all. I'm looking at your page now which I expect to be quite helpful.

 

[DaveHein]

Thanks for the link to the PDF. It is very informative. It appears that the sensor has a silcon filter to screen out the higher optical frequencies. The sun must appear dark in the very low frequencies that the thermopile is sensitive to.

The spec sheet for the part does say that it has a time constant of 30 msec, if I recall correctly. That seems surprising to me since it is essentially a thermal sensor that is heated by infra-red light.

I am looking into a method for stablizing and pointing a high-altitude balloon payload. It looks like this sensor along with a sun detector might be just the thing for that application.

Dave

 

[Monroe]

The "Quad Pod II" works very nicely you have to see it in action up close and watch the motor gimbal! We are working on the "Quad Pod III" with Dave In fact I have some extra long burn High H to low I motors I am about to cast for it. Dave made a great Load Cell we are going to test at the same time. Dave "you should do a write up on that one" I'm a member over at DIY Drone's thats another good site for an automatic pilot system that may have some use. I Dl'ed the paparazzi software about a year ago and that has interesting possiability's TOO!

Monroe
Team Prometheus

 

[dcshrum]

I am looking into a method for stablizing and pointing a high-altitude balloon payload. It looks like this sensor along with a sun detector might be just the thing for that application.

Dave

I've done that too

https://icbnn.wordpress.com/2009/01/02/launch/

This picture from the flight is just cool: https://francescobonomi.it/icbnn/res3bis_04_crop.jpg

 

[MarkII]

The "Quad Pod II" works very nicely you have to see it in action up close and watch the motor gimbal!

In the YouTube vid that JAL3 showed, if you watch closely you can see some gimballing of the motor during the flight.

MarkII

 

[dcshrum]

Thermopiles have a 20-30ms delay in reacting to changed radiation.

I might purchase this board -- https://chebuzz.com/paparazzi/index.php?main_page=product_info&cPath=2&products_id=1

The thermopiles are connected head-to-head on the board so that they give a zero voltage if they see the same IR radiation or a positive/negative if different.

I'm thinking 4 simple circuits that detect a change in voltage and actuate the 2 servos in either direction. I don't know what the response time would be for the voltage change to be detected, the servo motor to actuate and the motor to move.

Do you have any idea what kind of response time your quad pod was working with or did you simply build it and hope it was fast enough?

Feel free to send me a PM if you'd like to exchange emails.

 

[DaveHein]

I tried to minimize the response time in my control loop as much as possible. I used servos that can rotate 60 degrees in 0.1 seconds, and I sampled the gyros at a rate of 100 times/second. I ran some simulations where I introduced delay in the control loop and I ran into oscillation problems at about 30 or 40 msecs.

The time constant of a rocket is determined by the moment of inertia of the rocket and the amount of thrust. The Quad Pod has a moment arm of about 30 cm, and a mass of about 1 Kg, so the moment of inertia is around 0.1 Kg*m*m. If your rocket is much longer, and most of the weight is in the nosecone and motor mount (i.e., at the ends) then you could have a longer moment of inertia. As an exmple, if the rocket has a mass of 1Kg and is 2 meters long, and all of the weight is at the ends then it would have a moment of inertia of 1 Kg*m*m. With the same motor as the Quad Pod you would have a time constant that is 10 times longer than the Quad Pod's.

The Saturn V was roughly 300 feet long. As a very rough approximation, it would have a moment arm of about 25 meters. The moment of inertia is proportional to the square of the moment arm, so the Saturn V's time constant would be more than 6,000 times that of the Quad Pod. This is why slow moving hydraulic systems can be used to position the engines of large rockets.

Dave

 

[dcshrum]

I follow exactly what you are saying although I don't understand the exact definition of "time constant of a rocket"

I did just get back from lunch with my co conspirator on this project and we discussed that so long as we kept our weight towards the nose of the rocket we could have a lower sampling rate and reduced oscillations as the rocket attempted to track straight.

I just ordered a rocket book or two so perhaps I'll find "time constant of a rocket" covered in there somewhere.

I was going to ask how much torque you needed to but I found your web page shows you used the futaba S3114.

Did you gimbal the motor a fixed number of degrees from horizontal or did you vary the degree of gimbal based on how far away from level the rocket was?

Hopefully that question made sense

 

[bobkrech]

Dave

Stabalized rocket have been around for a while. The most common sensor is a store bought PA-2 Futaba horizon sensor. It rather plug and play using 4 photodiodes as horizon sensors. It's $59 for Tower Hobbies.

George Gassaway and John Pursley have both built stabalized rockets: George used sun-stabalization and John used the Futaba horizon sensor. A few google searches should find the information. Pursley wrote an article in Sport Rocketry in the early 2000's.

Folks have used a aft end gimbel and a pair of forward mounted microservos to push the top of the motor in the x-y plane. For sun seeking rockets, a simple quadrant diode alligned to the 2 servo axis works well and for the Futaba you simply have to align the photodiodes with the two-axis. Summing the opposite sensor will give a +/- error voltage that can be shifted and scaled to get the servos to push left or right.

Bob