I could use just a little guidance

This showed up on HackaDay today and might be of interest: https://www.instructables.com/id/Using-the-Intel-Edison-on-a-Cold-Gas-Reaction-Cont/?ALLSTEPS

Control theory is tedious, and long, and borderline torture; however, it is extremely satisfying once you write and implement a working controller.

Click to expand...

UhClem said:

This showed up on HackaDay today and might be of interest

Click to expand...

I could have used a little cold gas today. Did the fifth flight at White Sands. It's a bit warm here. Film at 11:00.

Jim

As advertised, here's a video of the fifth stabilization flight. The objective of this flight was to test the stabilization unit in the two-stage configuration. In this mode, the stabilization section is between the first and second stages and has to function as the interstage coupler. During the initial boost, the canards are near the CG. So, only roll control is effective during this portion of the flight (and remember, the idea is to slow down the roll rate, not necessarily to stop roll completely).

https://youtu.be/umoduWiQb-o

After burnout, the booster drops off and the stabilization spool and sustainer coast together. At this point, the canards are at the bottom of the rocket (instead of at the top as in all previous tests), giving both vertical stabilization and roll control. In the video, you can't see the sustainer after it separates from the booster, but it goes about twice as high as the booster (with the smoke trail). At 14 seconds, the stabilization section falls off. The sustainer motor would be lit at this point, but for this test, there was no sustainer motor present.

The video is from the nose cone, so you can't really see the canards in operation (but they were). I had another camera mounted closer to the canards (with a camera shroud from Landru), but the video file didn't save for some reason.

The flight path was a constant 6 degrees during the stabilization part of the flight (from 4 to 14 seconds). The actual tilt relative to vertical was a bit less, and the rocket moved down wind. There are a lot of factors contributing to this "equilibrium" tilt position, including 4 rail guides and 2 camera mounts hanging off the rocket. Those, plus small alignment errors for the canards, plus a hot day that might have affected the starting condition of the gyros just a bit, contributed to the slightly off-vertical path. With a clean rocket, I think it will go a bit more vertical. Otherwise, the flight was a success, with everything separating and recovering as planned.

Jim

I know this has little to do with the topic at hand but I can't resist asking about it anyways lol...At 1:45 or so in your video I see some straps that appear to be attached to the rail go flying off the rocket. How does that work?

DAllen said:

I know this has little to do with the topic at hand but I can't resist asking about it anyways lol...At 1:45 or so in your video I see some straps that appear to be attached to the rail go flying off the rocket. How does that work?

Click to expand...

That is just a Velcro strap that is glued to the rail and wrapped around the rocket. It just keeps the top of the rocket from flapping around when the rocket is initially placed on the rail and then raised. Apparently, it gets removed rather easily.

Jim

Here's a bit more data from the flight.

The Raven file is attached. The Raven was in the stabilizer spool. The data say it separated from the sustainer at a velocity of 250 ft/s (slower than I had planned due to the high overall drag of the rocket). After it separated, it coasted for about 9 seconds before the chute came out at apogee (I thought it might come to a stop much faster than that). The descent rate was only 17 ft/s. My simulation gave 25 ft/s. Not sure where I messed up there. Good thing there wasn't much wind. I need to downsize the chute a bit.

The tilt graph shows the tilt angles in the yaw and pitch directions plus the net tilt. The graph shows an initial tilt on the pad of 3 degrees. I think the rocket was actually more vertical than that (unless my level is bad), and that the temperature affected the accelerometers. The accelerometer offsets are set on the bench, but they will vary when the temperature changes. According to the Raven, the temperature inside the spool at launch was 109 F. It was a hot day! But now we know the magnitude of that effect (and with any luck, it will be cooler at Balls).

The tilt graph also shows some oscillation of the sustainer section after the booster separated. I don't know what caused this, but I suspect it was the various camera pods and launch lugs attached to the rocket. The rocket at Balls will have none of those things.

