Glider with GPS guided recovery

Discussion in 'Rocket Boosted Gliders' started by delta22, Jun 12, 2011.

  1. Jun 12, 2011 #1

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    Fellow CMASS flier Dave Lindbergh has been doing advanced electronics work and field tests in the pursuit of GPS guided rocket recovery.

    Details at: http://nerdfever.com/?page_id=695

    So far Dave's flights have used a steerable parachute controlled by a servo, logic unit and GPS.

    Over the last several years, there have been some successes in the pursuit of this very ambitious project. However, the complex and fast moving dynamics associated with steering a parachute continue to be a very tough challenge.

    We discussed the possibility of replacing the steerable parachute with glide control surfaces on the rocket.

    Dave invited me to collaborate with him on this idea and I immediately agreed.
     
  2. Jun 12, 2011 #2

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    The rocket is a delta wing designed for fast gliding.

    The flight profile we want to use is:

    1) straight up liftoff from normal rail

    2) near apogee, control unit (with input from GPS and altimeter) adjusts elevons to glide position and guides rocket on a fast glide back towards the pad

    3) at 100-200 ft altitude over the pad, elevons are raised to flare rocket and a conventional chute is deployed

    This rocket will be based on a LOC 3.1" diameter body tube and nose cone, and fins/wings will be 1/4" balsa with reinforcement at the edges.

    I built a 32% scale of this for testing.

    Attached are RockSim files of both these projects.

    View attachment GPS3.rkt

    View attachment GPS3small.rkt
     
  3. Jun 12, 2011 #3

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    Several pictures of the small test prototype.

    Built with LOC 1" body tube and 1/8" balsa fins/wings.

    Elevons are straight for liftoff and lifted for glide recovery.

    Parachute deploys conventionally by separating nose cone from body tube. However, a Kevlar line on the outside of the rocket connects chute to BACK of rocket body. This is to prevent the control surfaces from striking the ground during recovery. Line can be seen in last picture.

    IMG_0305.jpg

    IMG_0307.jpg

    IMG_0308.jpg

    IMG_0310.jpg

    IMG_0313.jpg
     
  4. Jun 12, 2011 #4

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    This design is intended for straight up liftoff from standard rail and a FAST glide recovery (think drogue replacement), followed by chute recovery.

    The rocket needs to glide fast enough to easily cut through the wind and guide itself back to the pad.

    As a result, planned wing loading is at the heavy end of the normal range for an RC plane.

    Attached are a couple screen shots of calculated values for our proposed full scale glider, from this web page: http://adamone.rchomepage.com/cg_calc.htm

    Design started by examining the very cool Delta Star rocket glider found here: http://www.randrmodelaircraft.com/Delta Star.htm Then design was modified for build simplicity and a fast glide.

    aero1.jpg

    wl1.jpg
     
  5. Jun 12, 2011 #5

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    Wing Volume Loading takes into account the fact that volume and weight scale with a cube factor and area scales with a square factor.

    It allows for meaningful comparison a wing loading regardless of scale. It accurately scales from small models to full sized airplanes.

    WVL is defined as (weight of model in ounces) divided by (wing area in square feet raised to the 1.5 power)

    Some typical values:
    sailplane 4-6
    aerobatic 8-10
    sport 10-11
    heavy warbird 11-13

    Plugging in values for this project:

    350 sq in wing area / 144 sq in / sq ft = 2.43 sq ft
    2.43 to the power 1.5 = 3.79
    48 oz rocket weight / 3.79 = 12.7 WVL

    Another definition of WVL uses the area and thickness of the wing instead of raising area to the 1.5 power. This method is not used here because these wings are flat.
     
  6. Jun 12, 2011 #6

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    Any and all input is welcome.

    Questions:

    > I made the model with about 5 degrees of dihedral, is this a good value?

    > Elevon control surfaces are each 10" wide (width of wing) and 3" long, is this a reasonable size?
     
  7. Jun 12, 2011 #7

    Wingarcher

    Wingarcher

    Wingarcher

    Sponsor TRF Sponsor

    Joined:
    Jan 18, 2009
    Messages:
    637
    Likes Received:
    0
    For a fast gliding airplane like this, those elevons are HUGE. Full width is good, but they don't need to be nearly so wide/thick/deep whatever the word is. The width will make it very twitchy as the trailing edge is so far from the hinge.

    Compare to the delta star / gamma star designs. The full throw (up for instance, from neutral) on my gamma star is about 1/4 inch at the trailing edge. While gliding slowly a full up elevator will stall the plane immediately. At landing you can work to slowly apply elevator, ending up in full stall and a "plop" into the grass.

    Will a vertical flight result in enough lateral distance for the GPS to get away from the RTL coordinates? The google maps I've seen produced by GPS autopilot tracks show very large circles around a point. The error in the GPS requires you to have a fairly broad buffer zone around a point or the autopilot will end up chasing the point all over the place.

