Mars Lander Rocket Quadcopter Project

Discussion in 'The Watering Hole' started by georgegassaway, Apr 8, 2019.

Help Support The Rocketry Forum by donating:

  1. Apr 8, 2019 #1

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    In 2016, researching more Multicopter info for my Lunar Module*, I found out that Arducopter-based Flight Controllers had a new Multicopter flight mode called “Throw Mode”. You arm the multicopter on the ground, it beeps that it is ready, pick it up, and throw it upwards into the air. When it detects it has reached apogee, it turns on the motors and stabilizes in a hover.



    So, I got to thinking about doing a Rocket Boosted Multicopter. To take off using rocket power only, then at apogee it would automatically stabilize into a hover and then I could fly it by R/C for the rest of the flight. BTW - I have checked with experts on an Arducopter forum and they confirm it should work. I decided not to design it to fold the arms, a complication and risk factor I didn’t want to get into.

    I got to thinking about other aspects of the design shape, and then hit upon a great ultimate shape for it. Far more than “just” a Rocket Boosted Quad. But I’ll save the info on the final appearance for a bit later.

    I began on it in early 2017, but the project got bogged down, in large part due to the parts for the shape, and accessing those parts later. I was also in the middle of the Lunar Module Quad project, it was flying but didn’t have the covering and other things to make it look nice yet, so I shelved this project to concentrate on that. Since that time, I came across someone who could solve the shape aspect in a fantastic way, with 3D parts (I’ll mention him later). So earlier this year, I started gathering up the parts, revising the design, and collaborating with the 3D guru about the other parts.

    Finally started the build. The foundation is a 1/8” plywood disc to hold the graphite tube arms for the Quad motors, a “tower” assembly to hold nearly all of the electronics above it, and a rocket engine mount below it.

    [​IMG]

    The key electronic parts needed. From top right down, a GPS receiver. “PixRacer” Flight Controller (F.C.) running Arducopter software. I would have preferred to use a "Mini-APM" F.C. as I have with other quads, but the APM's can't run the more recent Arducopter software (not enough memory), and it's the more recent software that has Throw Mode.

    “Racerstar” brand 20 Amp 4-in-1 ESC (the single board controls four electric motors, 3 wires to each motor). And along the left side, Spektrum 4649 Serial receiver designed for use with Quadcopters. It uses a single bus wire to communicate to the F.C. , rather than having a separate signal wire for every channel.

    [​IMG]

    I temporarily put the tower assembly together, to check for fit, layout, and what mods to make. The tower uses a lot of 3mm nylon standoffs (blue and red parts). I do not have all the parts I need, for the lengths I want, and have an assortment on order that should arrive soon. In the pics, some legs are missing or not all secured, it'll all be redone and complete once the assortment arrives.

    The 4-in-1 ESC is attached directly to the plywood disc, with short nylon standoffs. Above that, and on separate standoffs, is an anti-vibration mount to hold the Flight Controller. The light blue rubber pieces between the mount plates help to reduce vibration, allowing the F.C. to work better. The F.C. is attached to the top of the anti-vibration mount, with short black standoffs.

    [​IMG]

    Above that, a 1/8” basswood ply plate, to mount the rest of the tower to hold the battery and GPS receiver.

    The battery is a 1300 mAh 3S (11.1 V) Lipo. It will allow for several minutes of flying. I may also get a 600-800 mAh Lipo to reduce weight for some of the rocket boosted flights.

    In the photo, the GPS receiver is right above the battery. But I’ll be adding a thin plywood disk in between, to physically protect the bottom of the GPS receiver and will add some thick aluminum tape to the disc to provide a bit of EMI shielding.

    I temporarily slid the 3D printed motor mounts in place to show the basic assembly and layout of the Quadcopter. So glad that the 3D guru was willing to make those, otherwise it would be a PITA to mount to the 1/4" graphite arms. (the noses come off to allow bolt access to attach the motors).

    [​IMG]

    Among the things missing in the images, are the engine mount tube underneath, receiver, assorted wiring, switch, and low voltage audio alarm.

    Also not shown yet, is the 3D printed shell that it will carry. And it's actually that 3D shell that transforms this project to something special. Keeping that for a surprise for a bit.

    Soon, I’ll begin working with programming the Flight Controller with an Arducopter application. The old borrowed PC I had used for my other quadcopters is dead, and I’m not sure yet what I’ll use as an alternative to run the software.

    I plan to test fly this in 3 weeks or so (the big steps will be how the F.C. programming goes, and other things related to the electronics). First flights will be purely as a quadcopter. Eventually I’ll try rocket boosts. I’m torn between doing a “throw” test as in the video, and just “going for it” with a rocket boost.

