PDA

View Full Version : A 360 degree GoPro brushless gimbal



chatch
12th June 2014, 07:01 AM
Purchase Here (http://www.shapeways.com/shops/realfpv?sort=pricelh&section=Gimbals&s=0)

For the past 6-8 months I have been brainstorming optimal configurations for the gimbal system of individual camera types. Here is my optimized brushless gimbal with a purpose driven design - specifically for Hero3 and SJ4000 cameras.

The mounting bracket pictured is the first available. It attaches directly to 400Q and 400Q-X frames. For the GoPro to be in the best position, just flip around the clean plate. Now the area where the Hero3 used to sit is sticking out the back. Pull out the battery a bit to adjust the CG to perfection. Or, use a bigger/longer battery.

The gimbal supports a wide range of motors while maintaining balance. I put a lot of effort into making sure the motors are always under the smallest load possible - this saves energy and produces less heat. The first major step I took was to switch the roll/pitch positions around. If you can figure out why I did this, great! If you can't, you'll just have to trust me that I worked the numbers. Maybe a little too much time was spent on this particular design, but I want it to be just right.

Personally, I want to be able to fly just as crazy as I usually do with the gimbal attached. This gimbal supports motors with a hollow center if you prefer slip rings for 360 degree rotation. Movement is not physically limited to a set range of motion.

This allows you to keep using your FPV quad airframe just as you were before, whether the gimbal is on or off. For extremely aggressive flying, you will need to make sure your wires are secured properly.

I've accurately and precisely modeled everything down to the exact FOV of the GoPro, checking to make sure props are never in view. So here it is. CCD mount optional add-on.

So how does the camera stay attached? The simplest and easiest way is to use a strong rubber band. There is a place on the top of the bracket to hook the rubber band around, and another on the bottom.



GoPro Hero3
Use 2208 motors or equivalent weight motor for perfect balance(10mm slip rings).

SJ4000
Use 2208 motors or equivalent.





http://i.imgur.com/mJKyNOn.png

http://i.imgur.com/zzGv7pS.png




http://i.imgur.com/ijmOome.png

http://i.imgur.com/HloZaZi.png

http://i.imgur.com/7uKG0do.png

chatch
12th June 2014, 07:25 AM
I forgot to mention the gimbal is 154 grams with the Hero3 attached. This is about 100g less than non optimized gimbals.

chatch
12th June 2014, 08:44 AM
I have two full units on a local 3D printer and I will be posting pictures, video and footage later today or this weekend.

chatch
12th June 2014, 11:04 AM
http://i.imgur.com/lcQdsIg.jpg

chatch
12th June 2014, 11:04 AM
http://i.imgur.com/M9zy9Zr.jpg

criccio
12th June 2014, 11:26 AM
Subbed!

chatch
12th June 2014, 12:18 PM
http://i.imgur.com/2h5gV43.jpg

chatch
12th June 2014, 01:30 PM
http://i.imgur.com/tUIdaaf.jpg

chatch
12th June 2014, 05:44 PM
http://i.imgur.com/lCFDLL2.jpg

chatch
16th June 2014, 03:39 PM
http://i.imgur.com/jVBTo6h.jpg

Ho-Jazz
16th June 2014, 10:15 PM
Want one...

chatch
16th June 2014, 10:28 PM
I'll be sending you one for free, just sit tight :)

flyboy321
18th June 2014, 04:29 PM
This is looking awesome Chris! After taking my statics course last year, this design actually makes more sense than current ones. If you have any prototypes you want to get rid of, I've got a long arm 400Q that could use a little love :D

Ho-Jazz
18th June 2014, 06:43 PM
I'll be sending you one for free, just sit tight :)

Schweet!

agarabaghi
24th June 2014, 10:15 AM
Gimbal is mounted, just waiting on my alexmos board to run some test. Chatch is in Europe right now flying his spanky around, so im not sure he will be very active for the next few weeks.

Here is a picture with the motors mounted:

http://imageshack.com/a/img834/1366/pq6wc.jpg

Im new to gimbal boards so hang tight... hopefully i can get it up and running and report back soon.

Badlands
29th June 2014, 09:04 AM
Love your design Chris - i tried a similar design a few months back & although it worked well the gimbal goes nuts when i pan down via RC control. Its definately the 'reverse' gimbal design because my other regular home-made gimbals work ok with the same electronics (AlexMoss).
Is it a firmware limitation?

