New member, first build. Could use some advice.

[vinito]

Hi folks.
I just joined this forum today and liking it already. My intro thread is here:
http://www.multirotorforums.com/threads/howdy-from-kansas-city.24401/

I've been working on a build. Here's my list:
S500 frame
APM 2.6 controller w/ NEO-6M GPS & compass
power module
3DR radio telemetry kit
30A "simon series" ESCs (apparently blheli firmware and not simonk)
Tarot 2814/700kv motors
5000mah battery (so far)

So I just walked back in from the backyard attempt at hovering my new MC.
SUCCESS!
I can't say it was "pretty", but it wasn't too bad. Control is pretty twitchy. I think this thing will be a faster craft than I expected. I obviously need to tweak some PID but that too is a bit closer to decent than I expected. So something actually went well for a change. Yaaaay.

I have it set up with just two modes right now - stabilize and loiter. Stabilize is doing what it is supposed to I think, though just not great at it yet. When I switched to loiter, it was OK other than it slowly and steadily ascended. Probably from 5 to 15 feet in 5 to 7 seconds, so not awful but not exactly holding position. I tried it 5 times or so and it did the same thing each time. I'll look at the APM configuration web pages and see if I can glean anything. I guess I could maybe look at a telemetry log can't I? Maybe that would help point to something. I didn't want it to get too high of course and the whole session was probably no higher than 15 feet. Probably 4 to 5 minutes of hovering time though, which is a BIG step in the right direction.

My last look with Mission Planner showed that my compass is way off. I'm not sure what's going on but the last three or four times I've been looking at this, re-calibrating several times and repeating, it just keeps showing it to be off about 40 to 50 degrees! I really need to try it outside way away from everything and see if it's just field interference, but it seems like it should be doing better than that. The module is oriented correctly on the MC, and it's the external module housed in the GPS antenna, which is up on a mast. It also seems like my GPS doesn't like to function well either. Shouldn't it work indoors too at least to some degree? or is that just a no-no? It does seem to pick up at least 6 satellites when it's outdoors. I'm not worried about the GPS as much as the compass right now, I just thought I'd mention it seemed a bit weak maybe since the compass and GPS are tandem inside the disc enclosure thing in case it made a difference.

Well I'll leave it at that for now.

 

[vinito]

I looked into my compass issue a bit more.

After several routines of this to verify my findings to be solid, it does this every time.
Turns out my compass seems to be very sensitive to pitch and roll positions. After a new calibration, it isn't as bad as it was before I guess - only about 10 to 30 degrees off while sitting flat.
Then I noticed while moving the thing around that if I tipped it a bit in roll or pitch directions, the compass direction on Mission Planner would move quite a bit, i.e. if it showed to be off about 20 degrees, then a slight tilt in pitch or roll (maybe 5 or 10 degrees) would correct it or double it depending on which direction it is tilted. So basically the thing doesn't know which way it's going. And by the way, I would set it on a book or something and let it settle in for a while to make sure it wasn't an accelerometer issue. At least I guess I did - it showed direction on Mission Planner within a few seconds and I'd let it sit there for a couple minutes and it was rock-solid the whole time indicating a particular compass direction even if it was wrong.

So wassup wit dat?

Not sure what I need to try here at all. Goofy compass module or am I just ignorant of a configuration setting I need to tweak?

 

[Motopreserve]

For the compass, Have you done the "live" calibration in mission planner? Where you spin it and fill up the globe with dots?

The GPS is definitely not going to work well indoors. I'm using an older model with no compass - it caught some signal sitting right in the window once. Otherwise you need to be outdoors.

You might also want to post specific APM/3DR questions in that section so the experts are sure to chime in.

 

[vinito]

Yea I've done the live calibration a dozen or so times.

