Hello FliteTesters!
I watched Josh and Eric's Naze32 video and I was very intrigued by his angled-arm design. I almost exclusively fly the Naze, so I all but ignored most of the config and tutorial, and instead focused on his miniquad. I had to have one! I ordered the kit right away, and made sure my stash of spare parts was ready for the hit. Problem was, there were no articles or any details on Eric's build aside from the Naze setup, and apparently nobody else has tried the angled approach yet. So here we are, I guess the burden falls to me! lol
As soon as everything arrived, I started to dig in. That's a Naze32 Acro in a 3D printed case, DYS 1806 2300kv motors, 12A Afro ESC's (obviously lol), and the FliteTest parts.
Step one, cut the arms
I went with 96mm arm segments based on photos of Eric's machine. From the "bend" outward, the arms extend far enough past the bracket to accomodate motor mounts, and that's about it. As you'll see later, I end up changing my mind about having the segments equal length.
Once the mounting screw holes were marked and drilled, it was time to attach the new angle brackets. Problem was, I knew the 1806 motors weren't going to be able to handle a ton of weight, and it looks like Eric used 2204's or 2206's, so I was already at a power disadvantage. Although FliteTest typically frowns upon it, I went for glue over screws. In a good crash these brackets are likely going to be the first part to break anyway, so what difference does it make if they're glued! Besides, upon some zoomed in inspection, it looks like Eric ditched screws too!
Next step, slap on some motors.
However, I ran into an issue. My first thought was to use the standard FliteTest motor mount discs like Eric and many of us have used and love. But, fun fact: while M3x20mm screws are very easy to come by, M2x20mm are nearly impossible to find near me. I sourced them on Amazon and they were ridiculous... in the area of $7 for 8. Forget that. Time to fire up the 3D printer!
A buddy of mine and I came up with this
2.5mm thick, 24mm diameter with a 5mm hole in the middle, and 2 2.5mm holes that are 16mm apart center to center. On the ends are 12mm wide, 5mm long protrustions with a little nub on the tips. As you can see, they fit like they were made for the 1806 motors. Go figure! (PS, I have the 3D file available for anyone who wants it. As soon as we get an account created, my buddy and I are going to stick it up on thingiverse too)
As you can see, I used the button head screws that came with the motor, rather than countersunk flush mount screws. I did this so the screw heads would grip tightly to the wood beam and prevent side to side wiggle. With those screw heads protruding and 2 40lb zip ties, these motors aren't going ANYWHERE!
Now that the arms are built, time to do some soldering on the Electrohub plate and get the basic assembly screwed together.
It was at this point that I paused, leaned back and said, dang, this thing is way bigger than I thought it'd be. I sat a ZMR250 frame on top, and it was totally dwarfed.
Then I did the same thing with a T4 250mm mini that we just scraped off the bed of the 3D printer, and it felt equally tiny.
Bottom line is, it's too big. This isn't a mini race quad, this a mid size copter with teeny motors. So, let's blow it apart and break out the saw again!
I decided to lop off 42mm from the inside segment, which brought the beam right up to the edge of the bracket.
Once the mount holes were re-drilled, and everything was reassembled, it was much more reasonable size.
As it sits now, the shaft to shaft diameter is in the neighborhood of 320mm. Not exactly a "mini racer", but who's judging lol!
Upon futher zoomed-in close up inspection of Eric's build, it looks like he took a similar approach. I can't tell if his inner segments are shorter than the outside, or if he just has the whole deal mounted further in, but his brackets are "inside" the Electrohub disc just like mine now are.
Now that the parts of what makes this build special are complete, I won't go through all the details of how to wire ESC's and motors, plug in a receiver or program the Naze32 board (THAT is what Josh and Eric's video was for in the first place!). The only thing that I will mention, however, is that I chose to stick to bullet connectors throughout vs hard soldering because of the "this thing is going to explode on impact" concept of these wooden quads. During a crash they're SUPPOSED to break apart and save your fancy electronics, and in that case, I want my electronics to disengage and fall apart rather than rip motor wires or tear off foil pads. It is far easier to plug in a few bullets and pull out another zip tie than it is to replace a whole ESC or motor. All that said, here's the colorful bits :-D
Non-FPV flight-ready with no battery, we're sitting pretty at 364 grams.
My "Eric Special" is built, the Naze is flashed and tuned and the Taranis is standing by. Now I just need some daylight when it isn't snowing to go play!!!
