Glider with GPS guided recovery

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
For now I'm thinking of using Staples packing tape as hinges (one piece of tape on each side) to attach the elevons - anyone have thoughts on that?

That will do nicely, in fact, that is what I tend to use most often. This has worked on well over 100+ gliders over the years, and even all the way up to J motored versions 15-17 years ago. Just like you said, one piece on each side does the trick.

Thin 'chrome tape' is even better if you arent too worried about added weight, but that stuff does tear a bit too easily. Of course, two-plying an elevon with this plus the standard clear tape makes for an elevon not going anywhere.
 
I've just browsed the thread, so forgive me if some of this was covered.

Have you considered applying the gps control to a known flyable RC type plane, say a simple delta winged profile that is simple and cheap, take it up, and enable the gps and see if it can guide it the way you want.


Then try to get the rocket glider flying with rc control to figure out throws, mix and CG.

Then move the gps control into the rocket and apply your gradual approach to tuning the controls once you have a known quantity.

It seems you should separately validate the guidance and aerodynamics and then try to combine them.

Also, I think that having a small delta wing like this with a nose heavy post boost condition that may be likely unless you move ballast around will tend to require more elevon throw for pitch than for roll control. For roll control you have a very small span and I've found those more roll sensitive, even with some dihedral. I'm guessing you will run out of sufficient elevon authority before having problems with sufficient roll.

Frank
 
I agree. At least do some glide tests in a configuration that represents post-burnout. Stability margin is your friend on the way up, but it's your enemy on the way down. A severe pitch-down tendency with controls centered means that at a minimum you'll get a lot of drag from the elevons deflecting so far just to stay level. At worst, you won't have enough control authority to keep a controlled glide, so the rocket will come in ballistic until the chute comes out to zipper the airframe.
 
Have you considered applying the gps control to a known flyable RC type plane, say a simple delta winged profile that is simple and cheap, take it up, and enable the gps and see if it can guide it the way you want.

Then try to get the rocket glider flying with rc control to figure out throws, mix and CG.

Then move the gps control into the rocket and apply your gradual approach to tuning the controls once you have a known quantity.

It seems you should separately validate the guidance and aerodynamics and then try to combine them.

Yes, that is more-or-less the plan, except the that glider is a new design. But it's closely modeled on an existing kit.

Here's the test plan (posted here a few weeks back):

Test 1 - Small motor (500' or so). No left/right steering - steer straight. Experiment with pitch until we find the pitch setting that gives a nice stable glide (no stalling). Repeat until successful.

Test 2 - Small motor (500' or so). With fixed pitch setting (from step 1), gradually test increasing left/right steering commands to determine the maximum steering input that flies stably without flipping over. Repeat until successful.

Test 3 - Guided flight.

Also, I think that having a small delta wing like this with a nose heavy post boost condition that may be likely unless you move ballast around will tend to require more elevon throw for pitch than for roll control. For roll control you have a very small span and I've found those more roll sensitive, even with some dihedral. I'm guessing you will run out of sufficient elevon authority before having problems with sufficient roll.

Well, we'll find out.
 
At least do some glide tests in a configuration that represents post-burnout. Stability margin is your friend on the way up, but it's your enemy on the way down. A severe pitch-down tendency with controls centered means that at a minimum you'll get a lot of drag from the elevons deflecting so far just to stay level. At worst, you won't have enough control authority to keep a controlled glide, so the rocket will come in ballistic until the chute comes out to zipper the airframe.

Boris built and flew the 1/3 scale model of the glider to check that out - it seemed to fly OK, and we are trying to stick to the same CG and CP in the full-scale glider.

Still, it's a good thought. To minimize the chance of that happening, I'll rig the elevons to allow A LOT of throw in pitch for the first flight - that way we won't run out of control authority, if the design can fly at all. (I think I can get the elevons 90 degrees off the wing.)

Also, I have the electronics rigged up so I can trigger the parachute manually at any point. If it looks like it's coming in ballistic and we can't get it into a controlled glide, I can fire the chute early before it builds up a lot of speed.
 
Over the weekend I attached the elevons with Staples packing tape, and installed the control horns, clevii (?), and rods.

I haven't attached the rods to the servos yet - I want to run a little test cycle on the servos and then decide which horn to attach.

I also constructed a small piston for the parachute ejection. It's got 2 charge holders - the center one is the primary charge, the offset one is a backup (that can only be triggered manually, by radio command).

Then I drilled a little hole, just forward of the bulkhead, for the wires to go to the charges - they'll run along the outside of the body tube.