Now there's a new problem. I was just trying to set some PID values manually to get it to the point where I could control it good enough for an auto-tune eventually. The damned thing tried a flyaway!!! I lost all radio responsiveness and it started a slow ascent and drift like maybe it went into a GPS mode of some kind and automatically. Maybe I have a failsafe set wrong (I meant to just have it set to land since I'm keeping it very close right now). Anyway, it flew into a tree about 20 feet up and took a little tumble. Broke a landing leg and that's all I can see so far, so I was fortunate. This quadcopter stuff is damn frustrating. I'll keep with it for a while, but if it doesn't become a shorter list of problem after problem after problem, I can see just dumping it for something else. Life is too short, and my to-do list is three lifetimes long anyway.

Uugh.

edit to add:
Yea I checked battery and radio failsafes and the were both set to RTL instead of Land. I'm sure that's my error, but just another in a long line of issues is that my GPS when imported into Google Maps shows the path to be around 5000 feet or something. It's just a spire from my backyard rising into the stratosphere. In Mission Planner my altitude indicates pretty close to correct though. I need to figure out how to load and read logs - I'm having trouble with that. There are logs uploaded to my computer but they are .tlog extension and mission planner only allows to search for .log or .bin files, so that's not even useful it seems to me. Then again, even if I could load it I'm not sure where to look for what I want. I'd just like to know if the thing went into failsafe at some point in the flight so I'd know if that's what happened for sure, even though it seems like that's what must have happened.

UUghhhhhh.

Ok I figured out what happened. I found the proper log and played it back. I was by the house and going back & forth adjusting PID values so telemetry was on which made it easier I think.

Anyway, battery voltage dropped below 10.5V and went into RTL failsafe. I have reset it now to just land in failsafe, but first I have to fix a landing leg and generally check the thing out before further attempts. At least I know what happened, but it sure didn't have a lot of air time before dropping down. Is 10.5V higher than it should be or is that about right? I need to look into battery stuff more, and I guess I can search forums for that specific setting too.

I'll still complain about things going wrong, but honestly despite the depression of crashing it is very cool that with APM and Mission Planner, I was able to pull up data and analyze things to see what happened. Of course there's also the fact that APM made it crash too, but that was my fault anyway.

Oh well.

 

[Motopreserve]

OK, first: If it kicked into RTL why would it take off away from you? Did you launch from somewhere other than where you were standing when it went haywire? If RTL engages - it should be returning to the initial launch position, as long as you waited for it to get GPS lock.

Batteries: how many cells are you using? Regardless, you should stay on the safe side of the 80% rule. Be especially conservative for now until you get to know the batteries, and more used to the MR and its behavior. The 80% rule suggests not letting it drain lower than 80% (total battery or the lowest individual cell). So if you have each cell charged fully at 4.2 volts, multiply that by .80% you would get 3.36. Best to play it safe and not let any one cell dip below 3.7 (or multiply 3.7v by the number of cells you have) for now.

 

[vinito]

I'm not sure about the direction shift other than so far my GPS position on Mission Planner does seem to move around quite a lot - like a 50 foot radius indoors and 15 or 20 feet outdoors. The first thing it did in failsafe was slowly ascend like it's supposed to for RTL. At first it went pretty straight up, then at only about 20 feet it started to drift in position, then trees, then martial arts & gymnastics.

Also, I'm not sure how it picks up the launch position. I would assume that launch position is wherever it was armed, which it drifted away from in RTL. If launch position is determined somewhere else, I don't know where that might have been because I was inside the house, at the west end of the yard and then a little further east when I last launched it. The last place I was working with it, it was powered on for at least 10 minutes until the incident and I had armed, hovered, landed, disarmed, adjusted P values then repeated several times. At any rate, between my own ignorance and the strong possibility that the GPS isn't precise enough to be useful at the moment, I know I can't depend on any missions nor RTL nor geo-fences or the like until I sort that out.

As far as I know, I think I've set things up so far to ignore GPS most of the time other than I have a Loiter mode option enabled by a switch, which I haven't even tried for a while. And I've been making sure the switch is in Stabilize and not Loiter at each pre-flight check.

I'd just like to iron out PID and stability for now, then maybe wade slowly into figuring out GPS problems afterward.
My current goal after current repairs and checking is to get stability kind of close enough, then take it to an open field and see if I can get an auto tune done.