    * - Lunar Module Quadcopter thread: https://www.rocketryforum.com/threads/flying-r-c-lunar-module-quadcopter-project.137174/
     
    Last edited: Apr 9, 2019
    DaveW6DPS likes this.
  2. Apr 8, 2019 #2

    DaveW6DPS

    DaveW6DPS

    DaveW6DPS

    Well-Known Member

    Joined:
    May 28, 2013
    Messages:
    279
    Likes Received:
    121
    Gender:
    Male
    Location:
    Barstow, CA
    Following...
     
  3. Apr 9, 2019 #3

    dhbarr

    dhbarr

    dhbarr

    Amateur Professional

    Joined:
    Jan 30, 2016
    Messages:
    5,158
    Likes Received:
    375
    Any concerns about the props generating a substantial charge when under boost?
     
  4. Apr 9, 2019 #4

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    So now, the rest of the story. Because the project went from a technical design to something far more meaningful.

    When I was thinking about the design for the Rocket Boosted Quadcopter, originally it would be like a basic Quadcopter frame with a rocket engine in the belly and some fins added to help make it stable. Then I got to thinking of something more like a flying saucer with four arms sticking out. Hmm, like the old Centuri Flying Saucer?
    [​IMG]
    Well....interesting, but not interesting enough.

    Then got to thinking of some other old rocket designs and kits. And then it hit me like a bolt of lighting. PERFECT! Well, not easy, but this’ll be neat to fly even as a Quadcopter, the rocket boost being just the Cherry on top!

    A flying DeLorean! No, what I mean is….I’ll let Marty & Doc Brown explain it:
    [​IMG]
    [​IMG]
    [​IMG]
    (image above by SCIGS30 in the "Vintage Estes Mars Lander Build" Thread)
    https://www.rocketryforum.com/threads/vintage-estes-mars-lander-build.3349/

    Of course! The original model rocket “lander with legs”.

    It’s completing a circle. In later 1970 or early 1971, I got a Mars Lander kit. And was over my head. I had been in the hobby for a year or less. I wasn’t able to build it. I cannibalized the parts for some other stuff and that was the end of it.

    Probably lots who did build the Mars Lander played with it a bit, pretending to soft-land it onto the ground, moving it by hand. Well, with this, no pretending, no use of hands to soft land it, other than on the transmitter controls.

    So, I checked out a lot of forum postings to see how some built their models, usually the Semroc or Tango Papa kits, and some totally scratchbuilt models. I wanted to build it original size, no scale-up. So, I ordered a Semroc kit, figuring to use a lot of its parts such as the shrouds and nose cone. Started in early 2017.

    I knew the legs would be an issue. I wanted to be able to freely access the leg hinges and rubber bands or springs inside. And REAL hinges, with pivot points, not the original method with flexible tubing. Also it require take more force for the springs or bands due to the increased weight of the model. Well, I glued some of the shroud segments together and right off it was not going to go smoothly. Image below with a different quadcopter project to get a visual idea, without cutting slots into the tube

    [​IMG]

    And I kept trying mock-ups of the legs and springs and stuff, and it was talking a lot of time with little progress. I’d sort of hit a wall creatively and fatigued with the project, plus needed to get back to the Lunar Module which was still bare wood at the time. I realized I'd run out of time to complete both projects that spring. So I shelved the Mars Lander, planning to get back to it in late 2017 or early 2018. Uh, yeah. No.

    Later, I found out about the great 3D printed Mars Lander by Greg Glenn (mech-G) .
    https://www.thingiverse.com/thing:2670642

    [​IMG]
    WOW! I got in contact with him, and asked if he’d mind modifying a version of his 3D file to be used for a Quadcopter version. He was very willing, and has been an incredibly great person to collaborate with. Sometimes he’s a step ahead of me! Main changes were to eliminate the internal engine mount/baffle/recovery system parts as well as an internal mid bulkhead. And to make four openings in the Descent stage for the arms, while also moving the separation point down a bit. Also, he jumped all over making 3D printed engine mounts for the Quad’s motors, plus removable nose cones to allow access to the 2mm metric screws that hold the motors in place.

    Greg sent me a prototype set of parts last month. We both knew there’d probably be a need to make some tweaks later. And indeed there needs to be some, but not a lot. The legs work GREAT, slip into place easily, good hinge action. The rubber band system is great for the original, and easy to access and replace. Due to the weight of this model, the bands needed more leverage and Greg has modified the inner part of the legs where the hooks are, to get more leverage (I'm putting a BT-55 in it to hold the 24mm motor mount, so there's more internal space for the leg rubber band hooks to pivot thru than BT-60 allowed). A lot of the stuff that stymied me before, the legs and internal access and how to do it without messing up the shrouds, solved by using Greg's 3D printed parts!