And I installed a small 2-56 screw into the bottom of the body tube; this will be an alignment key to ensure the electronics bay is lined up straight with the tube - I'm putting in a video camera that will look out thru a hole, so I want to be sure the camera is aligned with the hole.

The plastic stuff forward of the screw is bubble wrap - since the forward portion of the electronics bay will be empty, I wanted to fill it up with something light that will displace air, so less air has to rush out thru the vent hole on ascent.

IMG_3744_elevonAndServo.jpg

IMG_3749_elevonSideView.jpg

IMG_3752_ejHoleAndKey.jpg

IMG_3757_AlightmentKeyAndBubbleWrap.jpg
 
Today (Monday) I mostly finished building the electronics sled. I moved around the PCB a bit from its previous position - I was concerned that the video camera would interfere with the telemetry radio, so I moved the camera to the opposite side of the PCB (the trailing side).

Then I got worried that EMI from the video camera might affect the GPS (since now it's an inch from the GPS), so I wrapped the whole video camera in aluminum foil - that should cut down the EMI (I hope).

My idea is to glue the plywood sled with the electronics into the coupler tube; there are no bulkheads on either end of the tube - it's held in place by a spacer on one end (against a bulkhead) and the alignment screw in the other direction.

I glued the sled into the tube, but it turned out the weight of the sled (only 125g) deformed the shape of the coupler tube enough that it wouldn't slide smoothly into the body tube. So I ripped it out and am now re-gluing it by holding the sled fixed and gluing the (lighter) tube onto it. The pix were taken as it's drying. I hope it doesn't bind.

The narrow slot in the tube is for the alignment pin (the 2-56 screw); the tube goes into the bay, slides forward, I install the spacer, and then slide it back into place.

IMG_3758_ebayBottom.jpg

IMG_3761eBayForwardEnd.jpg

IMG_3763ebaySide.jpg
 
It seems to be working - it slides easily into the body tube now.

In the picture, you can see the video camera on the right, wrapped in foil and hot-glued into place. The little square thing on the PCB to the left and below the camera is the GPS receiver.

The empty hole in the plywood sled is where the servo leads will run.

The whole e-sled drops into the bay, slides forward, I install the spacer (marked with flight direction; one end is sanded down to fit), then slide the sled backward against the spacer.

Also, a couple of pictures of the piston (not glued together yet).

IMG_3766_Esled.jpg

IMG_3764_PistonFront.jpg

IMG_3765_PistonBack.jpg
 
Boris built and flew the 1/3 scale model of the glider to check that out - it seemed to fly OK, and we are trying to stick to the same CG and CP in the full-scale glider.

Still, it's a good thought. To minimize the chance of that happening, I'll rig the elevons to allow A LOT of throw in pitch for the first flight - that way we won't run out of control authority, if the design can fly at all. (I think I can get the elevons 90 degrees off the wing.)

Also, I have the electronics rigged up so I can trigger the parachute manually at any point. If it looks like it's coming in ballistic and we can't get it into a controlled glide, I can fire the chute early before it builds up a lot of speed.

That sounds good. If you have a good hill where you live, and wind blowing up it, you can do some slope soaring for a sustained glide test.
 
I know you are getting lots of advice, some good and some bad...

I have literally thousands of R/C flights on Delta wing gliders (shuttles) of all sizes. A Delta performs different than a flying wing. Flying wings are sensitive in pitch while Deltas are sensitive in roll. George points this out. You will probably find you need about half the roll authority as you do pitch. The 4x ratio the other way for a flying wing seems about right.

You should build another model for strictly R/C. You shouldn't have any problems finding a local RCer to take it up on the back of an airplane and drop it off. You'll get all the data you need for stability, control (and trim). It would be a worthwhile one-day project. I wish I lived closer, I would do it for you.

Don't be turned off by the nay-sayers on the guidance. I worked on the problem 15 years ago for a Graduate thesis. 10 years ago a friend and I re-invented the software and hardware. Its not a difficult problem. Any micro controller hobbyist should be able to crunch right through it. It turns out an aircraft in gliding flight is about the easiest to control (given a stable model to begin with. See above on testing :) A parachute is slightly easier because you get built-in energy management; brakes. The logic for the glider isn't that bad and you should have already figured out energy management by this point.

Don't worry about feeding terrorists. They are much more sophisticated than you think. You are not inventing or creating anything that hasn't already been done for years.

Scott
 
Last edited:
I've finished installing all the electronics, as well as the ejection piston, into the glider.

