Replies to This Discussion

3D RoboticsPermalink Reply by Chris Anderson on October 27, 2009 at 4:44pm

John,

First, don't expect ArduPilot Mega until well into the new year. Second, it will be considerably more expensive than ArduPilot. Third, it won't necessarily work any better than the current ArduPilot for basic autopilot functions--it's just for people who want two-way telemetry and a more compact installation. Fourth, we'll continue to support the current ArduPilot. These are two different products, for two different customer classes.

We'll also be releasing a product that's a lower end of the market, an even simpler, cheaper autopilot for inherently stable aircraft. Also Q1.

ArduIMU is not part of the ArduPilot line. It's for people who want to integrate an IMU/AHRS into their own project.

The uBlox and adapter will be used in all of our autopilots.

Unless you're really trying to save money, I suggest you start with the current ArduPilot. It's mature, stable and will be supported as long as we can see. ArduPilot Mega will someday get there, but unless you're willing to deal with our teething pains for the next year, you'll be a lot happier with the current product.

• ▶ Reply

DeveloperPermalink Reply by Doug Weibel on October 27, 2009 at 5:12pm

Hey Chris,

Some input on Mega if the hardware isn't locked down yet. I hope you picked up on the common theme from Ryan Beale and Dan Goode (sp?). Using good quality gyros with on chip temperature measurement is really the way to go. Having worked with ArduIMU for a while now I think your comment that Mega may not work any better is almost guaranteed if the IMU is similar to ArduIMU. Bill P has done some wonderful work, but ArduIMU IS NOT a robust AHRS solution. No disrespect but I think Bill P's goal of using DCM for aerobatics is far fetched, unless it is used with a better sensor system. By that I mean decent temperature compensation so that you have some measure of stability in the IMU without relying heavily on compensation.

With a differentiated product line (ArduPilot Basic, ArduPilot, ArduPilot Mega) I think it is reasonable to go high end on the Mega hardware so that we have platform that can last through development of new ideas and several evolutions of software/algorithms. If you can get ArduPilot Basic w/ gps at say under $150, ArduPilot w/ gps at a bit over $200, I think Mega w/ gps at $300+ is in line.

• ▶ Reply

3D RoboticsPermalink Reply by Chris Anderson on October 27, 2009 at 5:28pm

Doug, good points and I agree with the sentiment. But we're not trying to compete with AttoPilot and above in IMU quality. Instead, we're aiming to be less than a quarter of the price and open source. We'll do the best we can with version one and we're using a modular design that will allow us to migrate to better sensors as they become available.

This is an evolving project. It won't be perfect on day one, but it will continue to get better, thanks to the community approach to development. Like Papararazzi, but "for the rest of us."

• ▶ Reply

AdminPermalink Reply by John C. on October 27, 2009 at 5:44pm

Thanks for the reply.

Teething pains don't bother me, but you hit on what is my current pain point. I want to be able to modify mission on the fly. I'm really interested in working on "On Point" camera tracking functionality as an add-on to the ArduPilot package.

I assume "considerably more expensive than ArduPilot" is still quite a bit less than the $800 attopilot, or more expensive Kestrel. I could probably get to half of that as a complete package with the right features.

I appreciate the efforts, and the update!

JC

• ▶ Reply

Permalink Reply by fred asiong on December 18, 2009 at 6:17pm

Chris,

I am very new to this ardupilot. Does it mean that ardupilot by itself will make my tri or quadcopter fly stable without the use of gps and 3 axis gyros ? Hoping for your kind answer.

• ▶ Reply

3D RoboticsPermalink Reply by Chris Anderson on December 18, 2009 at 9:24pm

No. Please spend some time reading more on this site. ArduPilot is not made for quad or tri-rotors.

• ▶ Reply

Permalink Reply by fred asiong on December 18, 2009 at 9:54pm

Thank you very. I thought this is like or better than the combination arduino 328/mega+5dof imu +idg300 that they used for aeroquad.(quadcopter). My intention on this electronic is for stability of flight nothing else.