************Flight Review Time!!!*************
We FINALLY have good weather here in the Baltimore area. In fact, the sky couldn't be bluer, and it is 50 today! I took my little creation outside, it flew for about 10 seconds and crashed hard. I put the battery back on, flew for another 20 seconds, and I barely made it back in one piece. Quickly frustrated, I put this quad down and broke out the orange 250 shown above. It also barely took off and once again went top down onto a mountain of snow after freaking out and oscillating out of control. Ok, must not be me lol. Quick trip back home and I found some solutions.
First, there needed to be a better way to mount the battery on the bottom of the Electrohub. Since I'm flying a 1300mah 3S, there was no good spot to mount it so it'd stay in the center of CG and still be solidly retained. So, my miracle worker designer buddy made me this battery strap holder on his 3D printer, and it works like a champ! It'll hold anything from a 1000mah up to however big the velcro will go, and since it is a flat plate with doublestick 3M tape on the back, it can be mounted to anything... multirotors, foamies, the mailman, whatever!
PS, ignore the servo and rubber band down there... I may or may not have plans to drop some little green army men with parachutes ;-)
So, now that the battery problem is solved (for all future aircraft I build!), time to figure out why both quads went oscillation crazy, freaked out and ate snow. As it turns out, it was the PID controller I chose for my Naze32.
Back up a couple hours to the crashes: I'm running Cleanflight with firmware 1.7.2, and PID controller 1 (rewrite), because forums online said controller 1 was "better" than 0, the original (old) algorithm. In Angle mode, the quad is utterly unflyable. It oscillates wildly and unpredictably. Switch to Horizon mode, and it smooths right out, but does not autolevel. Switch to Acro, and it acts just like Horizon... smooth and stable, but no autolevel. I gained a little altitude, switched back to Angle and attempted an autotune sequence. The board tried to do its self tune, but instead punched the throttle by itself and darn near flew away, tick tocking back and forth the whole trip. I regained control by quickly turning off autotune, flicking into Acro, and coming home safely.
Fast forward back home: I switched the PID controller back to 0 like I've always flown before, went outside to the parking lot and all was well with the world! Angle acts like it should, Horizon is Angle plus flips/rolls, and Acro is all manual deliciousness. I clicked into Horizon, ran through an Autotune session, landed and locked it in.
Here are the default PID's right after flashing the board with the new firmware (ignore the controller 1, the PIDs didn't change when I switched back to 0):
And here is what we ended up with after it self tuned. Fairly dramatic, I must say:
Ok, so now that I have a good flash, a good baseline tune, time to go tear up some sky! I went back to the highschool parking lot where I started hours prior, and I have to say, this little angle arm is a dream to fly! My current config is a 3S 1300mah battery, 1806 motors and 5x3 2 blade props, and it hovers 2 or 3 clicks below 50% throttle. It has plenty of get up and go above mid stick, and does not hesitate to shoot upward when told. Next time out I'm going to experiment with 5x3x3, 5x4, and 6x4.5 props, as well as a 4S battery. I'm sure any of those changes will make this thing a total rocketship. As it stands right now though, it is certainly no slouch. This little machine with PIDs shown above will happily do single or even double flips in any direction (double flips drop about 50'-75' in altitude, so make sure you've got some height in your favor). It flys faster in a straight line than any other mini I've flown, likely because of the increased thrust from the rear blades, and it is sickeningly stable. No matter what I did to it, in any mode, it just said OK. Random gusts of wind had zero impact, and no matter what I tried, I could not get this thing to feel any worse than a feather floating calmly on a still pond. This is the first multirotor of any size that I've ever felt comfortable doing double flips, and the most comfortable I've ever felt flying in Acro/manual. Now all I need is to figure out the best placement and mounts for some FPV gear!!
Eric, you're a genius and you have a winner on your hands!
Here is a quick clip of me testing the craft out. Full disclosure: I can count on 2 hands how many times I've flown something this small, AND it was pretty windy above the trees. So, sorry about the wobbles lol. At least I didn't crash!!
Log In to reply
Log In to reply
Log In to reply
I do have an electro hub sitting in my parts basket......all i need is a power pack .....
What was your flite time ?
Thanks for sharing this story , very inspiring.
Log In to reply
Flight times with a 3S 1300mah battery have been in the 7-8ish range, including high punch-outs and flips. If you fly more conservatively, I bet you could get closer to 10, and if you fly much more aggressively, it will be more like 6. All in all, I think that's great for what it is!
The video I posted was a little over 5 minutes, and I had tons of battery left when I landed... but a bus full of kids showed up for sports practice, so I had to land quickly.
Log In to reply
Log In to reply
Log In to reply
Log In to reply
One:
On a traditional multirotor with all props pointing down, there is a certain amount of X and Y drift because the machine is sitting on an invisible flat plane. Consider it like an air hockey puck. However, when all the motors are facing out slightly and forcefully opposing each other, there is far greater stability. This is wonderful for camera platforms and anything that needs to stay rock solid and in one spot.