Empty (no motor), the whole thing weighs 45.95 ounces (1303 grams). With the Roadrunner G80-10R motor installed, it's 50.6 ounces (1434 g).

Re the CG, empty it's about 1.5 inches forward of the mark on the bottom. With the motor installed, the CG is right at the mark (it just came out that way).

Boris - is this what you wanted, or do we need to worry about how to push the CG backwards?

In the first pic, you can see the servo wires running along the underside of the wing. I'm not 100% happy about just taping the wires to the wing with Staples packing tape - I'm concerned the tape might come off in flight (it doesn't stick so well to the fiberglass).

I'm half-tempted to dig a channel into the wing to bury the wires flush into the wing. (I'd fill it with epoxy then glass over it.) But I'm worried what that would do to the strength of the wing.

In the second pic, you see the servo wires going into the hole on the bottom to enter the electronics bay. I put some hot glue there to stabilize things; these' a little hole (just forward of the wires) where I plan to run the ejection charge wires. You can also see the camera lens peeking out from the hole just behind the wire hole.

The 3rd pic is a close-up of where I attached the extension cable to the servo lead.

IMG_3781_underside.jpg

IMG_3782_ServoWires.jpg

IMG_3783_servoExt.jpg

IMG_3784_servoRigging.jpg
 
Here you can see the e-sled installed in the e-bay.

It slots into the hatch, slides forward a bit, then I install the spacer behind it, and slide that and the e-sled backwards.

There are 3 1/8" holes (all reinforced with 1 layer of 3/4 ounce fiberglass - that's why it's ugly). One hole is in the spacer, and two in the e-sled, that are used to hold down the hatch cover - I'm using plastic removable rivets.

It seems pretty secure.

So the construction and electronics work is all done. (Unless Boris thinks I should bury the wires in the wing.)

All that's left now is some minor software work - mainly setting limits for the servos and, handling manual steering/manual pitch, and mixing the outputs to drive the elevons. That's just minor tweaks to what's already there - a few hours' worth. Then I want to do a lot of testing before flight.

I think we're on schedule to make the 10/22 launch.

IMG_3786ebayExposed.jpg

IMG_3787_ebayButtoned.jpg

IMG_3788_frontview.jpg
 
Hi,

This project look very well, and us I always say: "If it's look nice, it's should fly nice!"

Keep updating us on this!

I wish you the best! :wink:
 
Last edited:
Dave,

Looking very good.

Yellow masking tape would likely hold the wires in place. I have used it successfully on many flights to hold cameras and pods on the outside of rockets.

I would be concerned about cutting into the surface of the wings as the one layer of 1.3 oz FG present is structurally critical. Even if FG is replaced, "structural integrity would be a concern captain, she can't hold on much longer." Er, um, sorry couldn't help myself...

If the measured CG is right on the mark I put on the rocket, or just slightly forward of that point we should be in great shape.

Really looking forward to the 10/22/11 CMASS launch.
 
Last edited:
Yellow masking tape would likely hold the wires in place.

I only have the blue (low-tack) stuff. Bring some to the 10/22 launch and we can do it there (if you think the packing tape isn't sticky enough).

If the measured CG in right on the mark I put on the rocket, or just slightly forward of that point we should be in great shape.

It's right on the mark WITH THE MOTOR INSTALLED. Is that the way you measured it?
 
I only have the blue (low-tack) stuff. Bring some to the 10/22 launch and we can do it there (if you think the packing tape isn't sticky enough).

I will bring yellow masking tape. We can choose tape on the field.

It's right on the mark WITH THE MOTOR INSTALLED. Is that the way you measured it?

That is as intended, marked CG is target when 100% flight ready.

Kenn posted on cmass.org that the Amesbury field is being hayed right now, so we should be good to go 10/22/11.
 
Last edited:
That is as intended, marked CG is target when 100% flight ready.

Excellent! :D

I'm hoping to start on the software stuff later today. Will post videos of the servo actuation once I have that working smoothly.
 
I've been lurking on this thread and it looks like it's coming together nicely. Hope you have a successful launch! (Saturday?)

I saw a similar project about ten years ago when I was working at the Natn'l Severe Storms Lab, here in Norman. A guy from South Africa was working on a glider carring an instrument package, trying to get it to return to a "home base". Their State meteorology division didn't have enough funding to launch what are essentially single-use radiosondes (some do come back) twice a day (or more). He was hoping to build a recoverable, reusable glidersonde that would go up under a balloon. There were several professors and students from OU working on the project, but I've lost touch with them over the years.