The gyro graph is also included. There was a pretty good turn just after burnout. If you watch the video, you can see the booster turn relative to the sustainer in the time between burnout and separation. There isn't a lot of force between the parts at this point, but I believe the relative motion of the parts was caused by the torque from the canards as they resisted the turn.

Jim

Congrats on another great test flight.

I noted something at about 2:53 in the video, the close-up slo-mo. Seems that the first stage rotates a few degrees, first a tiny bit clockwise, then a lot more counter-clockwise, before it drops off.

This reminds me of the first time I tried roll control with Sunguidance, the coupler joint between the main body and guidance section was a nice smooth slip-fit, and the force of the roll control, when it over-controlled and locked-up, was so severe that it caused the guidance section to rotate a lot while the main body under it did not rotate as much (in the Cineroc film the main body fins appear to spin around and around, as did the ground).

Now in your case the roll control was very good, unlike that one.

And it flew well. So the slight rotation of the first stage in relation to the rest of the rocket may not be a problem. But I did want to point that out in case you'd missed it. Several ways it could be "keyed" to not rotate but still slide easily. But if it is not a real problem for flight then KISS may be best (in this case If it ain't broke, don't fix it).

After the booster sep, the model seems to have just a slight wobble in pitch and yaw as though that is the control system doing its job. Does the data correlate with that?

In any case, I'm really glad you added onboard video since data is nice, but it is easy to get lost in the numbers and analyze. With the video, you can just see it doing it's thing apparently the right way.

- George Gassaway

georgegassaway said:

Congrats on another great test flight.

I noted something at about 2:53 in the video, the close-up slo-mo. Seems that the first stage rotates a few degrees, first a tiny bit clockwise, then a lot more counter-clockwise, before it drops off.

- George Gassaway

Click to expand...

Yes, I noticed that. After burnout, there isn't much force between the parts. When the rocket rolled, and the canards tried to resist that roll, the torque turned the parts relative to each other. At least that's what I think explains it. I will probably key this on the three stage, or separate closer to burnout.

I really don't know the cause of the wobble. I don't think it is the result of the control action.

Jim

Jim, I assume you are using charges to separate the stages? If so, could you just use shear pins to eliminate the slip?

markkoelsch said:

Jim, I assume you are using charges to separate the stages? If so, could you just use shear pins to eliminate the slip?

Click to expand...

I could have done that, and could do that on the three-stage, but I chose not to. I will probably key the three stage, but I expect the first stage to drag separate (rememeber it has those fin crowns).

Jim

JimJarvis50 said:

I will probably key the three stage

Click to expand...

What do you mean by key the stage? Learning new things here...

SciAggie said:

What do you mean by key the stage? Learning new things here...

Click to expand...

What I mean is to install some mechanical means to keep the parts from turning against each other. One pic shows how I have done this with fins against the interstage coupler. They second pic shows what I will do for the three stager. Basically, you cut a notch or hemisphere into the end of the upper airframe, install it on the coupler (and install shear pins if there are any) and then fill the notch with epoxy (in black). When you separate the tubes, you have a notch that is an exact fit for the epoxy "key". I have done this using a dowel as the key, but the epoxy gives an exact fit and I prefer doing it that way. On my three-stager, every section of airframe is keyed in some manner to the adjoining airframe.

Jim

When you form the "key" what release agent or method do you use to prevent bonding the parts together? Thanks - a little off topic, but it is an interesting and useful technique.

SciAggie said:

When you form the "key" what release agent or method do you use to prevent bonding the parts together? Thanks - a little off topic, but it is an interesting and useful technique.

Click to expand...

I just put a little wax wherever I don't want things to stick. If I can, I'll catch the epoxy while it's still a little green and clean up the shape a little if any epoxy goes anywhere it's not supposed to. So far, I haven't glued any tubes together.

