75 replies to this topic

[neuron uav]

It would be very interesting to test the NetDuino be able to work with a lot of concurrent processes to mims time. Each of these processes have to communicate. It sure is not able to handle many processes. At the end some of them will be blocked. A not it?

[Dan Vanderboom]

Count me in to join a Netduino Quadricopters project. I'm looking to develop a commercial product, but would be happy to build it upon an open source hardware/software dev platform, and as such I have many free cycles to contribute. I've followed the very interesting conversation here. I'm not yet familiar with the demands of flight control, but I can see how a pure managed code solution could present a problem on such a resource-constrained device. As a long-time software engineer (25 years), I know that .NET code can be made to run quite fast; however, as I understand it, NETMF on Netduino runs interpreted which presents additional performance problems. Chris Walker - it was a pleasure to meet you at PDC10, and I'm excited to see your new products rolling out! Not sure you remember me amidst the crowd of excited devs there, but I showed you videos on my phone of some two-wheeled bots running around on my kitchen table (and not falling off). I'm interested to learn how to deploy and debug native code and call from interop on Netduino, and to see what kind of performance we could get out of this. If the main control loop runs native, all other code should be able to run managed (guessing, of course).  Does anyone have experience using the Critical Velocity gyro/accelerometer shield with Netduino? http://www.criticalvelocity.com/item.php?itemid=shield6 I'm not familiar with what makes an Arduino shield electrically compatible with Netduino or not, and I hesitate to buy Arduino shields because of comments I've read about having to switch pins and so on. Can anything be done in the future to reduce or eliminate this disparity? Is a Netduino driver available for this gyro/accel shield or a similar shield? Can anyone comment on how to evaluate the suitability of a gyro or accelerometer for quadricopter flight control, or will any of them suffice? Looking forward to working on this with some of you, doing a lot of fun experimentation, and generating a lot more buzz for Netduino, NETMF, and the UAV DIY community!

[Chris Walker]

Not sure you remember me amidst the crowd of excited devs there, but I showed you videos on my phone of some two-wheeled bots running around on my kitchen table (and not falling off).

Hi Dan, good to meet you again. [waving hands]

Does anyone have experience using the Critical Velocity gyro/accelerometer shield with Netduino? http://www.criticalv...?itemid=shield6

I just glanced at the datasheets, and it looks like all of them output 3.3V signals--so electrically they should be fine. [And even if they output 5V digital signals...you'd still be fine. It's just >3.3V analog signals that you should avoid.]

That said, it looks like the product has been discontinued.

On the broader shield compatibility question:
1. Netduino's analog pins read signals from 0-3.3V, not 0-5V like the AVR used on Arduino. If a shield outputs analog signals, it may need to be modified slightly to output less voltage. Not usually an issue.
2. Netduino's digital pins _are_ 5V tolerant--so electrically pretty much all Arduino shield should work.
3. You will need a driver for some Arduino shields (either C# or C/C++ code). As a corollary, there are a lot of native C drivers written for Arduino (not using Arduino's Processing language)...they just need to be ported to 32-bit ARM.

Welcome to the Netduino community, Dan.

Chris

[Brandon G]

Hi dan, we've made some great progress in the last week. Our gc is running once every 10 seconds and is a very fast sweep, thats with telemetry logging via my custom serialization class. We will be moving source to eith GITHub or codeplex in the next few days. I have no personal experience with the shield given that we are starting simple only using 3 degrees to get off the ground, we want to be stable with only a gyro, then we can get fancy later. our main thread can be found here http://forums.netdui...pter-for-netmf/ we are also going to start posting on rcgroups once we finally get some hardware in and show our first stable flight via video

[Dan Vanderboom]

@Chris, Okay, so as long as I avoid analog inputs >3.3v and find a native C/++ driver, I should be good. Good to know, thanks! @Brandon, I'm interested to take a look at your code, once you do make it available, and will be happy to contribute and test. I'll also start looking at the existing open source quadcopter software out there. What does your parts list look like so far? I noticed a few references to AeroQuad. Are you using this for the hardware platform as well as the software? It looks like ArduCopter at http://diydrones.com is the result of combining ArduQuad and ArduPilot to "combine the best of both". Have you looked at this platform as well? What I'm trying to achieve, in a nutshell, is a quadcopter design which is lift-assisted by a helium or hydrogen balloon, not totally unlike Blimpduino (http://bit.ly/eqyAw4) but more maneuverable. My aim is to greatly increase flight time by reducing the lift burden, and therefore power drain. The quadcopters I've seen so far are advertising 9-15 minute flight times, which limits commercial usefulness considerably.

[Dan Vanderboom]

I just read SparkFun's guide to buying accelerometers and gyros, mentioned earlier in this thread. Has anyone played with the Razor IMU? http://www.sparkfun.com/products/9623 At $125 it's somewhat expensive, but if it's solving for attitude locally and transmitting that serially, wouldn't that remove much of the real-time processing burden from the Netduino? Brandon, is your goal to squeeze all processing into a single microcontroller? Is the response to this data in terms of motor signals just as time critical? Is anyone aware of any examples or code to pull this data stream into the Netduino?

[Brandon G]

kind of a strange idea to combine hydrogen and a quad, whats the lift lb/vol? wonder if you could add hydrogen to the structure without compromising aerodynamics and not blowing up (i would suggest inert). We've taken a pretty good look at all the open source offerings, they are all good but all quite specialized and dont read well. long term goal is to run all on the same mcu but as the hardware and framework stand now, impractical. We are running gyro and motors on netduino, radio on another mcu. The razor is great but defeats the purpose of what we're trying to do, we want to make the predictions and math and understand it, not have it written for us. On the other thread we have another interested user that has written a netmf handler for razor. From here on out i would suggest you write to other thread

[Chris Seto]

9DoF (non stick) is not suitable in any way for quad flight without highly optimized firmware. The AHRS FW is not nearly capable of it.

[Miha Markic]

9DoF (non stick) is not suitable in any way for quad flight without highly optimized firmware. The AHRS FW is not nearly capable of it.

You mean because of the refresh frequency and the fact that it is sending data as ASCII?

[Chris Seto]

There are several reasons: A: As you point out, the stock FS is ASCII. Very, very inefficient B: The FW is not set up at all for immediate Ask/tell response C: Even when modded to ask/tell, the IMU will only respond every 25ms, without architecture changes.

[Miha Markic]

There are several reasons:

A: As you point out, the stock FS is ASCII. Very, very inefficient
B: The FW is not set up at all for immediate Ask/tell response
C: Even when modded to ask/tell, the IMU will only respond every 25ms, without architecture changes.

I am curios about C, what architecture changes you have in mind?

[Chris Seto]

I am curios about C, what architecture changes you have in mind?

Totally changing the way the 9DoF firmware works.... Frankly, there's not much in there I see fit to really keep other than the math.

[Miha Markic]

Totally changing the way the 9DoF firmware works.... Frankly, there's not much in there I see fit to really keep other than the math.

Ah, so you are talking about code. Great.

[Chris Seto]

Ideally, I'd kinda like to completely change the hardware as well, but that's a gripe for another thread..... Like the one about Yonghan and I building our OWN IMUs.

[antoniekruger]

Hi, did this project die or has it moved?

[Nevyn]

Hi, did this project die or has it moved?

Not sure it this project died but there is a rather nice example on The Wolf Bytes blog. You can also find forum postings in the Netduino Go! forum.

regards,
Mark