It's Gimbalquad

[OttoPilot]

So, one of the recent episodes with the FT270, combined with the FT homepage banner showing the FT270 rocking a 3 axis gimbal has inspired the heck out of me

I've decided to build a dedicated gimbal quad.

Not so much a quad with a gimbal attached to it...but more of a gimbal with a quad attached to it.


-Artists interpretation of what GimbalQuad might look like.

As the base I'm going to start with a 6" HyperLite Floss 2. These are what I fly for my race frames. They're easy to work on, durable, and most importantly light. I could cut my own custom frame...but the Floss 2 will be a great starting point.

I'll be starting with a 6" build using some F60 Pro 2207, 2200kv motors, and some spare DYS 30A ESCs. I can always expand to a 7" build if I don't get the performance I'm looking for...but a 6" will provide a good place to start, and has much more of a variety of parts and props available.

That brings me to the gimbal. I've purchased the Quanum FY Mini 3D PROS 3 Axis Gimbal. This things looks pretty robust (mostly aluminum...might be some opportunities to lighten it up) and stabilized the little session rock solid in the quick tests I did.

To make sure this is all going to work out and not be a colossal waste of time I took a look at some thrust and weight numbers.

A quick weight SWAG:
Gimbal + GoPro = 238.1g
Fully Built Floss 2 (5") = 292.0g
+3.6g difference per motor+longer arms (for the 6" build) = 15g
4s 1500 Lipo = 180g

Takeoff Weight = 725g (yikes).

Looking at data from MQTB the F60 Pros with HQ6x4.5 props will pump out 586g of thrust at 50% throttle (9 amps).

At 75% we're looking at 1093g at 21 amps so 4372g of thrust total...throw in 20% reduction for maneuvering budget (and because I don't really know what I'm doing) and we end up with:

~1875g thrust at mid throttle = 2.58 thrust to weight.
~3500g thrust at 3/4 throttle = 4.8 thrust to weight.

Compare that to my freestyle 5" Alien (600g w/ gopro, 2992g of thrust at 75% throttle) gives a 4.98 thrust to weight (without the 20% fudge factor thrown in above).

So, all of a sudden the 725g take off weight doesn't seem too bad. Obviously the massive increase in inertia won't allow for crazy acrobatics...but that's okay...that's not what this is about. This is about beautifully smooth, stabilized video that looks like it came from a Phantom...but done in a way a Phantom couldn't (or at least wouldn't) do.​

I think I can move forwards <mad scientist laugh here>

Next step is to get the frame ordered and figure out where I can shave weight off the gimbal (might try removing the aluminum mounting case and relocating the electronics to the FC stack). It'll all depend on how sensitive it all is to vibrations.

Sorry about the wall of text...trying to lay out my thought process so others and either duplicate it, or help me find the stuff I might be overlooking.

More to come soon.

-Tim

 

[Nathan_116]

Digging the revival of the gimbal quad on these racing style frames!!

 

[French]

Wow. Inspiring build. That's a lot of weight up high. Have you considered a top mounted battery and bottom mounted gimbal?

I hope you don't break something. That's a lot of weight in the middle and heavier motors on those tiny arms.

 

[OttoPilot]

Thanks Nathan, it should be an interesting experiment at the least.

French,

I did take a quick look at mounting the gimbal in the more traditional 'underslung' location but with the GoPro's field of view and the tilt at which I hope to fly at I'll be seeing more of the props and motors than I would like. Once I get her in the air it might be something I have to revisit.

And yes, with the extremely light weight frame I'm working with the assumption that I will be breaking things. This thing will absolutely not be able to take a hit like one of my race quads. I've even been considering mounting the gimbal to the quad using nylon bolts to act as fuse pins that will fail during a crash and separate the mass of the gimbal from the frame and vise-versa.

Did a little poking at the gimbal last night; it's looking promising.

I'll get some photos up tonight.

 

[French]

Sounds like you're thinking this through well. I like the idea of weaker gimbal mounts to break away. Best of luck!

 

[OttoPilot]

Well, tonights gimbal surgery went well.

A little disappointed that it looks like I'll be stuck using the stock mounting setup with 6 of the silicone bobbins...but it was fun to dig into how this thing works and to discover that the gimbal appears to have some hidden functionality (5v, gnd, and aux/video running to a picoblade connector hidden under the camera mount).


The gimbal and my 'ol reliable GoPro session. The bobbins mount through the six large holes in the flanges sticking out from the bottom.


Disassembled.

I was hoping to use the three screws (that hold the gimbal to the case) to mount it directly to a custom top plate I'll be making. Wires would be extended, and the 'control board' would be mounted to the top of the FC stack to isolate if from vibrations. This would have allowed me to get rid of the bulky lower case (21g w/ bobbins!) and keep everything looking clean.

Unfortunately the gyro for the gimbal isn't mounted in the lower control board...it's mounted all the way at the gimbal head (red arrow in the image below). This made complete sense to me once I saw it (no idea why I ever assumed the IMU was in the base...the processing and hardware required to do that would be an order of magnitude higher).

Unfortunately, that means that there's no good way to isolate the gyro from vibrations other than mounting the entire gimbal on the bobbins it came with. I might still experiment with the 'direct mount' I had originally planned but I think it's a lot higher risk given this configuration.