    So a few days ago I finally started on the Quadcopter part of the model, checking that it would fit inside of the Mars Lander body. I was quite relived to find out that the “tower” would not extend the GPS receiver so high up that the upper ascent stage half would not fit down over everything. Well, some tweaking will be needed to assure the fit, but had no way to know till I test fit the parts.

    So, here’s some pics with the prototype 3D Mars Lander parts:

    View inside of the Descent stage, showing the legs and rubber bands, plus an inside view of the Ascent stage shell.

    [​IMG]

    The Quadcopter 1/8” plywood base plate and arms test-fit into the Descent Stage. Provisions will be made in the next version of the Descent stage to add mounts for blind nuts to secure the two assemblies to each other (Greg has a great idea to modify the upper rubber band hooks to double as mounts for the nuts). No glue, for easy access and for easy replacement of potentially damaged 3D parts.

    [​IMG]

    So here it is with the two shrouds and Quadcopter together for test-fit. Items not shown are the footpads, engine nozzle, and an on-off switch.

    [​IMG]
    The photo above may be misleading, that it looks like it's mostly done. But so much of it is test fitting. There's SO much still to do to get it flight worthy.

    I'll feel a LOT better once I have the Flight Controller set up, and controlling the 4 in 1 ESC to spin the motors in the correct directions (for safety, such "tabletop" testing is done without any props on). Also need for the GPS receiver to work right, the "Throw Mode" programming requires a good GPS lock so it can go into a stable hover that does not drift downwind.

    The model is intended to fly on D12 power. It won’t boost very high, but hopefully well enough (anyone who sez put an F or bigger in it…. feel free to send me the motors :) ). There is a good chance that it may veer to one side if all four motors and props do not “windmill” the same (they are free to rotate in the airflow like a pinwheel). If one motor/prop rotated at a different rotation rate than the other three, it would cause a drag imbalance. If it becomes a problem, I have an idea for addressing it, clear spin tabs on the legs which would make it roll on boost to even things out (spiraling ballistic boost), but not have a significant effect for Quadcopter flight control. In case it veers off horribly at liftoff, and got horizontal under thrust…… the Throw Mode software SHOULD activate to stabilize it into a hover as soon as it begins to lose altitude.

    Of course, most of its flying will be done as a Quadcopter only, at least 90% of the time I expect (recharging a battery pack costs PENNIES, rocket engines cost DOLLARS). Rocket boosts will be a special thing to do, especially at rocket launches. And I'll fly it at NARAM, with rocket boosts and also non-rocket quadcopter take-offs.

    I really wish I’d done this last year. I have good incentive to finally do it this year because of the 50th anniversary. Not Apollo-11, but this is also the 50th Anniversary of the Mars Lander.
    [​IMG]
     
    Last edited: Apr 9, 2019
    neil_w likes this.
  5. Apr 9, 2019 #5

    Charles_McG

    Charles_McG

    Charles_McG

    Ciderwright

    Joined:
    Sep 12, 2013
    Messages:
    1,774
    Likes Received:
    185
    Location:
    SE Wisconsin
    I was hoping for a Mars Lander.
     
  6. Apr 9, 2019 #6

    boatgeek

    boatgeek

    boatgeek

    Well-Known Member

    Joined:
    Dec 27, 2014
    Messages:
    1,657
    Likes Received:
    331
    If you are ever out in Seattle with it during our flying season, I'd gladly pop a CTI 24mm 3G F79 into it. :)
     
  7. Apr 9, 2019 #7

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    There can be things from the past may have had an influence, that you forgot about. Years ago when I had the most basic idea about doing an R/C Lunar Module Quadcopter, I was checking out various groups. Found this thread on R/C Groups, by Dlazarus6660, about a paper-based Lunar Module model rocket. He scaled up 1/48 paper patterns to build a 1/32 model (I also scaled 1/48 patterns for my 1/16 LM).
    https://www.rcgroups.com/forums/showthread.php?1110754-Lunar-Module-Paper-Rocket#post13105778
    He posted some photos. Here’s one.[​IMG]
    And the last one he posted was…. this.
    [​IMG]I was not thinking of that photo for this project. I saw this photo once again after I made the decision, then recalled seeing it before - "Hmmm". So, when I was trying to think of a neater design for a Rocket Boosted Quadcopter, perhaps I subconsciously recalled that image from previous quadcopter LM model info searching. Or not.

    Moon Lander Quad, meet Mars Lander Quad.
    [​IMG]
    I’m not saying this tops the Lunar Module, it’s just sitting there. :)

    .
     