 

[Motopreserve]

Gotcha. The GPS (and therefore RTL) is going to need for the MR to sit outside and take hold of the satellites before you do anything. You need to power up the MR and let it sit for a bit. You should see on mission planner that you have lock (and there are lights on the bottom of the GPS (great place for them!) - and it should tell you a signal strength level I think. You might also need to do a compass calibration (spinning it around all axis) outdoors before you do anything else. Just to make sure its solid and was not effected by anything inside. It should be done away from power lines, interference, etc.

Unfortunately, if RTL is enabled and you somehow have it activated (loss of radio etc), then it will do its thing as long as it had some lock to GPS at some point - whether you use the GPS in any flight modes or not. That lock, if you haven't let it sit long enough, could be in a weird place away from you.

You could try flying without the GPS altogether if it makes you feel safer for testing and getting the PIDs dialed in.

 

[vinito]

Yea that's pretty much what I was guessing regarding GPS and such. Thanks for the clarification.

And for a little more clarity from my end since re-reading my post it sounded kind of vague, I am now set up to fly with no GPS at all and at crash o'clock it was not set up that way, thus the RTL on battery failsafe. I'll be keeping it low and clear of ground clutter for a while so I set it up to just land on failsafe for now. I should at least not have that particular, specific problem until I add complications later. So yes, flying without GPS altogether as you suggested is what I will do until I get a few higher priority tweaks ironed out.

I should have done a solid compass calibration outside today. Too many other things on my mind to remember I guess. The only ones I've done so far have been in the house and compasses of any kind never work well in here for whatever reason. I should have known better.

I need to figure out a landing gear hack now. I sure wish the 3D printer was working dangit. I have some 8mm carbon fiber rods which were earmarked for another project, but I got extra anyway so I think I may just build a whole new landing gear with it tomorrow. It should be a big improvement on the included ones anyway - they have too small a footprint, badly designed "hook toes" perfect for snagging the smallest patch of grass and obviously way to fragile. The thing broke right away in the tree branches ! It hadn't really even hit anything yet, and certainly not with any big forces. Immediately after it started tumbling I saw pieces flying, and it was that landing leg. Didn't take much of a blow at all to snap it.

The saga continues tomorrow.

 

[vinito]

Haven't fixed the landing gear yet, but I did inspect things closer and found that the only other thing that suffered in the crash was the antenna for the telemetry radio. It broke off the board. It was a little tricky, but I was able to hack it a little so I could solder it back into place and it actually turned out pretty decent. It's actually a bit more solidly mounted to the board now than before. Plus ever since I first attached it, I thought it was so long that it was cumbersome. So I cut off the outer half of the plastic antenna cover leaving just the wire to protrude and it seems to be a lot less suicidal-looking now. It really was just sticking out there asking to be broken off.

I have a family dinner to go to this afternoon (easter I guess) so I may not complete the new landing gear today, but if not afterward then tomorrow I'll have it back together. The short time away probably doesn't hurt. It may clear my head a little and nudge things back to the calm and objective side.

 

[Motopreserve]

Another thing I just thought of...

If you are doing PID tuning - I think you'll want to do the basic tuning in "manual" mode. All you need to do is hover and move slight amounts for tuning - so no biggie.

 

[vinito]

Thanks. That's what I thought too. Maybe it's because the thing with default setting is so twitchy I feel it is not to be trusted at all, but I figured I wanted to manually get some PID settings to make it somewhat close before setting it loose with an auto-tune.

I'm already back from the dinner. I guess everybody was busy or tired. But when I got back I looked at the broken leg and figured I had nothing to lose with a simple repair attempt. I had some black ABS filament so I cut some off to use as "welding rod", put an old tip on my soldering iron and set off to just weld it back together since it was a clean break. It worked !!. I think the legs might be made from ABS because it welded together really easily and it's as strong as it was before it broke, thought that's not saying much.