Will be interesting to see how you get on.


4834148342

chatch
29th June 2014, 06:26 PM
i have a couple boards in the mail i'll experiment with when i get back from italy. i just posted it on shapeways for those that want it right away. i will be posting sample videos and a build video etc. later next week.

chatch
12th July 2014, 04:15 PM
It definitely seems like a software issue. It seems to skip a beat on pitch axis

Badlands
12th July 2014, 06:09 PM
So you are getting similar results?
On my 'regular' gimbals I can hold the rig & roll it by hand until im tired but on this design the pitch freaks out at 90 degrees?
Shame cos this design is lighter & stresses the motors less.

chatch
13th July 2014, 03:34 PM
False alarm... It works amazingly. I had a bad pitch motor


we went through 4 motors but now it is LOCKED IN! I'll be posting sample video after I update the design files to make slight balance adjustments

chatch
13th July 2014, 06:31 PM
updated the file on shapeways.

superdr34
13th July 2014, 09:27 PM
What is the distance between the upper and lower plates on your frames? Wondering if can adapt to taller frame with spacing.

What are you thinking to recommend for motors since the problem with the one?

chatch
14th July 2014, 12:43 AM
it's 35mm. I don't have any different motors in hand so all i can suggest is to buy from a retailer known for good support. RMRC has replaced puffed lipos that are over a year old for free. They would have no problem exchanging a bad motor for a good one.

When I get the gimbal working on a test frame(next week) I'll post videos, settings and tips. I will be getting 3D printed parts over to RMRC. At that point I will begin looking at conversion kits for other frames(TBS, QAV, etc.). I really think this is how Hero3 gimbals should look and perform. Larger cameras(DSLR) have the motors in reverse because the moments are completely different around each axis.

Also, cutting the clean plate is not necessary. Just rotate it around. Use the extended section for a larger/longer battery, or to correct the CG.

flyboy321
14th July 2014, 09:48 AM
Chris, what is the cost for one of the 3D printed kits? I would like to get one of these going on my long arm 400Q

chatch
14th July 2014, 05:44 PM
after a bit of tuning, no balancing needed. will post PIDs so it works out of the box when i'm happy


https://www.youtube.com/watch?v=QOrN53muSxs

Derrick
14th July 2014, 06:02 PM
Looking good, have you determined what is causing the slight vibrations, Frame or Gimbal Tuning?

chatch
14th July 2014, 06:06 PM
it's just the gimbal PID settings. it could also be the low quality print used for this test gimbal(not as strong). the official shapeways one is very solid. this was printed on a cheap makerbot :p

chatch
14th July 2014, 06:14 PM
http://www.youtube.com/watch?v=AnA7jKXeoSE

agarabaghi
15th July 2014, 09:36 AM
https://www.youtube.com/watch?v=C6g3_2lOokU

These setting are getting better and better. Im re printing the main arm to see if the slight vibrations are from the arm or settings. Will keep everyone posted.

Badlands
15th July 2014, 01:11 PM
http://www.youtube.com/watch?v=AnA7jKXeoSE


Well that's working well...yours doesn't freak out with extreme roll movements?
What gimbal board/firmware is driving it?

agarabaghi
15th July 2014, 03:00 PM
Alexmos 3 axis *only 2 hooked up*, with SimpleBGC 2.4 GUI / Firmware. Quanum 2208 motors

superdr34
15th July 2014, 08:40 PM
Any vendors going to be selling this? I know it is available on shapeways. Kit with motors would be cool.

.Joel
16th July 2014, 05:06 AM
Chatch, the ones RMRC are going to be carrying are you printing them or they are? I bought one of their GoPro lens protectors, and whilst it's functional they must of printed it on the shittiest/fastest settings they could find as it needed to be sanded down to even fit smoothly. Other than that it worked perfectly.

agarabaghi
16th July 2014, 08:07 AM
which lens protector did you get. I am looking for one.

.Joel
16th July 2014, 01:01 PM
This one: http://www.readymaderc.com/store/index.php?main_page=product_info&cPath=165_300&products_id=2225
TBS also offer something similar http://team-blacksheep.com/products/prod:jelloguard_h3

Badlands
16th July 2014, 02:59 PM
Alexmos 3 axis *only 2 hooked up*, with SimpleBGC 2.4 GUI / Firmware. Quanum 2208 motors


OK I'm running Alexmos 2 axis with SimpleBGC 2.4 GUI / Firmware. RCTimer 2804 motors - the gimbal works well BUT if I hold quad in my hand & 'roll' it this design of gimbal goes nuts on the pitch axis after 90 degrees.


