Professional DIY Rig Project


Recommended Posts

So far I've resisted buying off the shelf UAVs in preference to building my own recreational rigs but thought what would it cost to build a truly professional multi rotor type UAV for aerial photography. I started to put together some thoughts on the bill of materials and would welcome any feedback or suggestions.
 
Frame. I would likely go with a hex carbon fiber frame. A quad is not redundant as if one motor or ESC fails it will fall out of the sky and it always surprises me that even the DJI Inspire and M200 are quads. With the right flight controller the Hex can continue to fly on 5 motors.
 
Flight Controller. Having had some good experience with the APM and Pixhawk hardware running arducoptor firmware I am inclined to choose the Pixhawk 2.1 for the flight controller. The "Cube" has built in redundancy as it incorporates  multiple sensors, can have 2 power sources and supports dual GPS/compasses.
 
 
Motors/Props and ESCs. http://store-en.tmotor.com seem to be the “Rolls-Royce” of motors for UAVS and they also sell ESCs and carbon props so this would be my first place to look. They publish very details specifications which should make it easier to work out what combination of motor/ESC and prop to select based on the total take off weight of the rig.
 
Camera/Gimbal. This one I’ll need some help with as I am not so familiar with some of the higher end cameras. All I know, at this early stage, is that its not a GoPro. The camera/gimbal would need to provide
 
  • control via the Pixhawk.
  • geo tagging of photos.
  • live video feed to ground station.
 
I really like the modular approach that these guys have designed in their latest surveying cameras.

https://www.mapir.camera/pages/kernel-cameras

A 4K version of one of these cameras to take regular video and high res photos would be great but I think they are just aimed at surveying for the moment. 

Link to post
Share on other sites
  • 1 month later...
On 6/25/2017 at 0:25 PM, Spitfire76 said:
So far I've resisted buying off the shelf UAVs in preference to building my own recreational rigs but thought what would it cost to build a truly professional multi rotor type UAV for aerial photography. I started to put together some thoughts on the bill of materials and would welcome any feedback or suggestions.
 
Frame. I would likely go with a hex carbon fiber frame. A quad is not redundant as if one motor or ESC fails it will fall out of the sky and it always surprises me that even the DJI Inspire and M200 are quads. With the right flight controller the Hex can continue to fly on 5 motors.
 
Flight Controller. Having had some good experience with the APM and Pixhawk hardware running arducoptor firmware I am inclined to choose the Pixhawk 2.1 for the flight controller. The "Cube" has built in redundancy as it incorporates  multiple sensors, can have 2 power sources and supports dual GPS/compasses.
 
 
Motors/Props and ESCs. http://store-en.tmotor.com seem to be the “Rolls-Royce” of motors for UAVS and they also sell ESCs and carbon props so this would be my first place to look. They publish very details specifications which should make it easier to work out what combination of motor/ESC and prop to select based on the total take off weight of the rig.
 
Camera/Gimbal. This one I’ll need some help with as I am not so familiar with some of the higher end cameras. All I know, at this early stage, is that its not a GoPro. The camera/gimbal would need to provide
 
  • control via the Pixhawk.
  • geo tagging of photos.
  • live video feed to ground station.
 
I really like the modular approach that these guys have designed in their latest surveying cameras.

https://www.mapir.camera/pages/kernel-cameras

A 4K version of one of these cameras to take regular video and high res photos would be great but I think they are just aimed at surveying for the moment. 

Building a professional rig is expensive, requires real engineering, and once you start down that hole there's no turning back.  We have been developing our own industrial drones, designing out the hobby grade and Chinese components  for more than three years.  We didn't really set out to manufacture our own drones, but for professional applications, off the shelf drones don't cut it.

Frame

Building in hardware redundancy starts with the frame.  We have six models of multirotors that can lift from 18-48 pounds AUW.  Anything larger than a quad is coaxial.  Hexacopter's are not generally hardware redundant.  The performance of our drones is measured at 50% throttle so if a Y6 or X8 has a motor failure and the drone hovers at 50% throttle or less, there's enough headroom on the opposite coaxial motor that you probably won't even notice there's a failure.  If it takes more than 60% throttle to hover and lose a motor you'll start to lose yaw authority.

The reason for this is that the thrust vector of the remaining motor is on the same axis of lift as the one that failed.  If you lose a motor on a Hexacopter, two motors on the opposite axis have to reduce thrust or the drone will flip.  So now you've essentially lost thrust from three motors. Depending on the payload, it can still fly but both yaw authority and total lift is greatly reduced.

The downside to coaxial motors is that there less efficient.  We have designed and manufacture our own coaxial motors [based on KDE motors], and by moving the props closer together we have gained some of that efficiency back but we have much more thrust than a DJI M600.

Also, our booms are detachable so the a drone that can carry 36 lbs AUW breaks down and fits into a case slightly larger than a Phantom and can be deployed much faster than an Inspire.  Depending on the commercial application how you ship and deploy can be very important.  

Flight Controllers

For flight controllers that cost under $4000, a PX4/Ardupilot based controller is the way to go.  We designed, manufactured and sold the X2, X2.1 and PixRacer.  We're no longer selling them but you can still get the PixRacer at https://store.mrobotics.io/  although I wouldn't recommend the PixRacer for larger commercial drones.  I'd recommend checking out this controller over the PH2.1.   https://emlid.com/navio/   The PH2.1 "Cube" was designed three years ago and I'm not sure how well they've been able to consistently ship?