    N
     
  8. Jun 13, 2011 #8

    hornet driver

    hornet driver

    hornet driver

    Well-Known Member

    Joined:
    May 2, 2011
    Messages:
    4,636
    Likes Received:
    0
    Sounds to me like you've thought this thing through. I'm not super tech savy so I won't address your gps recovery. That said, I think your on the right track. Your dihedral is probably a good value. Lets call it like it is --it's a good round number to work off of-esp with a delta. I've done a few delta's and while fast they are stable and forgiving for the most part. A couple things to watch out for ---it sounds like you know this--watch your CG. It looks like your carrying a bit of payload in the nose, remember your flying the wing not the control surfaces, they are just there for adjustment.The concern is not the size of the control surfaces but thier deflection. You want to fly the nose up --not twist it up or muscle it up. Accelerated stalls are no fun and its part of the territory with high wing loads, perhaps some kind of progressive deflection would help.Remember its a delta so speed is life--don't bleed it.I think you've got a great thing going here just don't rush it. By the way I know the whole CP-CG challenge and the delta is the way to go if you don't mind the speed---but then there's always the forward swept swing wing:rolleyes: Keep this posted I'm interested
     
    Last edited: Jun 13, 2011
  9. Jun 13, 2011 #9

    Wingarcher

    Wingarcher

    Wingarcher

    Sponsor TRF Sponsor

    Joined:
    Jan 18, 2009
    Messages:
    637
    Likes Received:
    0
    Hornetdriver makes a good point re control surface size. It's true, it IS the amount of deflection that matters. Not really measured in inches at the trailing edge, but in the total area that is deflected. A large and deep control surface will present a lot of area with very little trailing edge movement. Servos and the simple linkages used are limited in their resolution so it may not be possible to move a large area a super small amount.

    Find a hill and do some toss glides with your scale model to see how it flies. This will give you and idea of where your CG is/should be. A properly trimmed RC model will fly fairly well with the radio off.

    N
     
  10. Jun 15, 2011 #10

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    Thanks for the input gentlemen.

    I did a glide test three times with the small model by tossing it fast from a second floor window, much to my wife's amusement. It dropped 15" and traveled 30" horizontally each time maintaining a fairly steady attitude. The NC stuck firmly in the soft ground each time but the rocket is not damaged.

    Thought I had secured the nose weight adequately, but later found it was about 3/4" too far forward. The elevons (1" wide in this 1/3 scale model) were deflected 1/2" up.

    I realize in this position they are attempting to fight the nose up. This flight profile is halfway between a glide and a controlled fall.

    Will repeat this test adjusting weight position.

    This is the logic I used proposing large elevons, please let me know if some assumptions may not be accurate.

    The Delta Star has elevons that are about 1" wide and 15" long on each wing. It weighs about 24 oz and is 36" long.

    The 3" GPS rocket is proposed to have wings with a 10" span, weigh about 48 oz and be 48" long.

    Starting with the Delta's 1" elevon width
    x 3/2 (smaller GPS wing span)
    x 4/3 (longer GPS rocket)
    = 2"
    plus some more to adjust for weight and high wing loading.
     
  11. Jul 30, 2011 #11

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    Completed a series of "glide" tests.

    Rocket preparation:

    1) Tweaked RockSim model to exactly match the pictured small glider.

    2) Marked calculated Cp and desired Cg for a stability margin of 1.0 on rocket.

    3) Adjusted internal weight in small glider to match desired Cg and wing volume loading expected for larger 3.1" model.

    Test preparation:

    Repeatedly threw small glider from second floor window as previously described. All flights dropped 15ft to lawn and traveled roughly 30ft forward.

    Due to adjustments noted above and careful verification of actual Cg before each "glide", Cg was about an inch further back than the glide tests described in an earlier post.

    As seen in pictures the elevons were held in position with tape.

    What was varied in the test series was the width of the elevons. Between each round of glide tests, I trimmed 1/8" off the width of the elevons.

    Test results:

    1) Elevons 1" wide (in this small model, scales to 3.1" wide in larger GPS model). Threw rocket 3x, each time the rocket flared the nose up, then landed on its belly.

    2) Elevons 7/8" wide, 2x throws, very slight flare.

    3) Elevons 3/4" wide, 2x throws, steady attitude throughout flight.

    4) Elevons 5/8" wide, 2x throws, steady attitude throughout flight.

    5) Elevons 1/2" wide, 3x throws, rocket dropped its nose as flight progressed.

    This result suggests that for this small glider set up as noted, elevon widths in the 5/8" to 3/4" range worked best.

    Scaling this up by 3.1x for the GPS glider, this suggests elevon widths of about 1.9" to 2.3".

    This value is in line with the rough calculation noted in the previous post and moves in the direction suggested by Wingarcher.

    Picture is of glider after final trim and last glide test. I am pleased that it is still in good shape after plunging it's nose into the lawn well over a dozen times.