Jim

PS - The other benefit of doing this on my three stager is that it helps me remember how all of the pieces go together!

I found wax base lip balm works well for this and is easy to apply.

After the discussion of "keys", I made a few for the stabilization spool. On the upper end (between the stabilization spool and the 2nd stage), I used a dowel to make the key. There is a section of airframe that covers the bottom of the motor (the part that sticks out from the airframe and is the interstage coupler). The stabilization spool is shear-pinned to that (in the holes that are visible), but the key will provide a bit more security. On the bottom, between the 1st stage and the spool, I made an epoxy key. Still haven't accidently glued the tubes together!

So, I'm not planning any more test flights until after the Balls flight. Time to fade away to the 9th page of musty old threads. But, there is an excellent group of five rocketeers who will take the concept further than I have. At some point, I'm sure they will share.

Speaking of the Balls flight, the idea is to stabilize the second/third stages during the coast period prior to the second stage ignition. I have completed the simulations and submitted my information for Tripoli approval. The stabilization system brings down the projected altitude to around 130K, but maybe I won't have to walk so far to find the pieces.

Jim

One thing I learned from the Balls flight is that I need a better way to support the canards. My attachment method is axial to the servo hub, which isn't really how they are supposed to be used. I found the perfect bearings, though, and they're installed now. This makes the connection point much stronger.

Jim

JimJarvis50 said:

One thing I learned from the Balls flight is that I need a better way to support the canards. My attachment method is axial to the servo hub, which isn't really how they are supposed to be used. I found the perfect bearings, though, and they're installed now. This makes the connection point much stronger.

Jim

Click to expand...

Those look really nice, should work well.

Mr. Jarvis, you are miles ahead most of us in this hobby--keep pushing the envelope. /Tim

dixontj93060 said:

Mr. Jarvis, you are miles ahead most of us in this hobby--keep pushing the envelope. /Tim

Click to expand...

Why, thank you Mr. Dixon! Next year's project starting to take shape.

Jim

JimJarvis50 said:

Next year's project starting to take shape.

Jim

Click to expand...

Watching with interest! This is just so cool; thanks for keeping us posted, Jim! 130k' and climbing; way to go!!!!

-Eric-

Here's another thread that hasn't been updated for a while. Couple of things ....

Back in May, I asked for some beta testers to try out a stabilization project with the UDB5 controller. I believe there were six or seven volunteers, and three of them went on to do projects (the other guys are just a bit slower). Fortunately, several of the guys are pretty handy mechanically, which is important to success. Test flights are on-going, as we continue to try to figure out appropriate canard size, gains, etc. There are some interesting videos and data which I trust will get posted at some point. It's also possible that some commercial packages will be developed, which will be a big help for folks that want to attempt stabilization but don't have the time/ability to deal with the mechanical issues. That may be a ways down the road - hopefully not too far - but I would say that if there is anyone out there that would like to attempt a project doing their own mechanical design, it would be possible for us to support that.

The UDB5 control program continues to evolve. We're currently on the second generation of the control system. There have been some enhancements such as separate enabling of yaw/pitch and roll control and an algorithm that distributes the control responses to the canards so as to minimize the maximum angle of attack. The algorithm also allows additional response for roll control if the servo range for yaw/pitch is not fully used. The current firmware is still proportional control (on yaw/pitch angle and roll rate), and we may begin to experiment with some version of PID control.

My current plan is to conduct another test flight as soon as I can to test the second generation of the firmware at a higher overall velocity (up to about Mach 1.5). This would be another "two-stage" flight but with no motor in the upper stage, and guidance during the coast period. This is in preparation for the Balls flight this year.

Jim

Great job Jim!

The amount of info you have developed and made available via this thread thus far exceeds anything else I've been able to find. Consider me subscribed to this thread!

I am an engineer who recently "re-discovered" the hobby of model rocketry, and although I am a long, long way from this kind of work, you really have given me something to look forward to!

