ExPAC ver. 2 - Active Stabilization and Steerable Recovery Test Vehicle

Thank you guys for the response. Since Jim is here too I can ask a few more specifics

Does the angle of the canard scale with angle of attack, for instance very minor adjustments around 0°, and greater throw at larger angles of attack (to the stall limits of course)?

I'm still in the planning stage for mine (canard stabilization), but was planning on using a PID control loop for control, but wasn't sure besides flying it a bunch to tune the PID values to avoid oscillation. Access to a wind tunnel would be nice too I suppose.

I could see how just running the data from the gyro to adjust the canard angle should work well though...

Sorry if I'm rambling a bit...

Are you guys adding vibration dampening to the control board? Basically soft rubber etc between it and the rocket to rid the sensor of some of the huge spectrum of vibration during motor burn?

Jeffmhopkins said:

Thank you guys for the response. Since Jim is here too I can ask a few more specifics

Does the angle of the canard scale with angle of attack, for instance very minor adjustments around 0°, and greater throw at larger angles of attack (to the stall limits of course)?

I'm still in the planning stage for mine (canard stabilization), but was planning on using a PID control loop for control, but wasn't sure besides flying it a bunch to tune the PID values to avoid oscillation. Access to a wind tunnel would be nice too I suppose.

I could see how just running the data from the gyro to adjust the canard angle should work well though...

Sorry if I'm rambling a bit...

Are you guys adding vibration dampening to the control board? Basically soft rubber etc between it and the rocket to rid the sensor of some of the huge spectrum of vibration during motor burn?

Click to expand...

The response is just linearly proportional at present. I think it would be easy to put in a different function, but we haven't done that. I believe the unit could also be configured for PID, but we have elected not to do that either. It's just my opinion, but I think those types of things are refinements to be added after first getting a sound mechanical system in place. I haven't satisfied myself that I've gotten to that point yet. I don't have any vibration dampening, but it sounds like a good idea.

When I have a little time, I'll post more on the current version of the control system.

Jim

A few clear coats and the paint will be done, for now:




We didn't have time for the letter outlines, which will be added after the maiden launch. We will also be adding a durable epoxy clear coat to keep it looking nice after any touch-ups post launch.

Beautiful...

for a second there I thought you mixed up your photo with one of Nathans rockets! LOL

It's been a while since the last update. I got everything finished and flew the maiden flight at the 3/19 QCRS launch.
The steering bay was completed and laid out with the kite and control lines. To avoid damaging the servos with forces of recovery, even though their very robust, the kite lines go directly to the steering bay and the servos pull on loops in the lines to steer. When they are fully paid out there is a small amount of slack on the servo lines. They are left slack until after deployment to avoid putting stress on the servos. The third line in the middle of kite is what the manufacturer calls a “depower” line it is used to collapse the front of the kite and reduce its pull. This should allow me to use it to increase the decent rate from higher altitudes.



For the steering to work the bay has to hang horizontally below the kite. To accomplish this one of the kite lines is connected to each end of the bay and the mass of the rocket is hung from a Y-harness attached to both ends of the bay. Because of this the lines have to cross alongside the bay without snagging or tangling wile everything in loaded. A lot of taping and z folding is employed to keep everything in place.



The three control lines and the kite are folded into the burrito then the bay and burrito are slid into the rocket along the internal guide rails. Once it’s installed the canard bay is plugged into the connector on the top of the steering bay and the canard ban and nose are bolted on.

The first flight was flown on a CTI K515SK, and despite good weather not everything went as planned. Initially during the boost the canards seemed to stabilize the flight, but the maximum travel they were set for allowed one of the fins to stall. This caused the rocket to enter a spin that saturated the roll gyro causing the stabilization system to lose control. Because of this the rocket coned the whole way up and reached an apogee of 2500ft, 750ft below the simulations.

Then at apogee the steering bay and kite deployed rapidly from the rocket but one of the kite shroud lines tangled. This caused one side of the kite to fail to inflate fully causing a spiral on decent. I was briefly able to steer against the spin but the body components had begun to spin around each other. I would regain control of the kite only to have the spinning rocket pull the kite and cause it to sway and lose lift. Eventually I gave up and let the rocket spiral in.

Despite the failures I learned a ton from the three on board cameras, though the video made me queasy. I have changed the maximum travels of the canards to prevent a fin stall, reconfigured the shock cord to keep the components in line so they can’t spin, and rethought our folding technique to prevent the tangle we had. I will be launching it again on a CTI K940WT next weekend with Michiana. Hopefully we will have a much more successful flight this time.

Here is the video from the nose camera which ends shortly after deployment. I won't be posting the dizzying decent video, but you can see the kite tangle and start of the spin at the end of this video.

[video=youtube_share;_X7O1Ppl90M]https://youtu.be/_X7O1Ppl90M[/video]

Very cool what your doing! Really neat to get all that working, don't give up!
Just a quick view From the video it appears the canards move really far, like the correction would happen too far too fast then try to over compensate for the over correction and repeat. Do have a way of slowing or limiting the movement? Like a car in a fish tale.