I did another compass calibration outdoors. I turned it on and let is sit for a few minutes, then set Mission planner to start yet another calibration. After going out and moving it around as needed, came back in and clicked the button. Done. But no better and no different. Something has to be faulty as the compass direction on the Flight Data screen always has the thing pointing waaay not where it's pointing, and the module is oriented correctly. Minor rotations on pitch and roll (like +/- 20 degrees) change the compass direction dramatically, like 30 to 50 degrees. It doesn't matter which way I rotate it, the compass direction moves a lot. You expect that with yaw of course, but any rotation in any axis moves the compass a lot. That just can't be right. I've calibrated a couple dozen times and this is always the result. Unless there's an odd parameter off, then I'm assuming I must have a bad compass component.

Outdoors the GPS will get as many as 13 satellites that I have seen, but it still seems to drift around pretty good. I wonder if I should just bin the compass/GPS module and upgrade to a M8N and see if that works better.

Anyway, I may try to get some hovering in and look for a more stable PID despite the compass. It doesn't drift in yaw when I'm out there dinking around. So it holds a heading I guess, it's just not accurate. You would think that just the pitch and roll movements would rotate yaw but it doesn't seem to. Maybe it is and is just kind of centers around level and I just can't see it. That's entirely possible.

 

[vinito]

Well that does it. I think I'm not supposed to be into this.
Earlier I took the thing outside and did a compass calibration as I described. I armed it just to see the motors spin and it all looked OK. All I did was bring it in the house and set it down.
So I take it back out just minutes ago to do a little hovering and it won't arm. Come back inside and I've got bad GPS health, bad compass health and once it spit out a weak accelerometer message of some kind (can't remember what it specifically said).
I took the GPS module off and down to the bench to have a look. It took a hit and the connector housing is cracked, but no amount of fiddling gets rid of the error. I could try soldering the wires and skip the connector altogether, but it should work the way it is if I just plug it in. I may try that but I'm damned sick of this crap. It won't even let me arm it so all I can do is stare at it and think of the money I've wasted on a hunk of junkyard scrap materials. It just sits there and falls apart with no external help. Maybe I can take a blowtorch to it and make a sculpture statement.

I'm going to let it sit in the corner for a while and think about whether or not I want to deal with this anymore. I'm so frustrated with one problem after another after another after another after another that I'm seriously considering just filing it away to a mistake of middle age and move on. After a couple days or so I may decide to stab at it again, but there are so many whack-a-mole problems that IF I decide to try this again, I may just dump a bunch of money at it and buy a pixhawk setup with a better GPS and the whole rotten tamale. At the moment, I'm thinking it's very likely NOT going to go that way.

 

[dazzab]

Check the logs for how much vibration you are getting. The APM will be very flaky if not properly set up to avoid vibration. It's one of the most common issues that bites people. Install it on anti vibration foam, balance your props and then check the vibration level in the logs. Don't fiddle with the default PIDs until you have everything else settled in. It's not necessary. Most copters will fly just fine with the defaults. Changing them before you have some experience tuning can cause you more trouble than good.

 

[Motopreserve]

Dazzub!!! Finally some expert advice! .

I'm new to this APM stuff too - so my suggestions have been based solely on my research.

Just for clarification, is the vibration less critical in any particular flight mode? Or more specifically, which mode do you suggest using to hover in order get some vibration readings, that will have the least chance of drifting away (or worse - crash)?

Vinito, hang in there. I'm sure you're close.

 

[dazzab]

Vibration screws up the accelerometer's ability to give accurate data and to a degree the barometer. So it will effect all modes. The collective thinking with the APM is to run the most generic mode first and then move on to modes that require GPS for positioning. So do your testing in Stabilize, then try Alt Hold, then try Loiter or Pos Hold. As I said before, if your copter doesn't fly with the default PIDs reasonably then you either have a very unusual airframe or something is mechanically wrong with the airframe. Leave PID tweaking as your last task, if at all. Once you get Alt Hold and GPS modes happy give the auto tuning a go as it's very very good in 90% of cases.

 

[Motopreserve]

Great info. Thanks. I had planned on staying in Stabilize for the maiden anyway. But it's good to know how everything is working for future flights.

I'll check the vibe reading after the maiden to see where I'm at.

Appreciate the clarification.