Yours doesn't? Are you using 2 x IMU's?

agarabaghi
16th July 2014, 03:06 PM
nope. ill take a video of it in my hands tonight

superdr34
16th July 2014, 09:18 PM
How soon on RMRC stocking this you think? Pricing the same as Shapeways?

chatch
16th July 2014, 09:58 PM
it really depends on my supplier. i want the lowest price possible. i want it to retail for $49.99 but i'm not sure that can happen

Ho-Jazz
20th July 2014, 04:48 PM
Still want one...

superdr34
28th July 2014, 11:01 PM
Where you at with the gimbal and production?

chatch
29th July 2014, 12:35 AM
injection molding! but i am optimizing the parts first. i might have to split that main arm into 2 pieces.

superdr34
7th August 2014, 09:39 PM
You should make the arm that mounts to the frame able to accept some hex or round spacers. Then they could be cut to length to be able to fit any frame distance between the two plates? I would like to put one on a Disco LR. Maybe some mounting holes to mount the control board to also if there's room somehow? Just thinking out loud.

chatch
8th August 2014, 03:39 AM
i've been working out details on this, i might offer a dedicated disco mounting arm. at the moment i am tuning my 3d printer


http://www.youtube.com/watch?v=eaV1NqlTzVU

which will help a lot with this stuff. this one is printing ABS which is a lot stiffer. and once i'm done tuning i may be able to sell the gimbals a LOT cheaper. like $25 a set. i just have to work on print quality

superdr34
8th August 2014, 08:49 PM
If you make one for a Disco I would give it a try or one I can retrofit.

brianalexander
8th August 2014, 09:55 PM
Sounds bad ass. I'd love to add a gimble to my 400q-x

chatch
9th August 2014, 09:42 PM
yeah just give me a few days and ill print a bunch out. right now i'm just trying to get the print quality better

superdr34
6th September 2014, 11:40 PM
You done tuning that printer yet?

chatch
7th September 2014, 02:28 PM
Yep I have 5 gimbals printed.

Matty12
7th September 2014, 07:15 PM
I'm very interested too please Chris! Looks good!

brianalexander
9th September 2014, 05:58 PM
I'd like to be on the gimble list!

superdr34
9th September 2014, 08:26 PM
Count me in on one. I'll modify it to fit a Disco if I have too

Matty12
9th September 2014, 08:43 PM
Count me in on one. I'll modify it to fit a Disco if I have too

Here's a plan: Take your components off the Disco, put them on an ADS400q frame, install gimble, throw away disco ;)

Matty12
19th September 2014, 07:44 AM
Any chance of an update Chris?

superdr34
19th September 2014, 08:55 AM
Here's a plan: Take your components off the Disco, put them on an ADS400q frame, install gimble, throw away disco ;)

That would work except this Disco I'm running 15" props, 400kv and 6s trying to get the 20 plus minutes of flight time that's all. Otherwise I agree!

chatch
19th September 2014, 11:30 PM
hi, i just got back in town. will post updates shortly. i am working with superdr34.

chatch
19th September 2014, 11:32 PM
That would work except this Disco I'm running 15" props, 400kv and 6s trying to get the 20 plus minutes of flight time that's all. Otherwise I agree!

very soon you will have a 45 minute platform available from realfpv. the goal is 45 minutes but it may end up being around 32 minutes because of manufacturing limitations. i say 32 specifically because i've done the math and tried it out to verify.

gnrc
20th September 2014, 03:47 AM
very soon you will have a 45 minute platform available from realfpv. the goal is 45 minutes but it may end up being around 32 minutes because of manufacturing limitations. i say 32 specifically because i've done the math and tried it out to verify.