    Last edited: Apr 9, 2019
  8. Apr 9, 2019 #8

    BA_Incognito

    BA_Incognito

    BA_Incognito

    Banned

    Joined:
    Jun 19, 2018
    Messages:
    160
    Likes Received:
    62
    Maybe you could make your Mars lander look like the Roton?? Body shape is similar....;);););)

    [​IMG]
     
    MattJL likes this.
  9. Apr 14, 2019 #9

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    Not much new bout the model to post now, but I wanted to include some images of the motor mounts that Greg Glenn made.
    [​IMG]

    He designed the nose cone to attach by using a bayonet method. In this image, you can see two bolts holding the motor in place temporarily, there will be four (note the two unused holes).
    [​IMG]

    So, this won't be the first rocket boosted (VTOL boost only) Quadcopter.
    [​IMG]

    The QuadRocket website: http://www.quad-rocket.com

    "Available now", but the website is closed for maintenance.


    There's also this, posted in Makezine.
    [​IMG]

    Article: https://makezine.com/2018/03/21/drone-rocket-hybrid-lands-vertically/

    Using parts from an inexpensive quadcopter, 3D printed hinged so the arms fold for rocket boost. Water powered mostly. He also did one model rocket boost, but it flew ballistically and crashed up in a tree.
     
  10. May 1, 2019 #10

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    After a few setbacks and delays, I finally flew the Mars Lander Quad Sunday. Wind was 15mph, so it was tricky to try to fly closer to the camera. Pretty much I pointed the camera (on tripod) downwind, so when the model moves away it is usually drifting downwind.

    It flew for over 6 minutes, on a 1300 mAh LiPo. Ended up landing off-camera, so ironically no video of the landing of the Mars Lander. I did another flying session, with several touch and go landings, but the camera battery went dead.

    Monday, the wind was lower.

    Look Vern, no parachute!
    [​IMG]
    This video has better flying footage than the other one.

    Some more photos.
    Version 1 of the 3D printed Mars Lander Quad body, at left. Greg Glenn (Mech-G) sent me a modified set for version 2, to solve some minor fit issues and some other stuff, which was used for the model on the right. [​IMG]
    To see where the "front" is, I added a vertical stripe of red tape. I'll remove that and replace with some suitable marking for the front, when I add decals and markings.
    [​IMG]
    One thing I do not like is the front orientation makes it a "+" quad, with the legs in an "X". So, the front is aligned with a motor/arm. All other Quads I have flown have used "X" orientation for the arms/motors, and for the Lunar Module the front leg with ladder was the front (of course). I plan to make an adapter so I can mount the Flight Controller (and GPS module with compass) rotated at 45 degrees, to make it into an "X" Quad, with one of the legs being the front.

    So, phase 1 of flight testing went well.... it works and flies as a Quad. There still some adjustments to do to the flight characteristics. And some things were rushed in a messy way to get it flying, that I'll go back and do more neatly. Also a new lighter weight voltage alarm is not as loud as the ones I have used before. I found that it sounds louder when inside of the Descent stage, due to the air gaps in the leg openings, than when inside of the Ascent stage shell. So I need to add an extension cable to run the voltage monitor there.
    [​IMG]
    But the biggest hardware issue right now is the GPS Module. It's not working correctly. But I think I found out why, the cable's wiring is not correct. I need to rewire the cable. Hopefully after doing that, the GPS module will work correctly. It is useful for doing Return To Home (RTH, also sometimes called RTL), and also for Loiter mode (model hovers in place without drifting downwind unless piloted to move, the smoothest way to fly). But the biggest reason for this model is that Throw Mode requires the GPS got have a good lock on the GPS Satellites. Since I want to sometimes launch the Mars Lander ballistically using a rocket motor only, then for Throw Mode to automatically start the motors and stabilize it into a hover once it reaches apogee, I need for Throw Mode to be working properly. So, I need GPS to be working properly.

    Rainy and windy for most of this week, so it'll be a few days before more flying. Hopefully, the next flights will have GPS working. And after some successful GPS flying....try Rocket Boost with Throw Mode.
     
    Steve Shannon likes this.
  11. May 1, 2019 #11

    Rktman

    Rktman

    Rktman

    Eric Noguchi

    Joined:
    Mar 5, 2017
    Messages:
    661
    Likes Received:
    22
    I'm on board.
     
  12. May 5, 2019 #12

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    Update - I made an adapter plate so I could rotate the Flight Controller 45 degrees. So now the “front” when I fly it is one of the legs, and the props are at 45 degrees, in an “X” configuration which I’m used to flying (such as the Lunar Module with the ladder leg as the front). Before then, it was in a “+” configuration with the legs at 45 degrees and one motor as the front. I also added some shims so the legs have a wider angle. So, a bit more stability for landing, and a lower CG to the ground.

    Flew a few times Saturday the 4th. Got the GoPro working, so I could follow the ML as it flew (camera mounted on a hat. Sometimes a bit shaky). This is a pretty good video of the second flight of the day.