This was kind of cool to find though. A little picoblade connector (yellow arrow) was hidden and in-accessible underneath the camera mount at the head of the gimbal. A lot of times companies will use the same 'guts' of a higher priced product with more featurs/functionality in a lower cost product because it's cheaper for them to 'block off' the higher end features than it is to design a completely new configuration for the cheaper product.

Or at least that's what I was hoping for here.



Needless to say, I found this connector, notched out the back of the camera mount with my dremel so I could access it with the gimbal powered up (you need the GoPro installed to power on the gimbal or it freaks out due to the gains) and started probing all over the boards and connectors to see what functionality this connector had.

Sure enough I quickly discovered that the yellow and black connector listed as 'no effect' in the instruction manual (yellow circle) was directly connected to this connector along with ground, and regulated 5v power (not sure how much current I can pull on the 5v yet, but it will be useful no matter what). I also found that it would be almost effortless to run gimbal input voltage (7-17v) to this connector to power something that needs more than 5v.

Not sure how useful this will be for me...but considering the gimbal has pan/tilt controls I could see mounting a tiny whoop camera/vtx to the gopro to let a second person on the ground watch the gimbal shot or control pan/tilt. Not something I'll probably ever do, but cool to know I can if I want.

That's all for today. Still need to get the frame ordered.

More to come.

-Tim

 

[OttoPilot]

Well, despite Thanksgiving travel, an insane day-job schedule, and a ton of small biz. orders to fill:

Gimbalquad v1.0 is officially alive!

My top plate is currently 3d printed as a proof of concept (rather than wasting good carbon). Glad I printed the plate rather than going right to carbon since I found one measurement error (oops), and my first flight already has me wanting to redesign it in order to shift the gimbal further forward to get the motors out of the shot at lower tilt and bank angles (should be able to scoot it at least .25 inches further FWD, and any bit helps). Honestly it might stay printed for a while. The top plate doesn't see any major loads except in a crash...and this thing isn't designed to survive a crash anyways...we'll see.

As far as power goes: it has roughly what I predicted. Plenty of punch in a straight line, but you definitely feel the extra weight in the turns and have to steer it like a boat. I think some of that will fall out with some rate and PID tuning (stock KISS is surprisingly good on this yacht), but some of it will persist due to pure mass wanting to drag it all over the sky. It's definitely going to require a new approach to flying to get the shots I want out of it.

But enough of my yacking...glamour shots (forgive the filthy go-pro and bench).





Okay, maybe not enough yacking...

I also realized that now I have a whole different beast to tune: the gimbal itself. Going to need to dig into its settings to adjust 'follow' settings for certain axes to keep it from banging into itself and losing tracking. I also need to set up the KISS FC to allow mode settings or pan/tilt on the gimbal. It's currently wired...just needs programmed. So we'll see how well that works once I have a chance to play with it.

I promise...I'm done...here's the first flight footage.

First Outdoor Flight (just quick line of sight hop):

First FPV:

aaaaaand what happens when you do a flip out of habit and your gimbal locks, and loses sync, then re-syncs to a different 'down'.

You get some cool, unintentional, artsy shots:

It's been a fun project...but plenty more to come; especially optimizing the top plate layout, cutting it from carbon, and getting the gimbal set up properly with KISS FC pass-throughs.

 

[PsyBorg]

Getting in to this thread a little late seeings you already have this in the air. But the video semi confirms what I thought might happen. Quads when set up correctly ignore for the most part top mounted and bottom slung weight as long as the controllers are at the true cog and the distance to the motors are the same. Look up the Ted talks videos where they balance a broom stick on a quad as well as a full glass of champagne and do flips and rolls with out the objects shifting at all. The over all weight plays a huge part but that is only a matter of power management.

My first thoughts when I read you were going to use a floss frame was that the arms would be flexing like crazy under higher G turns and from the video it looks like I may be correct. As you noticed the quad flew well on stock pids which is kind of what you want on a AP ship. You don't want super snappy movements you want loose flowy changes but still be manageable to make the turns needed to get a shot. You want to lean more towards higher I gains and D gains (keeping aware of motor temps) then tight P gains.

I would serious keep those arms in the back of your mind as they will in my opinion be more a problem then pids with all that extra weight. Stiffer and longer arms would greatly benefit this build and I think bring it closer to perfection. Over all though great start and congrats on a successful first few flights. They show a lot and can be learned from for the next round of testing.

 

[OttoPilot]

Thanks!

Out of curiosity what in the video points towards arm flex? I didn't pick up any of the ordinary high frequency oscillations under load that I would have expected to see; especially with the longer arms on the 6".

I also agree, longer arms would definitely help out, unfortunately they would also move the motors further out into the gimbal'd camera's FOV which i'm really trying to avoid. I might wholly re-design the main plates of the frame in the end. Just not sure where this rig will sit in my hangar yet.

 

[Ocean]

I conclude with OttoPilot as well, I didn't see any flex. I would be very surprised if the carbon flexed more than a degree, if much at all. Even if they do flex I doubt it would present any issues. You would still 98% of thrust.

On my cinetank, the thrust angle is off by perhaps 10 degrees or so. There is no noticeable difference in flight.

Nice work, OttoPilot. I love that people are putting gimbals on acro quads now, there's lots of potential to film action sports I think.

 

[PsyBorg]

Ehh just erased what I wrote.. It sounds too argumentative over something that is perceptual. AS long as Otto is happy with how his gear behaves then that is all that matters. Happy Holidays gents.