I saw this today
https://www.igloocommunity.org/
Check out the product brief.
It's got:
GPS
3-axis accellerometer
3-axis magnetometer
3-axis gyro
pressure sensor,
and more. It has a microSD slot, so in theory, you could have in-flight video or stills.
 
I talked with Dave last night and he is continuing to work on programming the flight control system.

We are a go for first flight(s) at the 10/22/11 CMASS Amesbury launch.

First flight will use RC control to get a feel for elevon positioning and sensitivity. When that is established, later flights will fly under automated control. Control logic can be modified in the field.

We will start with RoadRunner G80 motors (108Ns). Will also have ProX G115 (141Ns) and G125 (159Ns) motors available.
 
Things are looking good for the first flight on Saturday.

Software work and testing are all done - the only thing left now on my list is to ground test the parachute deployment (piston system). I'll do that tomorrow.

Here's a short video showing installation of the electronics bay into the glider, and ground testing of the servos and elevons (starting at about 1:00).

[YOUTUBE]PZM2wFvz6Hg[/YOUTUBE]

Cheers,

--Dave
 
Been watching this thread with great admiration. Hope you have good weather tomorrow, and a great flight. We're all rooting for you.

:clap:
 
I ran some ground tests of the ejection system today (see video below).

It turns out that the parachute has to be packed in a specific way to deploy reliably, otherwise it stays stuck in the hollow nose cone.

That happened on the first couple of tests (at 100 and 200 mg of BP).

When I packed it differently for the last two tests (250 and 500 mg) it worked OK.

[YOUTUBE]UHQ0j1tGbHI[/YOUTUBE]

Unfortunately, 500 mg is too much - I caused a little zipper (see photo). I'm making up eight 250 mg charges for tomorrow. (Probably we won't use all 8, but best to have some extras handy.)

I'm repairing the zipper with a little patch - 2 plys of 2 ounce fiberglass. Should be OK to fly tomorrow.

--Dave

IMG_3802_MinorZipper.jpg
 
We got in 3 flights yesterday. Only one of them was really good, but we learned a lot from all 3; and the glider is in good shape.

We decided to name it the "Autonomy". Our goal for the day was to get some flight information about the proper pitch setting for a nice stable glide.

There is still a lot of data analysis to be done, but you can look at the videos below. The whistling noise on the audio track is a good indication of relative airspeed, I think.

FLIGHT 1 - ROADRUNNER G80

Video from the ground:
[YOUTUBE]1HHERRqYH2k[/YOUTUBE]

Video from the on-board camera (looking out the bottom of the glider):
[YOUTUBE]fo8EtW-tozQ[/YOUTUBE]

Immediately after apogee the glider went into a sharp roll to the left. This was because I forgot to turn off the navigation system, which hasn't yet been calibrated for the roll sensitivity of the glider. It commanded left yaw and put it into the roll. While spinning it lost lift and dived toward the ground.

The parachute deployed at 200 feet AGL as planned. No damage other than a very tiny zipper (similar to the one on Friday).

FLIGHT 2 - CESARONI G54

This was our best flight of the day. I remembered to turn off the navigation system.

Video from the ground:
[YOUTUBE]Jn77Qtw_mu8[/YOUTUBE]

And from the on-board camera:
[YOUTUBE]exEXqyR2ggU[/YOUTUBE]

The glider arcs over a little prematurely. Looking at the video, I think this is because the "straight" boost position for the elevons actually is slightly pitch-up, so the glider tends to arc up (toward the rudder) during boost.

I think we can reduce this in future flights by (a) adjusting the boost position a little bit pitch-down, and (b) adding a little bit of yaw command to the boost position to induce a slow roll during boost to even out any pitch-up/pitch-down tendency.

By chance the glider was right-side-up at apogee, and the transition of the elevon position from straight (for boost) to the glide position is very clear in the video. When the elevons go to glide position the glider does an Immelmann turn (due to the high airspeed at apogee - this was a pure accident) and then settles into a steady glide.

The really good news here is the glider does seem to be self-stabilizing due to the dihedral; regardless of what attitude it was in at apogee, if left alone it seems to settle into a stable glide.

During the glide, Boris was manually adjusting the pitch setting via the telemetry system. The effects of this aren't too obvious in the video, but I'm hoping to learn more when I look at the logged data (servo position vs. climb rate, airspeed, etc.).

Three seconds prior to parachute ejection, as planned the elevons go full-up to induce a flare for airspeed reduction. This is hard to see from the ground video, but the pitch change is pretty clear from the on-board video.