    Wind was decent before I flew but kicked up a bit. Not bad, just not as calm as I’d like for trying to hover close to myself to get some good shots. I settled for some close fly-bys to get screenshots from. [​IMG]
    [​IMG]
    [​IMG]
    Bad news on the original GPS module. It’s dead. My previous quadcopter experience was with an APM-Mini Flight Controller, with a GPS module that was plug-and-play. I am using a different controller (Pixracer), and got Mini M8N GPS receiver. I plugged the cable into both, but got no GPS reception. I dug into the issue further, and found that the wiring pinouts were different, the supplied cable needed to be rewired. I rewired it and then tried it, nothing. I think I fried it from the first attempt with the original cable. I confirmed Friday night it was dead. After flying the ML Saturday morning, I spent part of the afternoon trying to solve the GPS problem. I got out an older GPS module, not as good but works good enough, sorta. And triple-checked the wiring pinouts as I made up a cable for it. It worked. But a bit on the edge.

    So, late Saturday before dark, I tried it. Took off, went into Loiter mode, which should have held it in place over one spot, with no drift. It immediately veered left and out, towards trees. By the time I switched back to Alt Hold (manual stabilized control), it had buzzed thru the branch of a tree, stabilized itself, but ended up at a horizontal velocity and descending slowly as it hit the ground and flipped over. Fortunately, no real damage. A leg got flung off, because it pivoted to the same angle used for attaching it, and the rubber bands that act as shock absorbers catapulted the leg away.

    Why did it veer when I tried loiter? It had GPS lock, but the magnetic compass had an error, causing an effect called “Toilet bowl”. Where it tried to hold GPS position but due to compass error it moves the wrong way, then realizes it, tries to turn to where it should be, but since the compass is not accurate it never can, and begins a circling that becomes more and more extreme. But I have never seen to go so bad so rapidly as that.

    After reinserting the leg and attaching the bands, I flew it again. I flew it normally for a lot of the flight, then flew it to a more open area, higher up, and tried loiter again. It worked! It stayed in place, I walked up to it and only used the throttle stick to adjust altitude and roll it a bit. Got some nice video of that, and held the transmitter up to show it wasn’t being controlled. Unfortunately, the GoPro’s SD card got full right then, and did not record the model’s best landing ever (thanks to GPS Loiter assist).[​IMG]

    I flew it one more time before dark. Good takeoff, then when I tried loiter, it toilet bowled again, not badly, but enough to go back to manual. It was getting dark, and as the ML was in its last minute of flight, close to me, the voltage alarm went off. The low lighting allowed first one, then two of the three red voltage low LED glow to be visible thru the 3D printed body. Landed OK.

    So, I merged two short clips from the successful loiter, and the low voltage alarm and landing, into a 2 minute video below.

    I’ll have another major update on Monday.


    .
     
    Marc_G likes this.
  13. May 7, 2019 #13

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    Sunday, I finally flew the Mars Lander on rocket power only.
    [​IMG]

    Drove a half mile down the road to a larger area to fly from (no nearby trees). Set the pad up, set cameras up, everything. Then it refused to arm. Packed up, and came home. Mission Planner showed some errors I’d never seen before. Finally worked them out, one was a parameter I’d set that not needed. Then some LOG error. Don’t know why, so I disabled the onboard flight log. It would arm then, but I sacrificed the “black box” flight logging for analyzing any flight problems.

    Drove back, set up again. Prepped model. And reprepped. Finally all cameras running, and Throw Mode giving its “Ready” 3 beep tone that it repeats over and over that it is ready to be thrown.
    [​IMG]

    Pressed the launch button, it took off, weathercocked a bit, reached apogee of 60-70 feet, started to fall… and finally the motors started and it stabilized into a hover - IT WORKED!
    [​IMG]

    I wanted to fly it again and only have one 850 LiPo right now, so I wanted to land and save power for a second launch. I wanted to land on the road, a car came by, got distracted, but it was OK. Sett up to land it but it was windy and I got distracted noticing the D12 casting was sliding out and fell out. I should have aborted the landing to start over, but didn’t. It had too much sideways drift when it landed, and fell over, breaking a prop. I didn’t have a spare prop, so it was repack everything (I was afraid to leave all the stuff along the side of the road as it could be stolen), back to the house, replace the prop, then go back and set up again.

    Another thing I realized later….. I was into too much “Go Fever” mode. Having had so many issues crop up to prevent flying it, and a misfire, I was way more into “launch this rocket” mode, and not enough “Flight plan for R/C model” mode. What I SHOULD have had in mind to do once it powered up into a hover, was to put it into “Loiter” mode, and confirm if it was successfully loitering (No “toilet bowl” death spiral). If the loitering was working well, the wind would not have been a problem at all, I would have steered it over the exact spot I wanted it to land and then reduced throttle, and it would have come down vertically with no drift. Also, I was so excited that it had worked, and thinking about landing it “soon” to save battery power for flight 2, I rushed the landing more than I needed to.

