Servers by jDrones

Reference Designs [Expired]

referencedesign

(James Pattison) #1

Topic: Reference Designs

Proposal type: Hardware [x ] , Software [ ] , Other [ ] : _________________

Description: The designs/components in the wiki are quite dated, and often not complete/specific/definitive. I suggest either within the Dev group, or by holding a community competition, that reference designs for each supported vehicle type are defined. The build and setup wiki’s could then be very specific, and default params for the vehicle types could be made to support the reference designs. A potential additional benefit would be sale of build kits through the supporting stores, perhaps with a kickback to the project. A reference design update cycle (annual / bi-annual?) could be implemented to ensure that the designs remain contemporary over time.

Planned amount $$ (USD): unknown

Estimated time for completion: unknown


(David Buzz) #2

needs more information to be a complete proposal.


(James Pattison) #3

I didn’t want to solutioneer too much: more throw some ideas out and have them refined through discussion. What details do you want me to add?


(tridge) #4

I’d like there to be a reference quad at least. Might be something that jDrones could carry?
The Iris+ was the de-facto reference quad for quite a while. Maybe something of similar size?


(proficnc) #5

the Iris is being resurrected by the community. I have just received all the molds for the plastics from 3DR, and we have an updated main board with Pixhawk 2 in it.


(peterbarker) #6

Attributes required of a reference design:

  • relatively cheap (all-up budget of US$500 including batteries but NOT an RC transmitter)
  • repeatable (parts generally available worldwide)
  • robust (must be able to take a reasonable spill and not be a write-off)
  • replacement-parts (if a part dies a replacement should be easily available)
  • multi-source (parts should be available from multiple sources)
  • hackable (must be easily modifiable by the user for future expansion and playing)
  • Open (preference given to Open Hardware and components which run Open Source and Free Software)
  • sensibly sized (no trailer required for carrying it, but also see “hackable”, above)
  • expandable (related to “hackable”, above - the ability to add extra sensors, companion computers and the like)

(jpkh) #7

There is already one concept design that we are working on at jDrones to be as reference design. We try to make is at simplistic as possible and also as hackable as possible.


(skyscraper) #8

Well my only interest is in planes so the following applies to that

To @peterbarker s list, I wonder if should be added “easy to fly”, since a reference design is by nature likely to be used when you are starting out. For plane this tends to go against “robust” since flying wings are the most robust but can be more difficult to launch ( and if you can’t launch… ) , whereas an EasyStar ( or CloneX … cloneY … cloneZ) is easy to launch but not as simple in other areas ( fitting equipment etc)


(Gabriel DeVault) #9

Hi,

I am the designer of RotorBits sold through Hobby King. A modular “lego” or “mechano” type multirotor kit. I think these kits themselves would be excellent reference designs. You can even find compatible parts on thingiverse. I tried to work with HK to release rotorbits as open source but they don’t see the light.

I have a newer modular 'copter concept I am working on now. All 3d printable parts and round tubes. I am considering releasing this as an open source project. This could be expanded to include planes, and allow awesome “mix and mash” concepts combining parts from both.

Some pics from RotorBits project

RotorBits Printable Parts

RotorBits Example Builds

Any interest in this?


(proficnc) #10

And at the current runout prices… the Solo is perfect, with the community stepping up and building parts for it, it has a very long life ahead of it.


(James Pattison) #11

I think solo is great, but there is a bit of risk as until someone starts making bodies it isn’t very repairable. It’s almost the opposite of rotorbits - which is completely repairable but to be honest is (imho) a bit aesthetically lacking).
An “X” Iris is my suggestion (dead cats should be buried, not flown).
This builds on the body of work already done and would fly a bit better.
Thoughts?


(proficnc) #12

I actually now own all the moulds to Iris :slight_smile: so we can keep supplies going on that if wanted. Andy Jensen also has no shortage of Iris arms and other parts if people are interested.

the APUB group has Iris refresh plans coming, they involve a new main board that has the pixhawk 2 on it.

there are also heaps of second hand Iris’s around as well


(James Pattison) #13

Iris ‘refresh’ w/ ph2.1 has merit: simple integration of a companion computer is basically a must have for a reference design I think.
An X config would be better again…I’ll have a think


(James Pattison) #14

Thanks guys! It’s great that this is happening.
Has much consideration been given to performance requirements?
I think in order for a reference design to gain traction, it must satisfy two broad objectives:

  1. Be a suitable development platform, so developers and hackers can progress the project with a known, defined reference. Peter’s comments pretty much reflect a lot of what is probably needed there.
  2. Be desired by the community. I think some engagement would be useful in defining what that might mean. Should it carry a camera? What camera? A gimbal? What gimbal?
    Size? Endurance? Appearance? Configuration (quad? Hex? X? H? Deadcat?)

Ultimately, even if all the devs have a common baseline, if the community doesn’t embrace it then a lot of the benefits of having it can’t be realized.

For me, a 350 - 550 size quad, GoPro with a 3 axis gimbal, 25 min flight time, companion computer, backpack and air transportable, are things I’d like to see - basically it needs to do a bunch of what Phantom does, plus let me hack / try new things.
In Australia, getting all that whilst remaining under 2kg would be a bonus for a lot of people.
If Phoenix can get the endurance up a bit from Iris+ (Phil - I’m happy to help out with that), it could be a good option (I’m not a fan of the deadcat layout, but compromises have to happen somewhere I guess…).
Jani - is jDrones working with Phil, or is the work you’re doing leading to something different?

Regards,

James


(Luís Vale Gonçalves) #15

(Olivier Brousse) #16