 

[vinito]

Thanks for the replies, but I'm not kidding about reassessing whether this is worth it anymore. I can see why you would advise me to check vibrations and not mess with PID so soon. But the current problem and previous compass issues are there when it's stationary. Not arming means no possibility of checking for vibrations. The quad is slightly different than normal with a bit larger motors & props, though not too drastic.

In retrospect, when the thing went into RTL, once ino auto-land and a few times I switched it briefly into loiter, it actually smoothed out a lot. So the twitchiness I experienced apparently was more my input than PID, which was probably not too bad. More like a rate issue maybe.

Anyway, that's all immaterial if I walk away. It will be at least a couple more days before I calm down enough to objectively consider the situation and decide either way.

I mean look. SO far it's been over $400 and a few dozen hours time for a half-hours flying. I "enjoy the journey" as much as the next guy, but this path has been downhill rocks & thorns so far.

We'll see.

 

[Motopreserve]

Totally understandable that you'd be frustrated. But taking a break, stepping back and taking a breather is good practice. Typically you hit the problem with a fresh outlook and things come together smoother after a little break. It certainly does for me. I'm sure it's the same way with your machining. I know it is when I'm at the mill/lathe and I've tied myself in knots over some math I forgot 20 years ago.

if you decide you'd like to take another crack at it, there will be people here to help.

 

[vinito]

I've reached the point that I'm not so furious as before and if I had the parts in hand, I would probably be into taking another stab at it this weekend. At this point I still don't know if it was a GPS/compass module problem (which seems more likely since the case took enough of a hit that it cracked the housing), or an APM board problem, which is still possible. So maybe over the weekend I can connect some things in a somewhat systematic order and at least check some voltages, check for shorts and cracks in wires or traces, etc. The APM board is mounted to one of those anti-vibe mounts with the four rubber barrel isolators, so I'm guessing it's probably OK. Also regarding vibration, the props were balanced fairly well and there was very little obvious vibration when it was spinning, and that paired with it's isolated mount would seem to minimize the chance of vibrations being an issue for the APM board. But I will check that out anyway once I'm able.

In the meantime, I decided that maybe I'd just enjoy puttering around with a 250 of my own build, so I picked up a "cheap & dirty" CC3D board and made up a kit of parts for a simple sport flier. The CC3D has arrived and I've upgraded the firmware and fiddled with it via the GUI a little and it appears to function fine. I ordered one of the 250FPV ghost frames with the LED setup. Since I enjoyed flying the toys at night, this seemed to fill that want with style. I went with some 12A ESCs and 1804 motors figuring that would adequately fly this light frame and simple components fine for my skill level. I probably went a little over-weight with 1500mah lipo packs (I ordered two), but I think it should pull them around OK and again, I'm not exactly looking for a racer. So no GPS, no FPV, no telemetry other than what the receiver deals with. Just a little light, simple fun flier. Also ordered a receiver to go with my transmitter.

And speaking of radio stuff, I finally received the Taranis today, and that would be enough to inject a charge of interest into anybody I think. As I may have mentioned, I was into RC planes back in the '80s and this thing FAR surpasses the feel and capability of the best radio conceived of back then. The Futaba I borrowed from my buddy feels like a rusty cheese grater compared to this thing. Full disclosure, that Futaba felt kinda rough compared to the nice mid-level 7ch Futaba I bought back in the '80s too (which I pad more for back then than this Taranis cost me today by the way).

So I guess I probably won't bag it all up and set it at the curb after all. Not quite yet. The replacement GPS/compass module is on its way from china and should be here by the 4th of July maybe. I upgraded to a Neo M8N so maybe I can get better results than the one I had even if it was working properly anyway.

I think I may enjoy fiddling with this little 250 for a while. At least I have the ability to tweak the controls to my liking until I get some smoother, funner flying from it than the toys let you have. Too bad I'm not better at hacking. It would be fun to be able to reverse-engineer the Hubsan and set things up so a guy could tweak it to his liking. That would actually be a popular and fun hack I think.

 

[Motopreserve]

Sounds like you took the right kind of step back. Having fun always trumps stressing.

So hopefully you can come back to the APM with a fresh perspective.