Camera/Gimbal

Probably the best all around camera for professional aerial cinematography is the Panasonic GH4. Great form factor, lightweight, relatively inexpensive and shoots 4K video that can be edited together with much higher end camera's pretty seamlessly.  If you can afford about twice the price, ~$3500 then the Sony A7RII is great.  

Both of these camera's fly great on the Gremsy H3 gimbal.  With the GH4, the price for this setup would be about $3,900 and for that you'd have a gimbal that can easily be used handheld, on a jib, dolly, car mount or drone without having to recalibrate.  We use a Connex to stream 1080P 60 with no latency about 1300ft and we can control almost every aspect of the camera from the ground. Unlike the RoninMX which is heavy, temperamental and the horizon is off way more than it should be.

We also use a smaller gimbal and Sony a5100 for surveying.  

 

_D6A5329-Edit.jpg

_D6A5325.jpg

Y6 (1 of 1).jpg

Link to post
Share on other sites
On 6/27/2017 at 9:03 AM, Spitfire76 said:

From what I've seen from other posts in the forum this seems to be the company to get a decent professional gimbal from - 

https://gremsy.com

I need to do some more research on how well these gimbals interface to the Pixhawk 2.1 and Arducopter firmware.

 

You don't need to interface this gimbal with the flight controller.  It isn't necessary.   

006.jpg

Edited by Av8Chuck
Link to post
Share on other sites
10 hours ago, Av8Chuck said:

You don't need to interface this gimbal with the flight controller.  It isn't necessary.   

 

@Av8Chuck thanks for your feedback and photos of your UAV. I agree there is no need to connect the gimbal to the flight controller unless you want to be able to control it from a mission script or programmatically from a companion computer. My current setup below provides both gimbal pitch and pan control from my TX and from a mission script. I've also tested it by adding a Raspberry Pi companion computer and used a Dronekit API to set the pitch and pan angle of the gimbal.

Gimbal_001.thumb.jpeg.65cfb6c76f34b8ec78bf7c46b4e695da.jpeg

Link to post
Share on other sites

Your touching on kind of a philosophical difference between those that want the flight controller to manage all aspects of the platform and those that just want the FC to do "primary" flight control.

Whether the commands are from the RC/Tx, companion computer or GCS, I only want the flight controller to go where I tell it and stay there until I tell it otherwise.  Our "mission scripts" run on our companion computer.  

For aerial cinematography for example, we can plan or "record" a flight path, the drone will then fly the path as recorded and you can make changes to any of the parameters in realtime [and save those as well].   Then once you've setup your flight path you can "record" the flight of the camera/gimbal.  This enables several things, its makes it easy to setup very complex shots that are repeatable.  For timing, any of the parameters can be changed in realtime without effecting the path.  These paths can be explicit or they can be global, so once I have a mission that I know works for a given application I can simply set the start point and fly the mission anywhere.

We can also control [and "record"] every aspect of the camera in flight in realtime, exposure, focus, zoom, etc.. You won't be able to do that simply through AUX1/PWM from the Pixhawk.  This type of automation would require way more than 16 channels of PWM.

Link to post
Share on other sites
1 hour ago, Av8Chuck said:

Your touching on kind of a philosophical difference between those that want the flight controller to manage all aspects of the platform and those that just want the FC to do "primary" flight control.

Whether the commands are from the RC/Tx, companion computer or GCS, I only want the flight controller to go where I tell it and stay there until I tell it otherwise.  Our "mission scripts" run on our companion computer.  

For aerial cinematography for example, we can plan or "record" a flight path, the drone will then fly the path as recorded and you can make changes to any of the parameters in realtime [and save those as well].   Then once you've setup your flight path you can "record" the flight of the camera/gimbal.  This enables several things, its makes it easy to setup very complex shots that are repeatable.  For timing, any of the parameters can be changed in realtime without effecting the path.  These paths can be explicit or they can be global, so once I have a mission that I know works for a given application I can simply set the start point and fly the mission anywhere.

We can also control [and "record"] every aspect of the camera in flight in realtime, exposure, focus, zoom, etc.. You won't be able to do that simply through AUX1/PWM from the Pixhawk.  This type of automation would require way more than 16 channels of PWM.

I see your point. Having the flight controller handle just the flight control and a companion computer handle the gimbal and camera control does make a lot of sense. Also the ability to completely record a manual flight mission including camera and gimbal parameters and then be able to edit and playback that mission is really interesting. Is that companion software available or is that something you developed ?.  

Link to post
Share on other sites

That's something that we're developing.  I don't mind sharing our experiences, I'll probably produce a demo video of it flying missions in the next three weeks or so.  

I jumped into this conversation more because I think your headed in the right direction, not necessarily the same direction as us, but your putting in serious thought about what it takes to build a purpose built drone.  Once you start developing workflow for your application using a companion computer, the decision on which flight controller to use becomes largely irrelevant.      

Most people don't try to build a commercial drone, they just buy a DJI and think they've solved that problem.  If they're trying to use a DJI commercially their problems have just begun. You might never change anything about PX4/Ardupilot but you can change anything and everything about open source FC, Tx and companion computer, with DJI you can't change anything.

One platform enables innovation, the other is doing all it can to control it.  That will probably be a determining factor between success and failure of a commercial operator.  

  • Like 1
Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.