Upgradeable or completely new?
I should finally get my printed gimbal the next couple of days. It was kind if hard to find someone who was able to print it decently.

chatch
20th September 2014, 04:21 AM
It's completely different. 6 motors :D

superdr34
20th September 2014, 01:06 PM
Hmmmm. 30 minutes I'll take one.

gnrc
2nd October 2014, 08:12 AM
Got my gimbal today.
Sadly the screw patterns on my motors on the turning side do not match with the 3/4 hole pattern.
I am using dys 2208 Motors since they are hollow shaft :/

Guess I gotta poke a bit at the plastic now

viper522
2nd October 2014, 09:23 AM
I like everything except that the lens and IMU aren't centered on the roll axis. I've given up trying to adapt my TRD Beholder BLG to my older plate style ADS 400Q. I don't think this will fit that frame without a lopitoffame.

chatch
2nd October 2014, 03:10 PM
I don't think I've posted this yet, but I moved the IMU closer to center on the model I'm working with. As for the lens center, it's almost negligible

brianalexander
2nd October 2014, 03:43 PM
is there an expected release date?

Jrd
7th January 2015, 04:52 AM
Hi,

nice gimbal here, just a question: what is the benefit of this "reversed" design?
Thanks

chatch
7th January 2015, 05:23 AM
the moment around the pitch axis is greater, and the moment around the roll is reduced. they're closer to each other.

In traditional designs with larger cameras, the position of each roll and pitch motor makes sense in terms of moments about each axis. Whoever started the gimbal craze simply copied the idea without giving much thought to it.

here is a gimbal i put together for the Sony RX100. Notice the traditional motor locations -
http://i.imgur.com/PtQ7a0C.png

Jrd
7th January 2015, 06:21 AM
Cool, I've been trying to design some gimbal by myself and was wondering why all of them are like this.

One more question: someone mentioned here that IMU should be aligned with optics. Do you think it is also important that the motor axes go through the camera optics? Because it is pretty much what you get with traditional gimbal design (http://www.tarot-rc.com/images/TL68A08.jpg) and GoPro (but your roll axis goes rather through the camera center). Or this makes no difference?


Thanks, Jarda

chatch
7th January 2015, 06:30 AM
i don't agree that the optics should line up with the axis of rotation, unless the gimbal is designed so that it is balanced completely so that the CG is directly at the center of the lens. in the case of the GoPro, it's slightly to the left of the lens if you look at the face of the camera.

you certainly won't notice that the camera lens isn't at the center of rotation. The only time this could be a factor is for a camera with a telescopic lens. if you focused up close on an object far away while the gimbal changes its attitude, the angle of the camera lens would be maintained, but its position in space would translate slightly and you might notice the picture frame drift off center. in this extreme scenario you would practically need a telescope on the gimbal, and i think you would have to deal with deviations in space due to the copter(inaccurate altitude hold +- 1 foot, etc) rather than gimbal problems

you could balance a gopro gimbal so the rotation is at the center of the lens, but it would need some odd couterweight built in. i don't think any gimbal out there has this because it works fine without any of those considerations. you don't even need the gimbal in this thread - i've seen plenty of gimbals on planes work just great.

Jrd
7th January 2015, 06:43 AM
OK, seems legit :-)
But I like the possibility of mounting FPV camera aside of GoPro, so I am considering redrawing my own design to this.

Just one note. In multirotor design with front-mounted camera, one tries to move the camera forward enough (props out of view) while keeping as much mass as possible near the COG. Traditional gimbals have motors further back, so the copter can be balanced more easily (as well as the roll motor can be conveniently integrated into the frame (http://media.bestofmicro.com/J/L/460785/gallery/gimbal-stock_w_600.jpg))..

chatch
7th January 2015, 07:14 AM
all the extra dead weight is usually offset with a larger battery in the back. the most popular question I see asked with gimbal frames is "how large battery can i put on it?" - and putting all the extra dead weight up front seems like a smart move. the motors only weigh 40-60g i think.

Jrd
7th January 2015, 07:29 AM
If you want something as agile as possible, the motor position may play a considerable role.
But OK, cannot disagree with your opinion :-)

PS: looking forward to see the completed copter with the gimbal!

chatch
7th January 2015, 12:22 PM
For sure you want the motors as close in as possible for an agile copter. It only flies about 5 minutes with a standard battery though

Matty12
12th January 2015, 01:49 AM
Hey, is this gimbal available to buy anywhere yet, apart from print your own? Is it expensive from shapeways etc?