    And so, flight two. I had angled the rod slightly downwind to reduce weathercocking. It went well for the first 30 feet or so, then it veered HARD, got horizontal, pointed downwards about 30 degrees, and burned out. [​IMG]

    For whatever reason, the flight controller never started the motors, it crashed. BAD. Snapped one of the 6mm graphite arms, two others were pried off but I think can be reglued. Actually I’ve not even gone over it that closely since I picked up the pieces, put them in a box, and brought the box into the house. At some point i’ll take a photo of the carnage.

    I’ll rebuild it, but for now I’m not in the mood to go over it in detail. Several of the 3D Mars Lander parts will need to be replaced, but thanks to Mech-G who created it, that won’t require much work in the rebuild. Also, there was going to be a need for upgrading the body anyway. The key thing is the damage to the Quadcopter assembly. At least I do have replacement 6mm square graphite tubing for the arms, though only one seems to need total replacing. I do not know yet if the GPS receiver was damaged, or not. The Flight Computer was still merrily beeping away with error tones due to the crash, so at least it seems to be working (and being mounted to the anti-vibration mount, it didn’t receive as much physical shock on impact at a rigidly mounted board would).

    But I won’t REALLY know until I rebuild the Quadcopter, plug in the motors, and see if the motors all spin correctly (testing without props, of course).

    Now, WHAT happened? I theorized early in the project, this scenario. As the model’s vertical airspeed increased, at some point the props would begin to “windmill”. This causes much more drag than if the propeller is not rotating at all. So, what if one prop windmilled at a different rate than the other three? It would cause uneven drag, and the model would pitch or yaw on boost. I think THAT is what happened. In the various video camera views, it is clear the model did not go unstable (the CG was at or above the Quad arms, far more forward than on a normal Mars Lander). It veered HARD, then actually stopped veering before burnout. I think what happened is that at the moment it began to veer, that one prop started to windmill, causing a lot of drag towards the side it veered towards. Then after the veering began, I think other props finally began to windmill, equalizing the drag, which is why it sort of straightened out after it went past horizontal. Apparently on flight one, the four props mush have begun to windmill at about the same time, though maybe one began to windmill just before the others, so flight 1 may not have been 100% weathercocking into the wind.

    I would love to solve it by folding props, but that is not practical. Need pairs of CW and CCW rotating props, and have not been able to find any that are a close enough match. Folding the arms, too much hassle and overcomplication I don’t care to get into. A leading idea that I have to do a lot of proof of concept tinkering with would be to have the four props secured to each arm by an elastic thread or rubber band, with the other end secured to an external rod that would be retracted by a servo to release the elastic or rubber band ends to allow them to release the blades to let them rotate (Hey, kinda Rotaroc technology). It would have to be perfectly reliable in bench tests before I’d use it, since a quadcopter can’t fly on 3 props. If I did do that, then I’d need to do some active control using the transmitter for launch, to flip a switch to release the elastic/bands after burnout, before Throw Mode starts up the motors.

    So, Sunday was quite a day. A lot of frustration in multiple attempts to fly.

    The exhilaration of the first flight.

    The “oops” of the landing.

    And the crash.

    A few phrases come to mind, one from ABC’s Wide World of Sports: “The thrill of victory, and the agony of defeat”.

    And as regards the crash, Ben Roberto’s all-too-true philosophy: “If you’re gonna fly, you’re gonna crash”.

    So, here is a video of both flights, with three camera angles. This is why it's good to document special/new/experimental models with cameras as much as practical. Not just for when they work, more for when they don't.

    .
     
    Last edited: May 7, 2019
  14. May 7, 2019 #14

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    OK, I got the model out and took some photos.

    All four arms are damaged (two are split lengthwise) and will need to be replaced.

    [​IMG]

    Something significant to add, about launching this from a pad. I realized there MIGHT be some effect on the GPS receiver due to a steel launch rod running right next to it. Also, that the steel rod might affect the compass sensor. Fortunately, I had a 3/16" graphite rod 4 feet long to use for the rod, so I used that. And yeah, the original Mars Lander uses a 1/8" rod, but I intended this from the start to use a 3/16" rod for more stiffness and better launch. Also this model weighs over a pound at liftoff, more than triple the LO mass of the original.
     
    Last edited: May 7, 2019
  15. May 7, 2019 #15

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    [​IMG]
    Reconnected the motor that separated, powered up the electronics, and did an R/C test (props removed). The model responded, the motors ran fine. The GPS receiver lights up, but I didn't take it outside to see if it's receiving satellites. Even if that old 6M receiver is OK, I think I will order another M8N GPS receiver which should receive more satellites (including GLONASS), and may have a better compass.