Fredasiong

• ▶ Reply

Permalink Reply by Faisal Shah on October 28, 2009 at 9:56am

I would really really like to use the ArduPilot Mega, but the uncertainty in the time line is forcing me to design my own board. I intend to have a modular system, with the autopilot, AHRS/IMU, and magnetic sensors as separate boards (at least).

IMU separate so that it can be precisely placed as close as possible to the planes COG.
Magnetic sensor as a separate board so as to place it as far away from EMI sources.
GPS input
speed pressure sensors input
barometer input for altitude
input for thermopiles (x,y, and z)
Enough inputs for other sensors, range finder, etc ..

I'll probably get a panel of boards from GoldenPhoenix ... so I'll end up with a bunch of extras if anyone else is interested. I'll start another thread and ask for input from interested parties when I get some time.

I also want to use a flat layout for the boards (particularly the IMU), so I'm leaning to use the 6DOF Razor IMU from SparkFun - unless there is some way I can solder those lead-less packages at home easily. I'm a bit scared of diy stencil + solder paste + toaster over solder reflow method :) . Anyone here with experience in that? Is is necessary to get a stencil for the solder paste, or can i just put little dabs on the pads of the few components which i can't solder thru other means?

• ▶ Reply

Permalink Reply by Kim Skatun on November 12, 2009 at 2:13am

Well hopefully the new version will handle more than 4 servos.. At least 7 , preferably 9 would be great..
Different power connector as well would be great, the same one as eagle uses.. Then the board could meassure amps directly as well.
Add airspeed and altimeter aswell then to the shield so it can be used as vario, there are some glider pilots out here aswell.
Keep up the great work folks..

• ▶ Reply

DeveloperPermalink Reply by Jose Julio on November 13, 2009 at 7:32am

I think that only with one processor (atmega328 or 1280) we could handle all of the functions of an autopilot (IMU, Navigation, and RC signals/failsafe) and I am working on this line and now I have an atmega running IMU code, GPS parsing, stabilization assist mode, Radio signal reading and servo outputs. The "RC priority" function (manual mode) is totally implemented inside an interrupt so it´s safe (If main code hangs, manual mode is always working).
I´d like the idea of use an atMega 1280 because it has more serial ports but I think that with only one processor and a good optimized code it´s posible to do all this things and keep the hardware more simple (and cheap!).

JJ.

• ▶ Reply

Permalink Reply by anaximander on December 19, 2009 at 2:14pm

I'm seriously tempted to hold out for the Mega; the project I have in mind would seriously benefit from two-way comms and I don't think I have the skills to hash out a working comm system for the current ArduPilot on my own. One small question, though: what level of two-way comms would it be? Would it be limited to burst comms, for waypoint alterations and the like, or do you think it could handle real-time comms, for direct control to allow the pilot to dispense with the RC controller (assuming a suitable transmitter can be found)? The reason I ask is that I'm getting a little help from some friends on this project and they can't really fly RC planes with any real level of competence and to be fair, I'm not great either, and while we're doing our best to improve our skills it'd make us all feel much safer if we could simplify the job of flying somewhat. The ultimate goal is to put together a system controlled by a simple console-style two-axis thumbstick and a throttle slider, putting the control inputs through a stability augmentation program of some kind. If the Mega can handle that kind of real-time data streaming, then that'd be awesome. If not, then that's cool - I'll just have to go and read up a little so I can work something out myself. I'll do it someday; it's just that if the Mega can't handle real-time then it'll take a little longer :P

• ▶ Reply

DeveloperPermalink Reply by Doug Weibel on December 20, 2009 at 6:58pm

Nicholas - while the stability functionality will (almost certainly) be there, the idea of routing control through the two way comm will have no value in making it easier to fly. Stabilization will be a great help, but trying to fly with some poor thumbstick compared to a decent rc transmitter will just make the job harder. Also, keep in mind that "flying" is really easy compared to taking off and landing. That is where I would concentrate on making things easier...

• ▶ Reply

• ‹ Previous
• 1
• 2
• 3
• Next ›
• Page