Two:
A normal multirotor can only angle itself approximately 40 degrees in any direction before you lose lift ability. At 45 degrees, lift and forward thrust are equal and you can no longer climb, and beyond 45 degrees, thrust overcomes lift and you start to drop. With this design, by the time the front blades are at 40 degrees, the rear blades are at 60, providing far greater forward thrust than any other design possibly could.
And three:
Has to do with high bank turns and preventing motors from shutting down while yawing... I highly recommend you watch Eric explain it, I just can't figure out how to put it into words lol
Log In to reply
Log In to reply
Log In to reply
I've been trying to come up with a good way of making a dihedral arm on a 3D printer too, but I can't come up with anything stiff enough. The best 3D printed arms I've ever used are for the "T4 quad" (the orange guy pictured above). They're tubular, therefor very strong, but don't lend themselves well to being printed "bent". The best idea I've come up with is to modify how the arms attach to the body, rather than the arms themselves. Its still in my head though, not in CAD or even close to being printed. I'll post here if it works! Another option is to replicate the wood brackets on the printer, but I really want to support Flite Test... buy from them first, then duplicate on your own.
Log In to reply
Log In to reply
Both videos were shot with a Mobius B lens (wide angle), that was stuck directly to the top plate with double sided foam tape (3M VHB tape), plus a fat zip tie wrapped around for extra security. It is as close as you can get to being hard-mounted with screws. That said, I think the vibration jello wasn't all that bad!
Log In to reply
Log In to reply
Log In to reply
Log In to reply
I haven't found any negatives for the angled motors so far, and I think it is a brilliant design. DJI and others have been doing this on large camera rigs for a while, but this is the first I've seen it on something so small. Technically, if you were to put the whole rig on a thrust meter, there would be *some* reduction in vertical push because the motors are slightly opposing each other and not all pointed in the same direction. However, I can't tell the difference lol
The brackets are 2.75mm thick, 56mm from center to end on the acute side, and 12.5mm tall. I don't have a protractor or angle finder handy, but my calibrated eyeball says it is about 10 degrees
Log In to reply
Log In to reply
Log In to reply
Log In to reply
The parts shown above are printed with 2 walls at 30% lattice shaped infill, and a whole tray of stuff took less than an hour to complete. Speed and layer height are set to "low". There are more refined numbers, but we haven't dug into that yet... Just the default "low" so far. Hope this info helps!
Log In to reply
Also, how much, would you say, you have spent on this? Just so I have an idea.
Thanks again!
Log In to reply
I'll start with the bad news and follow up with some good news. Or as Patton Oswalt says, we're gonna take a quick trip to Mordor, then I'll bring you right back to The Shire.
Despite Josh's desire to be cheap, the costs involved with getting into a multirotor can be somewhat intimidating. For this build, I used the following base parts: Electrohub kit $38, Angle Arm brackets $10, motors $10 each, ESC's $10 each, and Naze32 $25. Add in incidentals like the 3D printed parts I used, glue, zip ties, receiver, wire, bullet connectors, XT60 plug, and shrink wrap, and we're in the $200ish area. That doesn't include transmitter, battery, or battery charger.
Ok, now for some positivity lol. Although you just dropped a good chunk of change on this one aircraft, all of the parts can be reused. The Naze32, motors, and ESCs can run on any small multirotor frame or airplane, and the Electrohub has almost endless possibilities from a tricopter to Josh's dragonfly to an octocopter camera rig (and everything in between). Once you’ve been doing this long enough, you’ll eventually amass a good stash of spares. For this particular build, the only things I purchased were the Angle Arm brackets and the ESCs… I had everything else on hand already.
Can you go cheaper? Sure! You could get a ZMR250 frame and lesser electronics, but then you only have a miniquad that will never be anything other than a miniquad. The beauty of this system is that you can change it up whenever you want and have a nearly endless supply of new toys. The other important factor is repair costs. If you crash any other quad, replacement arms come from the internet. If you crash an Electrohub, replacement arms come from Home Depot.
Anyway, this was my very long winded way of saying yes, I would recommend this quad to anyone!
Log In to reply
Log In to reply
Log In to reply
https://www.youtube.com/watch?v=2z9XI9Q0ECg
Log In to reply
I hope to hear more about this... ;)
Log In to reply
Log In to reply
Log In to reply
Log In to reply
http://www.thingiverse.com/thing:765813
Log In to reply
Are you able to share the STL files for the mounts?
Log In to reply
Log In to reply
Great article man!
Log In to reply
Log In to reply