I second the comment that this should be thread of the year.. (Do they have this sort of thing on this forum?) If it didn't make it for 2015, then perhaps it should be nominated for 2016..

All the best,

Dave

PS: do you have any detailed info on the derivatives used for roll vs pitch and yaw stabilization?

dlavallee said:

I second the comment that this should be thread of the year.. (Do they have this sort of thing on this forum?) If it didn't make it for 2015, then perhaps it should be nominated for 2016..

All the best,

Dave

PS: do you have any detailed info on the derivatives used for roll vs pitch and yaw stabilization?

Click to expand...

Thanks Dave, but I think any rumors of thread of the year are cut from whole cloth.

I don't really have any specific information on derivatives. PID control is something we're just starting to think about. However, I can share my current proportional control gains, and ask for comments if there are any. These settings are what I'm planning to test at my first opportunity and then use at Balls if the settings seem appropriate. Again, the stabilization module is attached to the bottom of the second stage and provides stabilization during the coast period before the sustainer motor lights.

For yaw/pitch, the canard response would match the angle of the rocket up to an angle of 7.5 degrees. Above that, the canard response would be pinned at 7.5 degrees. The objective of this approach is to minimize the canard deflection to prevent stalling them.

For roll, the canard response would reach 7.5 degrees for a 250 degree/sec roll rate and level out at 7.5 degrees at higher roll rate. This response would apply if there is also 7.5 degrees of canard response for yaw/pitch, such that the maximum response of the canards would not be more than 15 degrees. However, if the yaw/pitch response is less than 7.5 degrees, then the response allocated to roll could increase, up to a maximum of 15 degrees at higher roll rates.

I would also like to ask for comments on the position of the canards. The third pic shows the position of the three fin sets looking down the axis of the rocket (with the rocket sitting within the frame of the launch rail). I think I have two options. One is to align the canards with the booster and sustainer fins. The other is to offset the canards by an angle of about 22 degrees, where they would be in clearer air, which is the option shown in the pic. The canards would be about 12" below the sustainer fins, as shown in the fourth pic.

Jim

JimJarvis50 said:

I would also like to ask for comments on the position of the canards. The third pic shows the position of the three fin sets looking down the axis of the rocket (with the rocket sitting within the frame of the launch rail). I think I have two options. One is to align the canards with the booster and sustainer fins. The other is to offset the canards by an angle of about 22 degrees, where they would be in clearer air, which is the option shown in the pic. The canards would be about 12" below the sustainer fins, as shown in the fourth pic.

Click to expand...

i would have thought a 45 degree angle is going to be a lot better. Pretty sure that as the canards angle over their wake will interfere with the booster fins with unpredictable results. I suspect you don't want to do this as it will interfere with the tower, so you will need to modify the tower as well.

Hi Jim,

So active stabilization is only during the coast phase? Or during the booster phase as well? With booster attached, without, or both? I would imagine that having the canards active for both (booster phase and sustainer only) would add complexity due the difference in the control derivatives.. Forgive my ignorance in multi-staging as I am still very new to h.p. model rocketry.

Thinking out loud here.. what if the canards were aligned with the sustainer's fins, but ahead of them (if they are only used while the booster is attached)? Because the sustainer's fins are less crucial while the booster is attached, the turbulence is even less of an issue, plus it increases the torque moment that their deflection would provide with respect to the CG..

Still new to this, and aerodynamics are not my specialty so take any of my suggestions with a grain of salt!

-Dave

dlavallee said:

Hi Jim,

So active stabilization is only during the coast phase? Or during the booster phase as well? With booster attached, without, or both? I would imagine that having the canards active for both (booster phase and sustainer only) would add complexity due the difference in the control derivatives.. Forgive my ignorance in multi-staging as I am still very new to h.p. model rocketry.