I plan to look carefully at the GPS data from this flight (also correlating that with the two video streams). I hope to get a sense of the forward airspeed of the glider (at various pitch settings), whether or not we have enough control authority to do a full-stall flare (and how long it takes to do that), sink rate, etc.

FLIGHT 3 - A LARGER MOTOR (I FORGOT WHAT KIND - BORIS DO YOU REMEMBER?)

On the last flight of the day we tried a somewhat larger motor, aiming to get more glide time and a chance to try manual yaw control.

This time I worked the radio controls (probably a mistake) and Boris ran the camera.

Ground video:
[YOUTUBE]KroPWqxD-Rg[/YOUTUBE]

On-board video:
[YOUTUBE]xYZeyCOT-oY[/YOUTUBE]

Again you can see the glider pitching up during boost, causing a premature arc-over and too high an airspeed at apogee. It's more pronounced here than in the prior 2 flights.

It's hard to tell from the ground video (Boris, you need to zoom in more!), but you can see that at about 0:13 on the on-board video, as soon as I switched on the manual yaw control, I massively over-controlled it and the glider went into a sharp left roll, and dived that way toward the ground, much like in flight 1.

Clearly the glider is VERY sensitive in roll (as predicted by some posters here).

For our next flights (November 5 is the schedule), I'm planning to make the following changes:

* Greatly reduce the control range in yaw/roll to make it less sensitive.

* Adjust glide pitch setting based on data analysis (TBD).

* Add a little bit of yaw command in boost to induce a slow roll to even out any net pitch command during boost.

* Reduce pitch (very slightly) in the boost position.

* Reduce parachute deployment altitude from 200' AGL to get more glide time. Maybe 100'? What do you think, Boris?

Maybe other things once I've had a look at the logged data. We'll see.

Cheers,

--Dave
 
Last edited:
Dave, yes, I think definitely you want a neutral boost position, or whatever will give you a straight boost. A very slight roll on boost as you say might negate any wind or offset boost.

I think also what might help is to put in a slower rate of activating the glide position, so it is gradual which will help slow it down without a sudden pitch up. Maybe a 2-5 second delay from boost to glide position.

When I'm flying my models, I keep a neutral boost, then pitch over slightly till I'm in a horizontal position, then activate my glide trim.

I think for you since you can't tell what position you are going to be in, pitching over won't be possible reliably but a gradual application of glide trim will have the same effect of slowing down into glide speed/position.

I personally think that your ejection altitude is correct and wouldn't go lower, I think with a straighter boost you'll have the glide time you want.

Hope that helps.

Frank
 
Last edited:
Yea!! You really had a *GREAT* one on Flight #2 there! Hmm, to any of you here who see me fly a lot, did this look kinda familiar?? :wink:

It's also neat you got the crowd to really applaud you as well, I've pretty much 'desensitized' the bunch around here. And at a big ole CMASS launch, no less.

Seeing a larger glider boost like that is just something different and cool. Seeing it glide back is a real treat!

That slight pitch up on boost, a bit of roll will 'fix' that if you cannot just 'dial' it out. The glider I flew on an I195 earlier this year did just that, probably due to asymmetric drag, but it sure glided just fine at ejection! I'll adjust it a bit and refly it sometime soon before attempting a J motor in it.

Thanks again for all your info on this most excellent project, and good luck on continued success!:)
 
Dave and I were pleased with the results of these first three flights of a very experimental project.

The glider still in great condition and we collected a lot of good data. The elegant custom electronics performed well as did the airframe.

The motors we used were:
flight 1: Roadrunner G80 (115Ns)
flight 2: CTI/AMW H54 White Longburn (168 Ns)
flight 3: CTI/AMW H100 Imax (286 Ns)

The glider only went a little over 400ft up on the first flight so we moved up to H power for flights 2 and 3.

Dave's autonomous control system is powered up by placing a small magnet near the right spot on the airframe. During liftoff the elevons stay flat, at apogee elevons lift modestly for a cool backwards roll into glide position. Then a flare and chute deployment at about 200 feet.

For the second flight we planned that I would slightly vary elevon position during glide to evaluate effect on pitch. All other aspects of the flight were automated and worked flawlessly.

The well controlled glide in flight two was exactly the result we were hoping for.
 
Last edited:
Dave,

Awesome job on the electronics.

I think all your plans for refining the control logic sound good.

The chute did seem to deploy very quickly.

Could we try deployment at 150ft on 11/5 and adjust from there?

Boris
 
Back
Top