|
|
|
|
|
|
|
|
||
|
Quote:
current version 2.4 AQ for the 328 pro mini... with battery monitor builds to 27428, without builds to 26502, by my thinking thats about 900 bytes. Remember one of the reasons that AQ is larger than the others in code space is because we elected in 2.3 and up to convert everything to SI units. This has paid off in huge returns. It effectively lets the same PID values run on multiple frames. (e.g. most using 2.4, use the stock PID's due to this SI change). And when I say SI I mean everything, the DCM, (or Quaternions if you want to use them), sensors, flight controls, etc all return SI units. What was amazing to us was that after we did this, it was literally 2 lines of code to add our attitude (what other's called stable mode), and 2 more lines to add a velocity mode, and 2 more (to add position hold) (none of the later has been release yet, it's what I'm working on). I did of course have to add in the gps library, but really that was it. AQ is also for the most part object oriented with C++, not that I'm a complete fan of that approach it's really not ideal for an embedded solution, but in this case, because of all the platforms that we support, it's been worth the effort. It's pretty highly compartmentalized as well, one of the reasons we can take and build an AHRS only solution from it if we elect to... I think it's great that each software effort has it's own goals and niches for users. No problem there... Alex sorry to hijack the thread, but didn't want to give people the wrong idea before they try one vs. the other... I think this math was just off by a factor of 10 ... Alan |
|
|
Last edited by aadamson; May 30, 2011 at 07:44 PM.
|
|
|
||
|
Quote:
It would also help on Leadfeather's F-35 VTOL. |
|
|
||
|
|
|
You're right. I think your method would be best if you change the yaw to do what you said.
Yaw won't provide symmetrical lift but it may not be a big problem. You could also try standard quad+ control and standard quad+ motor rotations. The smaller fans will provide weaker yaw so you could angle them a little to provide more yaw and see what happens. |
|
|
|
||
|
Quote:
Here's where I got mine www.gressaero.com The camera is mounted on the radio and the noise is just tranfered through the mount form hands operating the gimbles and switchs. |
|
|
||
|
||
|
Quote:
Does this diagram illustrate your idea? |
|
|
Last edited by RCvertt; May 30, 2011 at 08:32 PM.
|
|
|
||
|
Quote:
See schematic for ITG-3200 Breakout board - resistors R1 and R2. |
|
|
||
|
||
|
Quote:
Having converting all sensors to SI is a great idea and I fully support it, but all those floating point calculations... maybe 16 MIPS AVR is not the right choice. I was thinking doing quite opposite - converting Alex's MultiWii code to use only integer math (long int should give enough precision) just for the sake of speed. Remember all of that is just my humble opinion. |
|
|
||
|
||
|
Quote:
|
|
|
Last edited by RCvertt; May 31, 2011 at 12:36 AM.
|
|
|
||
|
Quote:
The actual code to read the voltage is only a few bytes, it's the code to diddle LED's, beepers, blip the motors in an up down fashion and if necessary *autoland* you if you are pulling a battery too far down that takes the full 900 bytes. Seems pretty small for what it does... Yep, lots of floating point, and we just went to a new time executive. Instead of doing tasks at what I call a sliding window version of a time executive... We do it at exact intervals.... e.g. 5 mins after the hour, instead of 5mins after all the tasks finish ahead of it. This means that we have a minimum window of time to complete execution in, but it also helps dramatically PID loops and other timing related loops (AHRS) that expect specific hz of execution. Not trying to be everything to all people, and why we have an ARM project fully underway, but I think you'll see some good things from us in the coming weeks/months. Ok, I'm done with any other replies on this topic as I want to respect Alex's topic... but I do wish you'd get your facts straight before you put out propoganda that simple isn't true or contains only partial truths... feel free to take this up on the AQ forums if you have such strong feelings... Btw, Im in Atlanta, as well .. Alan |
|
|
||
|
|
|
Will one of the attached diagrams work for the accel? I tried to copy what the resistors do for the gyro.
|
|
Last edited by RCvertt; May 31, 2011 at 01:41 AM.
|
|
|||
|
Here is a short vid of MultiWiiCopter with PipoISU on my microquad:
Did anyone ever look further into into ADXL345? It seems to be more noisy than the "Bosch brothers" Jussi |
||
|
|||