I'm surprised that K515 didn't push it higher! That's a strong engine, must be some weight in that bird!

Still very cool, hope to get to see that fly one day

The rocket is made of fiberglass, all of the bulk plates and centering rings are linen phenolic and it has a significant section that is double walled for the steering bay to slide in. All of the hardware is stainless steel. The components include:
3 cameras
7 servos
2 RC receivers
1 Raven
2 2S lipo batteries (1 180mAh, and 1 1500mAh)
11 quick links (7 1/8", 4 3/16")
8 u-bolts
and a bunch of other miscellaneous stuff

Fully loaded with the motor it weighs approximately 21lbs.



As for the canard corrections I have reduced their maximum travel to about half which also effectively cuts the correction gain in half. I do have the ability to control the gain independently but I'm going to try these settings on the next flight.

Beautiful rocket, that is a beast. Good luck on the next flight!

Ravenex said:

...
As for the canard corrections I have reduced their maximum travel to about half which also effectively cuts the correction gain in half. I do have the ability to control the gain independently but I'm going to try these settings on the next flight.

Click to expand...

Are you using PID control schemes for the canards? If so, it may just need some tuning.

Currently it's proportional only, we are discussing using PI for roll and PD for pitch/yaw. But it will be a little while till it's implemented. I'm very familiar with control loop tuning as my job regularly involves programming and tuning high dynamics servo systems.

I'm extremely excited for this weekends launch with Michiana, we're expecting 70 degrees, sunny skies, and low winds. All of the improvement to the rocket from the last flight have been made and everything is packed. I think the first flight will again be on a CTI K515SK. If the canards don't cause it to cone it should reach 3200 ft. If all goes well I will attempt to repack the steerable kite and fly it again on a K520 or K940 to 3600 ft.

Waiver buster! Sweet...but maybe an extra 0 in there . Hope the flight goes well, really wish I could make it this weekend

Ravenex said:

I'm extremely excited for this weekends launch with Michiana, we're expecting 70 degrees, sunny skies, and low winds. All of the improvement to the rocket from the last flight have been made and everything is packed. I think the first flight will again be on a CTI K515SK. If the canards don't cause it to cone it should reach 3200 ft. If all goes well I will attempt to repack the steerable kite and fly it again on a K520 or K940 to 36000 ft.

Click to expand...

Typo? Good luck on the flight, I've learned a lot from this thread and your amazing build.

Typo fixed

So just a thought after viewing the flight video, I think there needs to be a priority system between gyroscope and inclination based corrections. So only correct roll until below 1Hz revolution on the gyroscope, and then start correcting the angle of the rocket to 0° (and roll simultaneously). And of course if it starts rolling faster, priority shifts back to only roll control. The other observation is that you should have a velocity dependent canard travel distance, this way at slower speeds you can still have higher throw, and at high velocity it appropriately scales to prevent high oscillation.

Looks like you are on the right path though

So due to a busy schedule and then some family medical issues it has been way to long since I’ve been able to update this thread. I’ve had two more flights with a lot of success and a good bit of failure. But on the whole the system is working well and I’m learning a lot about the potential failure points making improvements for the final design.

The first flight was at the 4-17-16 Michianna launch I mentioned in my previous post. For this flight I reduced the maximum travel and gain for the canards to half of what was used on the previous flight. Also there was a firmware update for the stabilization system to correct a glitch that caused twitching of the canards under certain conditions. We also came up with a new method for packing the kite to avoid the tangle we saw at QCRS.

The weather was fantastic, sunny, warm, and very calm winds. The rocket launched on a CTI K515-SK to altitude of 2940ft. On the way up there was still a lot of wobble but less than the last flight and the roll control was greatly improved.

On deployment the tail section rotated away from the upper airframe and the aggressive opening of the kite caused the tail to whip violently. This resulted on the shearing of the pins on the backup chute bay and a minor zippering of the airframe. The backup chute deployed and created a drag below the steering bay. This caused the bay to pull behind the kite and the kite to nose down, increasing the decent rate dramatically. The backup chute dragging also caused the steering bay to not untwist making forward facing video less forward facing.

Despite these problems the kite deploys fully and the steering worked wonderfully. Over the 153 second decent from 2800ft I was able to complete many controlled turns and landed the rocket only about 20ft from where we were standing in the field. Here is a video of the flight take from the ground by another flier at the launch. You can me and my dad in the video during landing, we’re on the left, I’m in the black shirt and hat.

[video=youtube_share;QLY7WcrzByI]https://youtu.be/QLY7WcrzByI[/video]

Here is the onboard video from the same flight from the nose mounted camera from launch until just after deployment.

[video=youtube_share;AS3WWiMFPYQ]https://youtu.be/AS3WWiMFPYQ[/video]

Very cool - there is a lot going on there and to land that right by you is a big accomplishment !

Ravenex said:

I’ve had two more flights with a lot of success and a good bit of failure.