Thinking out loud here.. what if the canards were aligned with the sustainer's fins, but ahead of them (if they are only used while the booster is attached)? Because the sustainer's fins are less crucial while the booster is attached, the turbulence is even less of an issue, plus it increases the torque moment that their deflection would provide with respect to the CG..

Still new to this, and aerodynamics are not my specialty so take any of my suggestions with a grain of salt!

-Dave

Click to expand...

Actually, the plan would be to activate roll control during the boost and during the initial coast period (after the booster drops off). Once the velocity drops to about 1000 ft/s, yaw/pitch control would be activated. This would continue for about 10 seconds or so as the sustainer slows to something like 600 ft/s. Then, the stabilization spool would be separated and the sustainer motor ignited. The strategy is designed to minimize roll (bad for a stabilization system) and to have the sustainer pointed mostly up when the motor lights.

It would be possible to use yaw/pitch control during the boost, but there are two reasons not to. First, the canards are located near the CG prior to booster separation, so they are not very effective for yaw/pitch anyway. If used, and if there was an angle from vertical, then the canards would cause some torque at that point in the air frame. I believe that this is the main reason why the booster and stabilization sections didn't separate in the Balls flight last year (although there is more to it than that). Second, in the case of the Balls flight this year, the CG would switch from behind to ahead of the canards at stage separation. We don't have the ability to switch the control directions with the current control program, and we don't really want to do that anyway.

I have devised the stabilization strategy above to point the sustainer vertical at the point where the motor lights, and to limit the application of stabilization to a relatively narrow speed range. It would be desirable to have stabilization through the entire flight, but I'm not in a position to operate with canards at over Mach 3. At some point, it would be interesting to vary the gain as a function of velocity (or flight time). One advantage of the approach I am using is that the canards (and their drag) are not present once the sustainer lights.

Jim

SpaceManMat said:

i would have thought a 45 degree angle is going to be a lot better. Pretty sure that as the canards angle over their wake will interfere with the booster fins with unpredictable results. I suspect you don't want to do this as it will interfere with the tower, so you will need to modify the tower as well.

Click to expand...

I'm not really sure if it would be better to locate the canards under the sustainer fins or out in clear air. It is possible that locating them under the sustainer fins would help with stall angle, but with a separation of 12 inches, I don't know if this would really help. But having them aligned might not be a problem either.

Regarding interaction with the booster fins, my belief is that with the long booster, that this would not be an issue. I am concerned about this for the test rocket, though, because the booster is much shorter and has 3 fins rather than 4. For the Balls rocket, there is a separate composite fin can. It would sit on a fiberglass ring that would be held in place using the bolts for the nozzle holder. It might be possible to configure this to allow the fin can to spin freely, although I suspect that as the motor heats, that the fin can would get pinned at some point.

Jim

Saw a fellow at the Mini-Midwest Power on 5/15 with a rocket that had canards "that looked familiar". I then overheard him say he was your beta tester. I didn't get a chance to directly see the launch or whether if flew successfully.
The rocket looked very nice and he demonstrated the canard motion on the ground by rotating/tilting the rocket with the system turned on. Kurt

ksaves2 said:

Saw a fellow at the Mini-Midwest Power on 5/15 with a rocket that had canards "that looked familiar". I then overheard him say he was your beta tester. I didn't get a chance to directly see the launch or whether if flew successfully.
The rocket looked very nice and he demonstrated the canard motion on the ground by rotating/tilting the rocket with the system turned on. Kurt

Click to expand...

That was my rocket ExPAC (https://www.rocketryforum.com/showt...lization-and-Steerable-Recovery-Test-Vehicle/) at MMWP. You can see the build details of the canard bay in that thread. It is currently using the board and code from Jim's project. I have yet to post the video or flight details from my last two launches do to a lack of time to edit the video. The MMWP flight went very well, despite the very high winds that caused most flights to weather cock badly. The rocket went up very straight with only a little roll, and the video shows that instead of arcing over the rocket back slid at deployment.