    So, it seems like the main damage from the crash is the need for replacing all four arms. Hopefully I can remove the remains of the glue, otherwise i'll make a new 1/8" plywood plate. Since the square graphite tubing split so easily, and was at risk of doing so in a quadcopter-only bad landing, I'll wrap some 3/4 oz glass cloth around the new ones to resist splitting. Photo below showing after and before.
    [​IMG]
     
  16. May 7, 2019 #16

    Leo

    Leo

    Leo

    Well-Known Member

    Joined:
    Aug 19, 2009
    Messages:
    1,737
    Likes Received:
    38
    George, fantastic project!
     
  17. May 7, 2019 #17

    timbucktoo

    timbucktoo

    timbucktoo

    Well-Known Member Staff Member Global Mod

    Joined:
    Jun 13, 2014
    Messages:
    5,342
    Likes Received:
    363
    Gender:
    Male
    Location:
    Cocoa Beach
    Cool project!
     
  18. May 12, 2019 #18

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    Have installed the servo for holding the props from rotating on boost. Rubber bands will prevent rotation. The servo retracts two rods (two bands on each rod), that will let the bands be flung free. I have a pushbutton on the transmitter that activates the release.



    Hope to have the rebuilt model flying as a Quad Monday if weather allows, and try rocket boosts by next weekend.
     
  19. May 12, 2019 #19

    BABAR

    BABAR

    BABAR

    Builds Rockets for NASA TRF Supporter TRF Lifetime Supporter

    Joined:
    Aug 27, 2011
    Messages:
    3,549
    Likes Received:
    211
    Very cool.
    I was watching your first test video and at 1:30 and 2:45 I was almost shouting

    “Don’t go into the pond! don’t go into the pond!”

    Which is of course exactly what would have happened had I been flying it.
     
  20. May 12, 2019 #20

    Heli Hacker

    Heli Hacker

    Heli Hacker

    Member

    Joined:
    Jan 15, 2010
    Messages:
    8
    Likes Received:
    1
    It's my understanding that some ESC's have a braking feature that makes a dead short across the motor and keeps it from turning.

    I guess that would require that the Pixracer software activate the braking during Throw Mode to be of any use in this situation.
     
  21. May 12, 2019 #21

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    20+ years ago I set "prop brake" on the ESC's of some electric sailplanes to make sure the props did not windmill. Though those were old "dumb" ESC's that had to be set by some sequences of swtiches and throttle movement with the transmitter.

    The Racerstar 4 in 1 ESC uses BL Heli firmware ( A VERY popular ESC firmware, if not #1 now), which can be programmed via a computer, by hooking up the model's Flight Computer via USB cable, and using BL Heli's "configurator" software to set things like prop brake. I tried it but can't get it to successfully connect via USB (USB does connect fine with the two types of software I use for programming the PixRacer flight controller, so it's not the hardware). Even downloaded and installed the Mac-related driver files to enable it. I'll try again, but I needed to have something workable now.

    Also I am not sure at all if prop brake would work with throw mode. I did find that some Muilticopter fliers are using "damping mode", which is a sort of version of prop brake that instantly changes the prop RPM when commanded, rather than allowing the motor/prop's inertia to slow down naturally when the RPM's are commanded to decrease. The good side effect is crisper response (good for lightning-speed racers, not me), and instant shutdown on landing. Bad side effects are increased current draw (reduced flight time), and the ESC's getting hotter, possibly to the point of burning out. So I'm a bit leery of it, but did want to at least check it out if I can get the USB connection problem solved.
     
    Last edited: May 13, 2019
  22. May 17, 2019 at 1:32 AM #22

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    I repaired the Mars Lander and flew it Wednesday.
    The Original Quadcopter's 1/8" plywood plate was replaced with a new one. The arms were changed from the sort of fragile 6mm square ones to stronger round ones of about .29" diameter.
    [​IMG]
    Mech-G (Greg Glenn) modified the file for the electric motor mounts from a square hole to a round hole and a bit thicker. I wanted to try to fly this soon (weather will suck the next week), so Ray King (who lives about an hour away) printed the new mounts and nose cones, I drove to his place to get them.
    Some signifincant upgrades - Replaced the 6M GPS receiver with a M8N mini GPS receiver by Holybro. So far, it has worked very well. Good GPS lock, and good compass so it can loiter reliably and do the other things that GPS is useful for on Multicopters.
    I revised the wiring, as the original wiring was piecemeal. Biggest thing, no power switch, just plug in the battery. Which was very frustrating when I needed to shut off and reboot the Flight Computer, needing to remove the Ascent stage shell to do that. Now it has a switch at the base, near a leg mount.

