Showing posts with label RTL. Show all posts
Showing posts with label RTL. Show all posts

Thursday, 29 August 2013

Current and voltage meter in the mail today!

This is what I got in the mail today. I ordered it from GoodLuckBuy and it was shipped out 10 days ago. This still means that the average time to delivery from China is about 9-10 days depending on holidays I suppose.

The little device in the middle gets hooked up and man-in-the-middles the power coming from the battery and tells the flight controller how much juice it has left. There is an excellent integration instruction at ArduCopters wiki but this is roughly how its meant to work shown in the picture to the right. There are some settings in Mission Planner that allows the flight controller to conduct an emergency landing if the voltage falls too low. Remember that I previously written that I've set my ESC's to cut power at 3V in order to spare the batterys health. This means that I should make an emergency landing prior to that. Other than an emergency landing is also to order the flight controller to RTL. However, we all know how that ended last time..
Talking about failing RTL's, I yesterday think I've got a clue why the GPS has been acting strange lately. By accident, I looked at the quad's heading according to the flight controller and it was 90 degrees wrong. This means that I've mounted the GPS 90 degrees wrong which makes the flight controller do things the wrong way.

Another update is that I received an e-mail this morning from the author of MegaPirateNG with a link to MPNG-3.0.1-R2 :). I still haven't downloaded the code but I sure will within the next few days.

I've also got another package in the mail about a week ago, which I still haven't written about. I will do so in a separate blog post due to its large content. I also have a few surprise upgrades so stay tuned ;)

 

Friday, 9 August 2013

Another one bites the dust

There has been a few days since I last wrote. To that, I can only say: laptop + morning coffee = less blogging and even less field testing of quad.

Today was however the first time since the resurrection that I've done some serious testing. I wanted to try out the RTL (return to launch) function that went AWOL last time I tried.
The first thing that happened today once I fired up the quad and was lifting off was that after 30 meters or so, one of the blades of the front left propeller broke off. I didn't realize this at first but what happened was that I heard a ill sounding noise as if something broke followed by a ill sounding motor engine. The quad then slowly started banking towards left and at that point, I couldn't stop it from flying further to the left. I made a controlled crash landing as I couldn't stabilize the quad. Once I located it again I quickly realized why it behaved as it did; it had close to no lifting power on the left side and therefore, I couldn't stop the motion towards left.

This is one of the 8x45 propellers I recently bought and from the breaking point it looks like it might had been delivered with a defect. I am no expert in this field but you can examine the picture for your self. Anyhow, these are cheap propellers and I guess I buy more as they may break. The positive outcome from this accident is that I now know that a failing propeller isn't an instant fatal crash but very recoverable.

I've considered buying these carbon fiber propellers instead of the plastic propellers I've been using until now but they are fairly pricy and I think I need more practice and less close calls to fund these. Carbon fiber propellers won't take any small nudges without splintering into pieces nor will the plastic once but the plastic once are six times cheaper..

Thursday, 25 July 2013

RTL fly away...

Yesterday I went away to find a field large enough to do some serious RTL (return to launch) and navigation testing. This requires space without trees too close. I found some nice fields and started testing. I successfully tested both RTL and three way point navigation before my last RTL went sour...

I flew quite far away as I previously had done and hit the RTL. The normal procedure then is for the quad to gain height to 15 meters, which is default. Then head to its home position and descend to about 3 meters and loiter. This time, instead of turning towards home, it started drifting to the left. I quickly intervened to save the quad by counter steering rightwards. The quad then rotated 90-180 degrees making the direction impossible to see from this distance and disappeared behind those trees according to the drawing. I then ran towards the highway and never saw it again.


I've since then been covering the grounds around the point where it disappeared for about 8 hours in total, checking trees and the ground where it might have crashed without result.

It might have kept going in that direction ending up on the other side of the field as the program does not include anything about landing, but keeping 15 meters until the destination has been reached. It might also have taken a short trip on the field, then turned around ending up somewhere else. I still think the highest possibility is that it sitting in one of those trees along the highway.

Later today or tomorrow, I'm going to print posters asking people to contact me if they find my quad. Hope that gives some results. At this point, I don't think I can afford a new one, which leaves me with a pile of spare parts that will not be used..

Wednesday, 24 July 2013

The new propellers was a success

Last night I tried out the new 8 inch propellers along with a fresh motor to replace the last faulty one. The x550 now spins like a cat and flies like a bird. I trimmed the flight controller for the new propellers, which wasn't all that much to change and it's very stable.

After the basic trimming I started working on navigation and RTL (return to launch). These also has a set of P, I and D parameters for navigation. There is a good guide over at the ArduCopter wiki. However, I didn't have much success in trimming the RTL and the x550 always overshot or navigated wrong. Once both batteries were drained, I figured out that I'd trimmed using wrong parameters, dooh! I had been trimming the HLD_LAT/LON instead of LOITER_LAT/LON. Will have to do this at some other point where I can find a good field to try it out at.

Link to the guide: https://code.google.com/p/arducopter/wiki/AC2_loiter_PID