Edit: Check that, it's kinda hideously expensive :(

chatch
12th January 2015, 04:49 AM
It's really expensive on shapeways. I can print it myself for cheap, $15 for you guys. But let me take another look st the design, I've gotten a lot better since starting this project

Matty12
12th January 2015, 07:19 AM
Thanks Chris!

matn1
12th January 2015, 05:28 PM
Hey Chris,
I ordered the (prohibitively expensive) Shapeways version of this gimbal frame.
One issue I notice right away when fitting it on the ADS 400Q-X is that the front left motor blades actually end up hitting the gimbal :confused:. Either I've assembled the frame totally wrong (I did install 2 spacers on the dirty frame part - perhaps I wasn't supposed to use those?). Any idea what I should do to address this? It feels like I need to get some kind of half inch spacers to raise the motors up so that they clear the main gimbal arm. If anyone has a suggestion, fire away!

Cheers,
Mat

Edit: here are pictures of the issue
http://i.imgur.com/SBUeDgv.jpg
http://i.imgur.com/Af87ei3.jpg

chatch
12th January 2015, 07:41 PM
Hi, please make sure the arms of the quad are flush against the top dirty plate. This should fix it! Let me know how it turns out

Matty12
13th January 2015, 12:28 AM
What size props are you using Mat? Are your arms 'facing forward'?

chatch
13th January 2015, 01:24 AM
It looks like you're using 2212 motors. The frame/gimbal is for 2216 stator. I'll mail you a few new arms to try.

matn1
13th January 2015, 11:38 AM
Hey guys!
Raising the arm right up to the top dirty plate resolve the issue, albeit with a clearance of about 2mm from the gimbal. What didn't help is that this specific blade seems to have been warped downward. Unwrapping it (they are gemfan plastic) seems to make it fine.
I am using Lumineer 2208-80 motors actually. My arms are facing forward as I wanted to get away from the awful asymmetry (and the jumpy yaw issues stemming from it) from my previous TBS disco frame. I'm using 9 inch props on 980kv sunnyskys. I would have to angle the arms outward in order to fit 10 inch I suspect, but I had been running 9 inches on my TBS with those motors as that was the recommendation I had seen at the time. I suspect that rotational velocity centrifugal forces will likely align the props straight, but then again I ought to buy CF props for this guy seeing as I am flying a FC-for-dummies (naza) on it and am usually unlikely to crash it short of mechanical/power related issues.

I only raised a single one of the arms at this point. I will probably raise the others too, but I suspect it won't make much of a difference.

Doh: yeah those were the gimbal motors. The quad's motors are sunny sky 2212 980v...
I take it you mean I should use something like this: http://www.getfpv.com/fm2216-11-900kv.html ?

chatch
13th January 2015, 04:06 PM
if it fits, then that's great. I modeled the gimbal assuming everyone flies with 2216 1100kv motors. the 2212 980kv work too I guess. If the prop doesn't hit anything then you don't have to worry about changing motors.

Matty12
18th March 2015, 03:29 AM
I test flew the gimbal setup last night, using a cheap SBGC board.
The gimbal was working nicely, tuned adequately but when the quad is flying vibrates horribly. I added ferrites everywhere, moved the wires around and balanced my props but didn't really get anywhere. Finally I turned down the D values 25% in the FC (a Vector) and it seems to be able to hover with only a few slight shudders.
I know a few people were getting a lot of vibrations with the printed versions (Chatch kindly provided mine). Did anyone get better results with stiffer parts? (ahem! I mean stiffer 3d printed gimbal parts)

Matty12
19th March 2015, 10:28 PM
(still searching for a solution)

Matty12
24th March 2015, 12:43 AM
I was finally getting some good footage from the gimbal and it fell apart!!!

Sections of the 3d printed arm separated in flight, I guess it's back to the drawing board. I hope you get a chance to get some proper moulds made Chris!

viper522
24th March 2015, 08:16 AM
I was finally getting some good footage from the gimbal and it fell apart!!!

Sections of the 3d printed arm separated in flight, I guess it's back to the drawing board. I hope you get a chance to get some proper moulds made Chris!

Everything I've printed and put on a multirotor not in compression has come apart eventually. It's a bummer when it happens.

Matty12
27th April 2015, 08:42 PM
Changed over my gimbal to aluminium arms. It's working quite well now!
I'm not sure if this project is still a thing, I suspect not, it would be a great product though!61499

michelros
27th June 2015, 12:18 PM
Hello , thelast hours , I have read this post and I am impressed by the design! Really good innovative job!The last months I 'm looking for a gimbal which can be 360 ​​degrees ROLL axis . Have any of you havea video that shows this? I'm also curious about the board and softwarethat's used. I planning to build one next week so any help would be great. Thanks!