    The Ascent Stage shroud was pieced back together.... but I didn't find all the pieces. Mech-G is working up version 3 of the 3D printed model parts, which I won't have for a bit but also I didn't want to RISK those parts till after at least one more rocket boost.
    [​IMG]
    I finally got it in flyable shape by Tuesday night, after dark. I happened to wake up about 20 minutes early Wednesday morning, so I quickly got dressed and took it outside to test fly it as a quadcopter. It flew well, so I knew I could plan to do a rocket boost after work.

    Shortly before 8 PM, I drove to where I boosted it before and set up. Had it ready, cameras rolling, when a small plane flew by, so I had to wait. Probably a good thing, before the plane left, the Controller did a unique set of beeps that it does when it has been too long after power up (I need to check into the parameters list and see if I can change that from say 5 minutes to 50 minutes).

    So, I had to turn the model off, then back on. It took some time for it to beep it was fully ready, I tried the rubber band release servo (to keep props from rotating) and it did nothing until after the controller indicated it was ready. Well, makes sense - the receiver only has one bus wire for all the channel signals, not a bunch of wires for every single channel. So it's the controller (Pixhawk, running Arducopter) that is decoding the signals, including the signal to the rubber band release servo. I confirmed the servo worked, by pressing the button and all 4 bands flung away.

    That is when the video begins, the moment I get the beeps the Pixhawk is ready, and press the release button on the transmitter. Then I hook up a new set of bands, and got ready for launch.
    [​IMG]
    It took off nicely on an E15, and weathercocked a bit. Not as much as the GoPro camera makes it seem to be, it has a very wide view that is almost fish-eye style (I have not been able to set it to a more normal view yet). At burnout, I pressed the button and said "button, button, button". And that's when I heard the "Zzzzwhirrr" sonud of the quadcopter motors starting up, and the model stabilized upright.

    So, it WORKED! Again. The rubber band system apparently worked well, no signs in uneven windmill drag (at some point I will add a video camera to a dowel above the nose, to look down, and see what happens on boost). The GPS worked well. I used Loiter a lot, and also used Return to Home a few times to fly it back closer. Unfortunately, the wide angle view does not show a lot of the flying, I could see it by eye but the camera didn't show much till it got closer (say within 100 feet).
    [​IMG]
    I finally brought it in closer, and eventually prepared to land it on the road. It's a country road with little traffic, so I looked to be sure it was clear before landing. I used loiter to assist in making a smooth landing on the road. As it was landing.... the Voltage Alarm beeper went off! Was kinda funny. But if it had not gone off, I was going to take it off again for a minute or so and land again.

    So, it flew VERY well. E15 is the way to go with it. Coming up sometime in a couple of weeks, I'll have the new version 3 parts and add decals to make it look better.

    And so, the video:
     
    Last edited: May 17, 2019 at 11:13 AM
    MECO likes this.
  23. May 17, 2019 at 2:44 AM #23

    neil_w

    neil_w

    neil_w

    Three Points Rocket Foundry TRF Supporter

    Joined:
    Jul 14, 2015
    Messages:
    5,236
    Likes Received:
    501
    Location:
    Northern NJ
    Very cool George!
     
  24. May 17, 2019 at 5:22 AM #24

    burkefj

    burkefj

    burkefj

    Forum Supporter TRF Supporter

    Joined:
    Jan 27, 2009
    Messages:
    2,701
    Likes Received:
    94
  25. May 18, 2019 at 12:03 AM #25

    Heli Hacker

    Heli Hacker

    Heli Hacker

    Member

    Joined:
    Jan 15, 2010
    Messages:
    8
    Likes Received:
    1
    Version 3 components are on their way.... - Mech-G

    upload_2019-5-17_18-1-0.png
     

    Attached Files:

    georgegassaway likes this.
  26. May 18, 2019 at 3:32 AM #26

    georgegassaway

    georgegassaway

    georgegassaway

    Lifetime Supporter TRF Lifetime Supporter

    Joined:
    Jan 18, 2009
    Messages:
    4,179
    Likes Received:
    135
    Oh, nice photo of all the version 3 upgrade parts!

    Yes folks, Heli Hacker / Mech-G is the 3D guru who's helped SO much with the Mars Lander part of the project, and modifying the parts to make the model more practical. The parts in that pic will replace the damaged parts, plus the upgrade features (including dummy nozzle)

    Also, a belated shout-out to Ray King, who lives across town, who 3D printed a new set of motor mounts, from a new set of Mech-G's files, when I rebuilt the model with larger round arms in place of the square arms. I drove to Ray's place Sunday night to get those parts in time to be able to test fly the rebuilt model this week, before the wind got worse after Wednesday.

    The MASA launch was postponed to next Saturday the 25th. So I'm planning more rocket boosting for then, and hopefully someone can get some better (closer-up) video and liftoff photos. Also, if I have time, apply decals.
     
    Last edited: May 18, 2019 at 3:40 AM

Share This Page