Click to expand...

There was a second flight?

Jim

JimJarvis50 said:

There was a second flight?

Jim

Click to expand...

Yes, at QCRS 5/15, I apologize for not getting you the data yet. As I mentioned my wife was sick for a few weeks, I still haven't even pulled the stabilization data from that flight, I will try to do that tonight. Despite the high wind that had everyone else weather cocking badly my flight went up very straight, in fact it back slid at apogee.

[video]https://youtu.be/P4QbyOHm4PE?t=14m13s[/video]

Sorry for the long delay between updates. As I mentioned the flight at QCRS was on 5-15-16, ExPAC flew on a CTI K520 to 3300ft. There were strong winds on the ground, and at altitude. Most of the flights of the morning were weather cocking badly and then drifting back to the pads on decent. ExPAC exhibited no weather cocking and only a slight wobble, the tuning on the stabilization seemed to perform much better. At apogee the kite deployed nicely and I was able to make a number of controlled maneuvers directly overhead. After descending a bit the kite seemed to hit a higher velocity wind layer and stopped making progress into the wind. Despite my best efforts at this point the rocket drifted backwards about ½ a mile at which point I depowered the kite causing it to more rapidly descend. In hind sight I should have depowered it immediately upon encountering the higher wind speeds and allowed it to descend below them before inflating it to steer back. The rocket was recovered easily, however one of the side plates of the steering bay had broken off and the upward facing camera was missing. This seems to have been the result of a poorly prepped epoxy joint and the bay impacting the airframe at some point. Further analysis of the flight data confirmed my suspicion that the kite I have been using is grossly oversized (2.5 sq m.) for the rocket resulting in the aggressive openings that have been plaguing the project. Despite having a projected area equivalent to a 70” round chute and a it descends at only 5.5 ft/sec with a load of 19.7lb. Even when deflated the decent rate was only 19 ft/sec. With a fixed glide ratio this slow decent rate, I’d like to see 18 ft/sec, results in a forward penetration of 30% of what was expected. I ordered, and have since received, two smaller custom kites for testing a 1.4 sq m. and a 0.8 sq m.

The final flight of ExPAC ver.2 was on 6-18-16 at the WOOSH launch at Bong Recreation Area, WI. I replaced the side panel and missing camera from the steering bay. I had also intended to use a smaller kite which I paid to expedite shipping on, but it didn’t arrive on time, and I planned to add GPS telemetry in the kite bay after I received a new transmitter but I had some power supply issues. These two things may have made a world of difference.

ExPAC flew on a CTI K940 to an estimated 3500 ft. The boost was awesome and the stabilization was near perfect with only a slight roll and it backslid at apogee before separating. Unfortunately that’s where the perfect flight ended. The kite burrito snagged in the airframe for nearly 15 seconds and the rocket was able to gain some speed before the kite wiggled free. The resulting aggressive kite opening caused the shock cord to zipper the airframe which then cut clean through it in two places separating the upper and lower airframe sections from the steering bay and kite.

The tail section nosed down and slammed into a shallow pond crushing the av-bay and forcing it into the payload tube below it and forcing the tube down into the tail section. The stabilization and nose section fell flat hitting the gravel road and bounced several feet. The steering bay and kite drifted far down range before I “depowered” the kite and it spiraled to the ground. The tail and nose sections were recovered a few hundred feet away. Unfortunately 8 hours of searching and 8 miles of walking with two people failed to find the steering bay and kite.

The only data that could be recovered was the video from the nose camera:

[video=youtube;aY5f9s-P6No]https://www.youtube.com/watch?v=aY5f9s-P6No[/video]

The tail section will be cut off and rebuilt. The av-bay needs to be completely rebuilt, the R/C electronics survived but the Raven altimeter was destroyed. The nose section survived largely undamaged. Despite landing on two of the canards none of the servos were damaged. One canard cracked its fillets but is still sturdy enough I can’t separate it from its base plate and two of the plastic aesthetic protrusions cracked off of the airframe. The largest loss however was the steering bay which included 3 servos, a receiver, 1500mah battery, 2 cameras, a 5V power supply, and the 2.5m kite.

These setbacks have forced me to accelerate my plans to move on to rebuild the rocket as ExPAC ver.3 and move closer to commercialization of my steering and stabilization products. As such I have been working on improvements to the manufacturability and ease of use as well as component reductions. I will be starting a new thread shortly dedicated to the Rev 3 build.

Pictures of the crash damage:

I'm so sorry. This was a great project. I'll be watching the ver. 3 thread too.

Thank you, when I can find the time I will start the new Rev.3 thread as well as a thread about my progress on my stabilization module I feel is ready for testing by others.

Sorry to hear about the crash and loss.
I haven't worked with stabilization but did put together an RC controlled steerable chute. I still haven't got it working right. It's always deployment issue that bite me, not the controls.
Good luck with v3

For those who are following this project I have finally posted a thread for ExPAC ver 3.x:
https://www.rocketryforum.com/showthread.php?143076