    IMG_2167.jpg
     
    Last edited: Aug 13, 2011
  12. Aug 1, 2011 #12

    jms

    jms

    jms

    Well-Known Member

    Joined:
    Jun 5, 2011
    Messages:
    298
    Likes Received:
    0
    Very cool build! What kind of electronics are you using? What kind of GPS module is the final build going to use?
     
  13. Aug 1, 2011 #13

    Adrian A

    Adrian A

    Adrian A

    Well-Known Member

    Joined:
    Jan 22, 2009
    Messages:
    2,105
    Likes Received:
    5
    Interesting project. I have an ambition to do something similar, but with forward-stowed wings that would swing out. I know a guy who cuts RC wing foam cores, and I'm working on some prototypes of RC aircraft electronics.

    Are you thinking about doing anything to shift the CG back for glide mode? If you have the CG far enough forward for good rocket stability, you'll need an awful lot of elevon deflection (and associated drag) to keep the pitch controllable.
     
  14. Aug 9, 2011 #14

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    I'm doing the easy part, building the rocket glider, so I am not familiar with the details of the electronics.

    Dave has a lot of information posted at: http://nerdfever.com/?page_id=695

    I know he is planning to use the same electronics he has been using in the guided parachute flights.
     
  15. Aug 9, 2011 #15

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    I started with a R&R Aircraft Delta rocket glider as a starting reference point.

    As it is an excellent flier with a delta wing and no Cg or Cp shift when transitioning from rocket boost to recovery glide, I'm hoping to accomplish something similar.

    For this project, a fast moderately steep recovery is desirable.
     
  16. Aug 15, 2011 #16

    mr_matt_II

    mr_matt_II

    mr_matt_II

    Well-Known Member

    Joined:
    Jul 26, 2010
    Messages:
    159
    Likes Received:
    0
    I have a gamma star (smaller version of the delta star) and while I have not flown it I know you are going to get a pretty big forward shift in Cg between the start of boost and burn out. No idea how this affects flight qualities yet, after I fly her I will give a report.

    You guys might want to look at the JPADS program for inspiration!

    http://faculty.nps.edu/oayakime/ADS...ADS Programs Overview and NATO Activities.pdf
     
    Last edited by a moderator: Aug 15, 2011
  17. Aug 15, 2011 #17

    Dave92F1

    Dave92F1

    Dave92F1

    Active Member

    Joined:
    Dec 28, 2010
    Messages:
    42
    Likes Received:
    0
    I'm using the GlobalTop PA6B GPS. Here's a link:

    http://www.gtop-tech.com/jsf/module...puid=3fd4de25b0dd3bf98c210cb99566759047d54dec

    I'm getting a blog posting ready for NerdFever.com with the latest status on the project. I had some good flights in July and learned a lot. It turns out my main problem has been tangling/twisting of the parachute control lines - I'm going to have to redesign the attachment of the payload bay to the parachute to fix that.

    But that problem won't affect the rocket glider that Boris is building - the idea of the rocket glider is to test out the GPS steering system without the complications of the parachute deployment and tangling issues.

    --Dave
     
    Last edited: Aug 15, 2011
  18. Aug 16, 2011 #18

    dlb

    dlb

    dlb

    Resident Crazy

    Joined:
    Jan 24, 2009
    Messages:
    1,141
    Likes Received:
    0
    Gender:
    Male
    Love the designs, I seen something like it before.

    Go's up nice and glides like a cement brick, was very hard to control.
    RIP!

    Black_Wolf_Glider_rear.jpg

    Black_Wolf_Glider_side.jpg
     
    Last edited: Aug 16, 2011
  19. Sep 5, 2011 #19

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    At the 8/13/11 CMASS launch in Acton, MA the small prototype glider flew twice on C11-3 motors.

    Both flights went up fast and arrow straight.

    I had intended to put the CG 1 caliber (in this case 1.0") in front of the CP. Due to the use of a lot of wadding, which pushed the parachutes and the CG slightly more forward, the actual CG was 1.25 caliber in front of the CP.

    This model is 1/3.1 scale of the planned GPS guided rocket. It did not have any active control, this was just a test lift off of stability.

    The small model's control surfaces were fixed in a flat position. Used "elevon" sizes that will scale to 2" wide in the 3.1x larger rocket.

    IMG_5730.jpg
     
  20. Sep 5, 2011 #20

    delta22

    delta22

    delta22

    Well-Known Member

    Joined:
    Jan 21, 2009
    Messages:
    1,015
    Likes Received:
    0
    In addition to some other preparations for the arrival of hurricane Irene last week, I put my larger rocket projects on top of tables.

    Normally they sit just off the floor of my basement, but the risk of loosing electrical power for days at the same time as heavy rainfall could have resulted in serious basement flooding.

    It takes two sump pumps and a dehumidifier to keep my 24' x 26' basement dry.

    These plus other rockets not shown here represent well over 500 hours of work.

    Fortunately we did not have heavy rain here in southern Massachusetts and our house only lost power for one night.

    IMG_5889.jpg
     
    Last edited: Sep 5, 2011

Share This Page